Canny vs Productboard

Agent access

1. ai-native userPoint an agent at llms.txt or agent-oriented docs

   weight 2 · round drawn
   Canny✓fullprobed9/10

   A direct probe confirms llms.txt exists and returns 200 with developer documentation content at developers.canny.io/llms.txt, and Canny also publishes agent-oriented docs (MCP server docs, API reference) explicitly designed for LLM/agent consumption. Missing for 10: independent third-party confirmation that agents successfully consume this llms.txt in practice.

   • [probe] “PROBE llms.txt: HTTP 200 at https://developers.canny.io/llms.txt # Canny Developer Documentation > Developer resources for integrating with…”
   • [claimed-docs] “Install Docs | Canny”
   • [claimed-docs] “Canny API Reference”
   • [probe] “official MCP server documented at https://canny.io/features/mcp”

   proof ↗⚑ flag

   Productboard✓fullprobed9/10

   A probe confirms Productboard serves a working llms.txt at developer.productboard.com/llms.txt (HTTP 200) with a description of the API and instructions to append .md to any doc page for agent-readable versions, which is exactly the agent-oriented docs pattern the story asks for. Missing for 10: independent/community confirmation that agents actually consume this llms.txt successfully in practice.

   • [probe] “PROBE llms.txt: HTTP 200 at https://developer.productboard.com/llms.txt # Productboard Public API Documentation > Productboard Public API …”

   proof ↗⚑ flag

2. ai-native userRun the product headlessly / in CI for automation

   weight 2 · round to Productboard
   Canny~partialprobed3/10

   Canny exposes a REST API reference and API-driven install/SDK docs, which implies some programmatic/automatable access, but there is no explicit documentation of CI usage, a CLI, or headless automation workflows; the MCP server is described as requiring interactive secure sign-in rather than API-key/CI-friendly auth. missing for 10: explicit CI/headless usage examples, CLI or scriptable automation docs, non-interactive auth for MCP/API use in pipelines.

   • [claimed-docs] “Canny API Reference”
   • [claimed-docs] “Add Canny's MCP server to ChatGPT, Claude, or Cursor with a secure sign-in. No API keys, no engineering.”
   • [probe] “PROBE openapi: all candidate paths 404 (https://developers.canny.io/openapi.json, https://developers.canny.io/swagger.json, https://develope…”

   proof ↗⚑ flag

   Productboard~partialclaimed6/10

   Productboard's REST API supports personal-access-token authentication, cursor-based pagination, and webhooks, all of which enable scripted/headless use in CI or automation pipelines (docs-43, docs-51, docs-5). However, there is no explicit documentation of CI/CD integration, SDKs, or example automation scripts targeting headless pipelines. Missing for 10: explicit CI/CD pipeline examples, official SDK/CLI for automation, and documented rate-limit/queueing guidance for unattended batch jobs.

   • [claimed-docs] “Use a personal access token to authenticate API requests. Best for internal tools, scripts, and getting started quickly.”
   • [claimed-docs] “Create, search, and manage feedback, documents, customers, and product data through Productboard's REST API.”
   • [claimed-docs] “For large result sets, both the list and search endpoints support cursor-based pagination.”
   • [claimed-docs] “The Webhooks API allows you to subscribe to events in Productboard and receive real-time notifications when data changes.”
   • [claimed-docs] “The Productboard REST API uses cursor-based pagination for all endpoints that return a list of results.”

   proof ↗⚑ flag

3. ai-native userPlug MCP servers into this product so it can use their tools

   weight 3 · round drawn
   Canny—none0/10

   All evidence shows Canny exposes an MCP *server* that external AI tools (ChatGPT, Claude, Cursor) connect to — the reverse direction. There is no evidence that Canny itself acts as an MCP client that plugs in external MCP servers to use their tools.

   • [claimed-docs] “Add Canny's MCP server to ChatGPT, Claude, or Cursor with a secure sign-in. No API keys, no engineering.”
   • [claimed-docs] “Connect Canny to ChatGPT, Claude, and Cursor. Ask questions about customer feedback, take action on it, and ground your AI workflows in real…”
   • [claimed-docs] “Canny MCP Server: Customer Feedback in Your AI Tools”
   • [probe] “official MCP server documented at https://canny.io/features/mcp”

   proof ↗⚑ flag

   Productboard—none0/10

   All evidence describes Productboard shipping its own MCP server so external agents/tools can call Productboard's capabilities (docs-10 through docs-13, probe-3) — this is the reverse role. There is no evidence that Productboard itself can act as an MCP client and plug in external MCP servers to use their tools.

   • [claimed-docs] “Now you can use any agent to add new features, reschedule an initiative, or start building off a product specification.”
   • [claimed-docs] “Update custom drivers and other data fields on features and initiatives (like value, risk, impact, confidence, and effort) to automate prior…”
   • [claimed-docs] “Add features to initiatives, move an initiative to a new release, or create key results beneath an objective.”
   • [claimed-docs] “Access your latest product specifications from coding agents and AI prototyping tools to initiate the product delivery process.”
   • [probe] “official MCP server documented at https://www.productboard.com/integrations/mcp-server/”

   proof ↗⚑ flag

4. ai-native userConnect an agent via an official MCP server

   weight 3 · round to Productboard
   Canny✓fullprobed8/10

   Canny documents an official MCP server that connects to ChatGPT, Claude, and Cursor via secure sign-in (no API keys), enabling agents to query feedback data and take actions like updating statuses and replying to customers. Missing for 10: independent/hands-on third-party corroboration beyond vendor docs.

   • [claimed-docs] “Add Canny's MCP server to ChatGPT, Claude, or Cursor with a secure sign-in. No API keys, no engineering.”
   • [claimed-docs] “Ask your LLM what customers asked for and understand the context behind each request. Dig into top themes, ARR impact across ideas, and quer…”
   • [claimed-docs] “Move ideas through your workflow without leaving your LLM. Update statuses, link deals, organize requests, and reply to customers.”
   • [claimed-docs] “Connect Canny to ChatGPT, Claude, and Cursor. Ask questions about customer feedback, take action on it, and ground your AI workflows in real…”
   • [claimed-docs] “Use Canny as the data layer for your AI workflows. Draft PRDs from real requests and build the features with Claude, generate reports for le…”
   • [probe] “official MCP server documented at https://canny.io/features/mcp”

   proof ↗⚑ flag

   Productboard✓fullprobed9/10

   Productboard ships an official, well-documented MCP server with a single connection endpoint (mcp.productboard.com), browser-based sign-in via existing login, OAuth with PKCE/Dynamic Client Registration, admin controls for enabling MCP and role-based access, and documented use cases for reading/writing features, initiatives, specs, and prioritization data via coding agents and AI tools. Missing for 10: independent/third-party hands-on corroboration beyond vendor docs.

   • [claimed-docs] “The Productboard MCP server lets your coding agent read product specs straight from Productboard — and write back to them.”
   • [claimed-docs] “You connect your agent to a single address: `https://mcp.productboard.com`. Signing in happens through your browser, using your existing Pro…”
   • [claimed-docs] “Under **Who can use it**, choose which roles may connect an agent — for example, makers with admin access, makers, contributors, or viewers.”
   • [claimed-docs] “Authenticate MCP (Model Context Protocol) clients using Dynamic Client Registration and the authorization code flow with PKCE. No client sec…”
   • [claimed-docs] “Turn on Enable MCP for this workspace.”
   • [claimed-docs] “Create and update features, subfeatures, initiatives, releases, objectives, key results, companies, users, feedback, documents, and more dir…”
   • [claimed-docs] “The Productboard MCP server bridges those divides by allowing you to employ the AI tools and agents you’re already using to access and updat…”
   • [probe] “official MCP server documented at https://www.productboard.com/integrations/mcp-server/”

   proof ↗⚑ flag

5. ai-native userUse an official CLI

   weight 2 · round drawn
   Canny—none0/10

   Evidence shows Canny has a REST API and an official MCP server, but nothing indicates an official CLI tool for AI-native workflows; no mention of any command-line interface anywhere in the docs.

   ⚑ flag

   Productboard—none0/10

   Productboard documents a REST API, webhooks, and an MCP server, but no evidence anywhere mentions an official CLI tool for AI-native workflows.

   ⚑ flag

6. ai-native userDrive the product through a documented public API

   weight 3 · round to Productboard
   Canny✓fullprobed8/10

   Canny publishes a documented public API Reference (canny-docs-16) with a full developer docs site (canny-docs-15, canny-probe-1 llms.txt confirming developer docs), plus SDK/widget install docs and an official MCP server for LLM-driven interaction (canny-docs-3, canny-docs-9, canny-probe-3), enabling AI-native users to query and act on data programmatically. Missing for 10: an accessible OpenAPI/Swagger spec (canny-probe-2 shows all candidate OpenAPI endpoints 404) and independent hands-on corroboration of API usage.

   • [claimed-docs] “Canny API Reference”
   • [claimed-docs] “Install Docs | Canny”
   • [probe] “PROBE llms.txt: HTTP 200 at https://developers.canny.io/llms.txt # Canny Developer Documentation > Developer resources for integrating with…”
   • [claimed-docs] “Add Canny's MCP server to ChatGPT, Claude, or Cursor with a secure sign-in. No API keys, no engineering.”
   • [claimed-docs] “Connect Canny to ChatGPT, Claude, and Cursor. Ask questions about customer feedback, take action on it, and ground your AI workflows in real…”
   • [probe] “official MCP server documented at https://canny.io/features/mcp”
   • [probe] “PROBE openapi: all candidate paths 404 (https://developers.canny.io/openapi.json, https://developers.canny.io/swagger.json, https://develope…”

   proof ↗⚑ flag

   Productboard✓fullprobed9/10

   Productboard documents a comprehensive REST API v2 with authentication (token and OAuth PKCE), cursor-based pagination, CRUD across features, feedback, documents, customers, teams, webhooks, and configuration/discovery endpoints, plus quickstart and Postman guides, and an llms.txt probe confirms machine-readable docs are live. This is a thorough, well-documented public API suitable for AI-native/agent-driven use, reinforced by a dedicated MCP server for agent access. missing for 10: no independent/third-party corroboration of real-world API usage or hands-on developer confirmation beyond vendor docs, and an OpenAPI spec file was not found at expected locations.

   • [claimed-docs] “Create, search, and manage feedback, documents, customers, and product data through Productboard's REST API.”
   • [claimed-docs] “For large result sets, both the list and search endpoints support cursor-based pagination.”
   • [claimed-docs] “Create, read, update, delete, and search customer feedback, complete with text content, tags, owners, source metadata, and customer attribut…”
   • [claimed-docs] “List and search feedback with structured filters by owner, creator, tags, date range, and status. For large result sets, both the list and s…”
   • [claimed-docs] “Use a personal access token to authenticate API requests. Best for internal tools, scripts, and getting started quickly.”
   • [claimed-docs] “Authenticate users via the OAuth 2.0 authorization code flow with consent screens and refreshable tokens.”
   • [claimed-docs] “The Productboard REST API uses cursor-based pagination for all endpoints that return a list of results.”
   • [claimed-docs] “Get up and running with the Productboard API v2 in Postman in under 5 minutes.”
   • [claimed-docs] “This Quickstart walks you through your first API call using Productboard's REST API v2. In just a few steps, you’ll authenticate, send a req…”
   • [probe] “PROBE llms.txt: HTTP 200 at https://developer.productboard.com/llms.txt # Productboard Public API Documentation > Productboard Public API …”
   • [probe] “PROBE openapi: all candidate paths 404 (https://developer.productboard.com/openapi.json, https://developer.productboard.com/swagger.json, ht…”

   proof ↗⚑ flag

7. ai-native userIssue scoped/least-privilege API credentials for an agent

   weight 2 · round to Productboard
   Canny—none0/10

   Canny's docs mention a general API Reference and an MCP server that uses 'secure sign-in, no API keys' rather than any scoped or least-privilege credential system for agents; there is no evidence of granular API scopes, role-based tokens, or agent-specific permission controls.

   • [claimed-docs] “Add Canny's MCP server to ChatGPT, Claude, or Cursor with a secure sign-in. No API keys, no engineering.”
   • [claimed-docs] “Canny API Reference”

   proof ↗⚑ flag

   Productboard~partialclaimed5/10

   Productboard offers multiple credential types for agents (personal access tokens, OAuth 2.0 authorization-code flow with consent screens, and PKCE-based OAuth for MCP clients) and lets admins restrict which roles may connect an agent, giving some coarse-grained control over agent access. However, there is no documented support for fine-grained scopes (e.g., read-only vs write, per-resource permissions) on tokens or OAuth grants, so true least-privilege credential issuance isn't evidenced. Missing for 10: explicit scoped/permission-limited token or OAuth scope configuration, documentation of granular read/write or resource-level access control for agent credentials, and independent confirmation of scope enforcement.

   • [claimed-docs] “Under **Who can use it**, choose which roles may connect an agent — for example, makers with admin access, makers, contributors, or viewers.”
   • [claimed-docs] “Authenticate MCP (Model Context Protocol) clients using Dynamic Client Registration and the authorization code flow with PKCE. No client sec…”
   • [claimed-docs] “Use a personal access token to authenticate API requests. Best for internal tools, scripts, and getting started quickly.”
   • [claimed-docs] “Authenticate users via the OAuth 2.0 authorization code flow with consent screens and refreshable tokens.”
   • [claimed-docs] “You connect your agent to a single address: `https://mcp.productboard.com`. Signing in happens through your browser, using your existing Pro…”

   proof ↗⚑ flag

8. ai-native userBuild against official SDKs

   weight 2 · round to Canny
   Canny~partialprobed6/10

   Canny provides an official JavaScript SDK for identify calls, a documented REST API reference, and even an AI-friendly llms.txt file for developer docs, all of which support building against official SDKs. However, probes show no discoverable OpenAPI/swagger spec (404s across all candidate paths), and evidence only shows one language SDK (JS) rather than a broader official SDK ecosystem. missing for 10: OpenAPI/machine-readable API spec, multi-language SDKs, independent developer corroboration of SDK robustness.

   • [claimed-docs] “Install our JavaScript SDK to identify your users. This way, when they leave feedback on Canny, it'll be tied to their existing user account…”
   • [claimed-docs] “Canny API Reference”
   • [probe] “PROBE llms.txt: HTTP 200 at https://developers.canny.io/llms.txt # Canny Developer Documentation > Developer resources for integrating with…”
   • [probe] “PROBE openapi: all candidate paths 404 (https://developers.canny.io/openapi.json, https://developers.canny.io/swagger.json, https://develope…”
   • [claimed-docs] “Install Docs | Canny”

   proof ↗⚑ flag

   Productboard~partialprobed5/10

   Productboard publishes a well-documented REST API v2 (quickstart, Postman collection, auth via PAT/OAuth, pagination, webhooks) and an MCP server for agent integration, which lets AI-native builders programmatically interact with the product. However, there's no evidence of official language-specific client SDKs (e.g., Python/JS packages) and the OpenAPI spec probe returned 404s, suggesting no auto-generatable SDK artifact. Missing for 10: official SDK packages in common languages, published OpenAPI/swagger spec, independent developer confirmation of SDK usage.

   • [claimed-docs] “This Quickstart walks you through your first API call using Productboard's REST API v2. In just a few steps, you’ll authenticate, send a req…”
   • [claimed-docs] “Get up and running with the Productboard API v2 in Postman in under 5 minutes.”
   • [claimed-docs] “Use a personal access token to authenticate API requests. Best for internal tools, scripts, and getting started quickly.”
   • [claimed-docs] “Authenticate users via the OAuth 2.0 authorization code flow with consent screens and refreshable tokens.”
   • [probe] “PROBE openapi: all candidate paths 404 (https://developer.productboard.com/openapi.json, https://developer.productboard.com/swagger.json, ht…”
   • [claimed-docs] “The Productboard MCP server lets your coding agent read product specs straight from Productboard — and write back to them.”
   • [claimed-docs] “You connect your agent to a single address: `https://mcp.productboard.com`. Signing in happens through your browser, using your existing Pro…”

   proof ↗⚑ flag

9. ai-native userSubscribe to events via webhooks

   weight 2 · round to Productboard
   Canny—none0/10

   The evidence pack includes API reference, integrations, and MCP server docs but contains no mention of webhooks or event subscription mechanisms anywhere. This is an applicable axis for a feedback platform with an API, but no evidence supports it, so verdict is none.

   ⚑ flag

   Productboard✓fullclaimed8/10

   Productboard documents a dedicated Webhooks API with explicit create/list/delete endpoints for subscribing to entity change events and receiving real-time notifications, matching the story directly. Missing for 10: independent/hands-on corroboration of webhook reliability and payload/event-type documentation depth beyond the reference stubs.

   • [claimed-docs] “The Webhooks API allows you to subscribe to events in Productboard and receive real-time notifications when data changes.”
   • [claimed-docs] “Subscribe to entity change events and receive real-time notifications at your endpoint.”
   • [claimed-docs] “Creates a new webhook subscription to be actively notified on each change in the specified entities.”
   • [claimed-docs] “Permanently deletes a webhook subscription. No further notifications will be sent after deletion.”

   proof ↗⚑ flag

Agentic features

1. ai-native userGet AI-generated insights and suggestions from my data inside the product

   weight 2 · round to Canny
   Canny✓fullprobed7/10

   Canny provides an official MCP server enabling users to ask their LLM for insights (top themes, ARR impact, roadmap queries) directly from feedback data, and to act on it (update statuses, draft PRDs, run agents) — matching the AI-native insight-generation story. missing for 10: no independent/hands-on validation of insight quality or accuracy, and no first-party in-app AI analytics dashboard beyond the MCP-based LLM interaction.

   • [claimed-docs] “Ask your LLM what customers asked for and understand the context behind each request. Dig into top themes, ARR impact across ideas, and quer…”
   • [claimed-docs] “Move ideas through your workflow without leaving your LLM. Update statuses, link deals, organize requests, and reply to customers.”
   • [claimed-docs] “Connect Canny to ChatGPT, Claude, and Cursor. Ask questions about customer feedback, take action on it, and ground your AI workflows in real…”
   • [claimed-docs] “Use Canny as the data layer for your AI workflows. Draft PRDs from real requests and build the features with Claude, generate reports for le…”
   • [probe] “official MCP server documented at https://canny.io/features/mcp”

   proof ↗⚑ flag

   Productboard~partialclaimed6/10

   Productboard automatically detects insights within feedback and links them to related feature ideas, showing native AI-generated insight surfacing inside the product, and the MCP server lets AI agents act on that data (prioritization, specs, initiatives). However, evidence is limited to marketing claims with no hands-on validation of insight quality or breadth of AI suggestions beyond feedback-to-feature linking. Missing for 10: independent/hands-on verification of AI insight accuracy, broader AI-generated suggestions (e.g., roadmap prioritization, summarization) beyond feedback linking, and detail on how insights are surfaced in UI.

   • [claimed-docs] “Productboard automatically detects insights within feedback that can be linked to related feature ideas.”
   • [claimed-docs] “Now you can use any agent to add new features, reschedule an initiative, or start building off a product specification.”
   • [claimed-docs] “Update custom drivers and other data fields on features and initiatives (like value, risk, impact, confidence, and effort) to automate prior…”
   • [claimed-docs] “Access your latest product specifications from coding agents and AI prototyping tools to initiate the product delivery process.”

   proof ↗⚑ flag

2. ai-native userSet up automations that run autonomously in the background

   weight 2 · round to Canny
   Canny~partialclaimed6/10

   Canny's Autopilot feature automatically captures feedback from Gong, Intercom, Slack, Fathom, Fireflies, and review platforms without manual triggering, and native integrations (Jira, Linear, GitHub, etc.) keep statuses synced automatically — these qualify as background automations. However, this is limited to feedback ingestion/sync rather than a general-purpose agentic automation framework the user can configure for arbitrary autonomous tasks, and the MCP-based actions (updating statuses, replying to customers) appear to require active LLM invocation rather than running unattended in the background. Missing for 10: evidence of user-configurable, scheduled/triggered autonomous workflows beyond feedback capture, and confirmation that MCP-driven actions can run without a live agent session.

   • [claimed-docs] “Send new conversations to Autopilot to capture feedback you haven't tracked yet.”
   • [claimed-docs] “Automatically capture feedback from Gong, Intercom, Slack, and dozens of other sources, or directly from customers on your Portal.”
   • [claimed-docs] “Autopilot captures feedback from support, sales, and review platforms automatically. Native integrations keep the rest of your stack in sync…”
   • [claimed-docs] “Autopilot pulls feature requests out of your Fathom call transcripts automatically.”
   • [claimed-docs] “Capture product feedback from your Fireflies sales call transcripts.”
   • [claimed-docs] “Send feedback to Jira, link it to existing issues, and sync statuses. Only for Jira Cloud.”
   • [claimed-docs] “Send feedback to Linear, link it to existing issues, and sync statuses.”
   • [claimed-docs] “Send feedback to GitHub, link it to existing issues, and sync statuses.”

   proof ↗⚑ flag

   Productboard~partialprobed4/10

   Productboard exposes webhooks, a REST API, and an MCP server that let external agents or scripts react to events and modify data, which are the building blocks for automation, but there's no documented native feature for scheduling or running autonomous background workflows within Productboard itself — all examples show agent-initiated or event-triggered actions rather than self-running background automations. missing for 10: a first-party 'automation' or scheduled-agent feature, evidence of unattended/recurring execution, and confirmation that MCP-driven actions can run without a live agent session.

   • [claimed-docs] “The Webhooks API allows you to subscribe to events in Productboard and receive real-time notifications when data changes.”
   • [claimed-docs] “Now you can use any agent to add new features, reschedule an initiative, or start building off a product specification.”
   • [claimed-docs] “Update custom drivers and other data fields on features and initiatives (like value, risk, impact, confidence, and effort) to automate prior…”
   • [claimed-docs] “Access your latest product specifications from coding agents and AI prototyping tools to initiate the product delivery process.”
   • [probe] “official MCP server documented at https://www.productboard.com/integrations/mcp-server/”

   proof ↗⚑ flag

3. ai-native userDelegate tasks to a built-in AI assistant inside the product

   weight 3 · round to Canny
   Canny~partialclaimed4/10

   Canny ships 'Autopilot,' an automated AI feature that captures and categorizes feedback from support/sales/review sources without user intervention, and users can 'send new conversations to Autopilot' to have it process feedback — a limited form of task delegation to a built-in AI component. However, this is not a general-purpose conversational assistant embedded in the Canny UI; the richer AI interaction (asking questions, taking actions, drafting PRDs) is delivered via MCP connecting to *external* LLMs like ChatGPT/Claude/Cursor, not a native in-product assistant. Missing for 10: evidence of an interactive, conversational built-in assistant UI, an example of delegating arbitrary multi-step tasks to it, and independent/hands-on confirmation of Autopilot's task-delegation behavior beyond marketing copy.

   • [claimed-docs] “Send new conversations to Autopilot to capture feedback you haven't tracked yet.”
   • [claimed-docs] “Automatically capture feedback from Gong, Intercom, Slack, and dozens of other sources, or directly from customers on your Portal.”
   • [claimed-docs] “Autopilot captures feedback from support, sales, and review platforms automatically. Native integrations keep the rest of your stack in sync…”
   • [claimed-docs] “Autopilot pulls feature requests out of your Fathom call transcripts automatically.”
   • [claimed-docs] “Capture product feedback from your Fireflies sales call transcripts.”

   proof ↗⚑ flag

   Productboard—none0/10

   Evidence covers an MCP server that lets external AI agents (Claude, ChatGPT, coding agents) call into Productboard's data and workflows, plus automatic insight detection — but nothing describes a built-in, in-product AI assistant/chat that a user can delegate tasks to. MCP is an external-agent connector, not an assistant embedded in Productboard's own UI, and insight-detection is automatic, not user-delegated task execution.

   • [claimed-docs] “Now you can use any agent to add new features, reschedule an initiative, or start building off a product specification.”
   • [claimed-docs] “Update custom drivers and other data fields on features and initiatives (like value, risk, impact, confidence, and effort) to automate prior…”
   • [claimed-docs] “Add features to initiatives, move an initiative to a new release, or create key results beneath an objective.”
   • [claimed-docs] “Productboard automatically detects insights within feedback that can be linked to related feature ideas.”
   • [probe] “official MCP server documented at https://www.productboard.com/integrations/mcp-server/”

   proof ↗⚑ flag

4. ai-native userOperate the product with natural-language commands

   weight 2 · round to Productboard
   Canny~partialprobed6/10

   Canny ships an official MCP server enabling natural-language interaction via ChatGPT, Claude, or Cursor—users can ask questions about feedback, update statuses, link deals, and draft PRDs using plain language instead of UI/API calls. However, this is scoped to feedback/roadmap management rather than full product operation, and there's no independent/hands-on verification of the NL command experience quality. missing for 10: independent corroboration of MCP usability, breadth of natural-language commands beyond feedback workflows, evidence of reliability/limitations in practice.

   • [claimed-docs] “Add Canny's MCP server to ChatGPT, Claude, or Cursor with a secure sign-in. No API keys, no engineering.”
   • [claimed-docs] “Ask your LLM what customers asked for and understand the context behind each request. Dig into top themes, ARR impact across ideas, and quer…”
   • [claimed-docs] “Move ideas through your workflow without leaving your LLM. Update statuses, link deals, organize requests, and reply to customers.”
   • [claimed-docs] “Connect Canny to ChatGPT, Claude, and Cursor. Ask questions about customer feedback, take action on it, and ground your AI workflows in real…”
   • [claimed-docs] “Use Canny as the data layer for your AI workflows. Draft PRDs from real requests and build the features with Claude, generate reports for le…”
   • [probe] “official MCP server documented at https://canny.io/features/mcp”

   proof ↗⚑ flag

   Productboard✓fullprobed8/10

   Productboard ships an official MCP server (mcp.productboard.com) plus Spark natural-language interface that let any AI agent (Claude, ChatGPT, coding agents) read specs and create/update features, initiatives, releases, and objectives via natural language, and Spark supports natural-language queries against Slack/Amplitude/Notion. This is first-party documented and includes install/auth/OAuth flows and worked examples. Missing for 10: independent/hands-on third-party verification of the natural-language command experience and coverage of edge cases or limitations in real usage.

   • [claimed-docs] “Now you can use any agent to add new features, reschedule an initiative, or start building off a product specification.”
   • [claimed-docs] “Create and update features, subfeatures, initiatives, releases, objectives, key results, companies, users, feedback, documents, and more dir…”
   • [claimed-docs] “The Productboard MCP server lets your coding agent read product specs straight from Productboard — and write back to them.”
   • [claimed-docs] “You connect your agent to a single address: `https://mcp.productboard.com`. Signing in happens through your browser, using your existing Pro…”
   • [claimed-docs] “Pull a spec, talk through the plan with your agent, and build. No setup beyond the connection.”
   • [claimed-docs] “The Productboard MCP server bridges those divides by allowing you to employ the AI tools and agents you’re already using to access and updat…”
   • [claimed-docs] “Query Amplitude analytics through natural language in Productboard Spark.”
   • [claimed-docs] “Search, message, and act in Slack through natural language in Productboard Spark.”
   • [probe] “official MCP server documented at https://www.productboard.com/integrations/mcp-server/”

   proof ↗⚑ flag

Api quality

1. ai-native userExplore an interactive API reference with runnable examples

   weight 2 · round drawn
   Canny—none0/10

   Canny has an API Reference page (canny-docs-16) but there is no evidence it includes runnable/interactive examples (e.g., try-it console, live code execution), and probes for a standard OpenAPI spec file all returned 404, suggesting no discoverable interactive spec.

   • [claimed-docs] “Canny API Reference”
   • [probe] “PROBE openapi: all candidate paths 404 (https://developers.canny.io/openapi.json, https://developers.canny.io/swagger.json, https://develope…”

   proof ↗⚑ flag

   Productboard—none0/10

   Evidence confirms Productboard has API documentation (developer.productboard.com, llms.txt) but no mention of an interactive reference with runnable/try-it examples, and probes for an OpenAPI/Swagger spec (which typically power such consoles) all returned 404. Missing for 10: any documentation or screenshot of an interactive console, runnable code snippets, or Swagger/Redoc-style UI.

   • [probe] “PROBE llms.txt: HTTP 200 at https://developer.productboard.com/llms.txt # Productboard Public API Documentation > Productboard Public API …”
   • [probe] “PROBE openapi: all candidate paths 404 (https://developer.productboard.com/openapi.json, https://developer.productboard.com/swagger.json, ht…”

   proof ↗⚑ flag

2. ai-native userDownload a machine-readable API spec (OpenAPI or equivalent)

   weight 2 · round drawn
   Canny—none0/10

   Canny has an API Reference page but there is no evidence of a downloadable OpenAPI/Swagger spec; explicit probes for openapi.json, swagger.json, and similar well-known paths all returned 404.

   • [claimed-docs] “Canny API Reference”
   • [probe] “PROBE openapi: all candidate paths 404 (https://developers.canny.io/openapi.json, https://developers.canny.io/swagger.json, https://develope…”

   proof ↗⚑ flag

   Productboard—none0/10

   While Productboard documents a REST API v2 and even a Postman quickstart, no first-party page offers a downloadable OpenAPI/Swagger file, and a direct probe of the standard OpenAPI spec locations (openapi.json, swagger.json, etc.) on developer.productboard.com all returned 404, indicating no machine-readable spec is actually published for retrieval.

   • [probe] “PROBE openapi: all candidate paths 404 (https://developer.productboard.com/openapi.json, https://developer.productboard.com/swagger.json, ht…”
   • [claimed-docs] “Get up and running with the Productboard API v2 in Postman in under 5 minutes.”
   • [claimed-docs] “This Quickstart walks you through your first API call using Productboard's REST API v2. In just a few steps, you’ll authenticate, send a req…”

   proof ↗⚑ flag

3. ai-native userTest against a sandbox environment without touching production data

   weight 1 · round drawn
   Canny—none0/10

   No evidence of a sandbox/staging environment or test mode separate from production data; all docs reference live API, widgets, and MCP integrations acting on real feedback data.

   ⚑ flag

   Productboard—none0/10

   No evidence pack item mentions a sandbox, staging, test workspace, or non-production environment for API/MCP testing; all documentation refers to production workspace access and live data operations.

   ⚑ flag

4. ai-native userRely on versioned APIs with a documented deprecation policy

   weight 2 · round drawn
   Canny—none0/10

   Canny has an API Reference and install docs, but there is no evidence of API versioning or a documented deprecation policy; the OpenAPI spec probe returned 404s at all candidate paths, suggesting no machine-readable versioned spec either.

   • [claimed-docs] “Canny API Reference”
   • [probe] “PROBE openapi: all candidate paths 404 (https://developers.canny.io/openapi.json, https://developers.canny.io/swagger.json, https://develope…”

   proof ↗⚑ flag

   Productboard—none0/10

   Evidence describes REST API features, pagination, webhooks, and MCP server capabilities but contains no mention of API versioning scheme or a documented deprecation policy; OpenAPI spec probes also 404, offering no discoverable version metadata.

   • [probe] “PROBE openapi: all candidate paths 404 (https://developer.productboard.com/openapi.json, https://developer.productboard.com/swagger.json, ht…”

   proof ↗⚑ flag

Reporting

1. ai-native userUse AI to declutter raw feedback and surface the most impactful opportunities

   weight 3 · round to Canny
   Canny~partialprobed6/10

   Canny's Autopilot and MCP features let AI surface themes, ARR impact, and top requests from raw feedback across sources (support, sales, reviews), and the MCP integration lets users query/act on this via ChatGPT/Claude. However, evidence is entirely vendor-authored marketing copy with no independent validation of how well the AI actually declutters/prioritizes feedback, no detail on the underlying scoring/impact methodology, and no hands-on or third-party corroboration. missing for 10: independent/hands-on evidence of AI decluttering quality, methodology detail behind 'ARR impact' scoring, third-party validation.

   • [claimed-docs] “Ask your LLM what customers asked for and understand the context behind each request. Dig into top themes, ARR impact across ideas, and quer…”
   • [claimed-docs] “Send new conversations to Autopilot to capture feedback you haven't tracked yet.”
   • [claimed-docs] “Automatically capture feedback from Gong, Intercom, Slack, and dozens of other sources, or directly from customers on your Portal.”
   • [claimed-docs] “Use Canny as the data layer for your AI workflows. Draft PRDs from real requests and build the features with Claude, generate reports for le…”
   • [claimed-docs] “Autopilot captures feedback from support, sales, and review platforms automatically. Native integrations keep the rest of your stack in sync…”
   • [probe] “official MCP server documented at https://canny.io/features/mcp”

   proof ↗⚑ flag

   Productboard~partialclaimed5/10

   Productboard's docs claim automatic AI-driven insight detection linking feedback to feature ideas and a centralized searchable feedback repository, which directly supports decluttering raw feedback, and its MCP/API layer lets automation update prioritization fields (value, risk, impact, effort). However, there is no detailed documentation of an AI feature that explicitly ranks or surfaces 'most impactful opportunities' beyond basic insight linking, and no independent/hands-on evidence validating this capability. Missing for 10: dedicated AI opportunity-scoring/ranking feature documentation, independent user validation of insight quality, and clearer detail on how 'impactful' is determined.

   • [claimed-docs] “Productboard automatically detects insights within feedback that can be linked to related feature ideas.”
   • [claimed-docs] “Access one searchable repository for all product requests and feedback from colleagues and customers.”
   • [claimed-docs] “Ensure developers are working on what really matters by sharing customer context and only sending fully-validated features into delivery.”
   • [claimed-docs] “Update custom drivers and other data fields on features and initiatives (like value, risk, impact, confidence, and effort) to automate prior…”

   proof ↗⚑ flag

2. product-managerAccess customer context alongside feedback so I prioritize what really matters

   weight 2 · round drawn
   Canny✓fullclaimed8/10

   Canny lets PMs identify users/companies via SDK, attach custom fields, and view ARR impact and themes alongside feedback, plus MCP integration lets PMs query customer context (themes, ARR impact, roadmap) directly from LLM tools. Missing for 10: independent/hands-on validation of how customer context is surfaced in the actual UI (only docs/marketing evidence, no third-party corroboration).

   • [claimed-docs] “Install our JavaScript SDK to identify your users. This way, when they leave feedback on Canny, it'll be tied to their existing user account…”
   • [claimed-docs] “Ask your LLM what customers asked for and understand the context behind each request. Dig into top themes, ARR impact across ideas, and quer…”
   • [claimed-docs] “Custom Fields | Canny”
   • [claimed-docs] “Install Companies | Canny”
   • [claimed-docs] “Connect Canny to ChatGPT, Claude, and Cursor. Ask questions about customer feedback, take action on it, and ground your AI workflows in real…”
   • [claimed-docs] “Use Canny as the data layer for your AI workflows. Draft PRDs from real requests and build the features with Claude, generate reports for le…”

   proof ↗⚑ flag

   Productboard✓fullclaimed8/10

   Productboard associates feedback with users/companies and links notes to features to quantify demand, provides a single searchable repository for requests/feedback, and surfaces automatic insight detection with citations tying feedback to evidence, directly supporting customer-context-informed prioritization. missing for 10: independent/hands-on validation from PM users of the customer-context view and quantitative proof of prioritization impact.

   • [claimed-docs] “Associate feedback with users and companies, and connect it to features, components, and other product hierarchy entities.”
   • [claimed-docs] “Ensure developers are working on what really matters by sharing customer context and only sending fully-validated features into delivery.”
   • [claimed-docs] “Access one searchable repository for all product requests and feedback from colleagues and customers.”
   • [claimed-docs] “Productboard automatically detects insights within feedback that can be linked to related feature ideas.”
   • [claimed-docs] “Understand critical feedback trends and act fast on new opportunities. Productboard includes citations for every finding, so decisions are a…”
   • [claimed-docs] “Notes can be linked to features to quantify demand and inform prioritization.”

   proof ↗⚑ flag

3. ai-native userAsk an AI assistant questions about customer feedback themes and their revenue impact

   weight 3 · round to Canny
   Canny✓fullprobed8/10

   Canny's MCP server explicitly lets users ask their LLM about customer feedback themes and ARR/revenue impact ('Dig into top themes, ARR impact across ideas'), connecting via ChatGPT/Claude/Cursor without engineering work. This directly matches the story's ask-AI-about-feedback-and-revenue-impact use case. Missing for 10: independent/hands-on verification beyond vendor docs, and concrete examples of the AI answering real revenue-impact queries.

   • [claimed-docs] “Add Canny's MCP server to ChatGPT, Claude, or Cursor with a secure sign-in. No API keys, no engineering.”
   • [claimed-docs] “Ask your LLM what customers asked for and understand the context behind each request. Dig into top themes, ARR impact across ideas, and quer…”
   • [claimed-docs] “Connect Canny to ChatGPT, Claude, and Cursor. Ask questions about customer feedback, take action on it, and ground your AI workflows in real…”
   • [claimed-docs] “Use Canny as the data layer for your AI workflows. Draft PRDs from real requests and build the features with Claude, generate reports for le…”
   • [probe] “official MCP server documented at https://canny.io/features/mcp”

   proof ↗⚑ flag

   Productboard~partialprobed3/10

   Productboard's MCP server lets AI agents create/update features and initiatives, and it auto-detects insights linking feedback to features, suggesting some AI-native querying capability over feedback data. However, there is no evidence of an AI assistant answering natural-language questions about feedback themes or their revenue impact—the Analytics API is scoped to workspace usage metrics, not customer feedback/revenue correlation. Missing for 10: evidence of a conversational AI/analytics feature that surfaces feedback theme summaries, revenue impact metrics, or any linkage between feedback and revenue data.

   • [claimed-docs] “Now you can use any agent to add new features, reschedule an initiative, or start building off a product specification.”
   • [claimed-docs] “Productboard automatically detects insights within feedback that can be linked to related feature ideas.”
   • [claimed-docs] “The Analytics API provides access to usage data that helps you understand how Productboard is being adopted and used across your workspace.”
   • [probe] “official MCP server documented at https://www.productboard.com/integrations/mcp-server/”

   proof ↗⚑ flag

1. ai-native userPerform bulk operations across many items at once

   weight 2 · round to Productboard
   Canny—none0/10

   Canny's docs show per-item actions (update status, link deals, sync to trackers) via its API and MCP server, but no evidence of any batch/bulk endpoint or MCP action that operates across many items at once — missing for 10: bulk update/merge API, batch MCP tool calls, any documented multi-item operation.

   ⚑ flag

   Productboard~partialclaimed6/10

   Productboard's REST API supports cursor-based pagination for list/search endpoints and explicitly mentions 'bulk operations' for document management, plus general create/update/delete across features, feedback, and other entities that an AI-native user (via scripts or MCP-connected agents) could loop through for bulk actions. However, there's no documented native batch/bulk-update endpoint (e.g., a single call to update many items at once) outside of the one document-specific mention, and the MCP server examples focus on single-item actions rather than explicit multi-item bulk operations. Missing for 10: dedicated bulk-update/bulk-create API endpoints across all entity types, explicit bulk-operation support in the MCP tool set, and independent evidence of an AI agent performing true bulk operations at scale.

   • [claimed-docs] “Create and manage editor documents and product document boards, with support for bulk operations, document templates, and thread and comment…”
   • [claimed-docs] “For large result sets, both the list and search endpoints support cursor-based pagination.”
   • [claimed-docs] “List and search feedback with structured filters by owner, creator, tags, date range, and status. For large result sets, both the list and s…”
   • [claimed-docs] “The Productboard REST API uses cursor-based pagination for all endpoints that return a list of results.”
   • [claimed-docs] “Create and update features, subfeatures, initiatives, releases, objectives, key results, companies, users, feedback, documents, and more dir…”
   • [claimed-docs] “The Productboard MCP server lets your coding agent read product specs straight from Productboard — and write back to them.”

   proof ↗⚑ flag

2. ai-native userDefine rules that trigger actions automatically on events

   weight 3 · round to Productboard
   Canny—none0/10

   Evidence shows Canny automatically captures feedback from integrations and syncs statuses, but there is no documentation of user-defined rules/triggers that fire actions on specific events (e.g., an automation/workflow rule engine). Missing for 10: explicit rule-builder or trigger-condition-action documentation, evidence of customizable automation logic, examples of event-based automation beyond passive data capture.

   • [claimed-docs] “Automatically capture feedback from Gong, Intercom, Slack, and dozens of other sources, or directly from customers on your Portal.”
   • [claimed-docs] “Autopilot captures feedback from support, sales, and review platforms automatically. Native integrations keep the rest of your stack in sync…”

   proof ↗⚑ flag

   Productboard~partialclaimed4/10

   Productboard's Webhooks API lets users subscribe to entity-change events and receive real-time notifications, and the REST/MCP APIs allow subsequent actions to be scripted in response — but this requires building custom integration code rather than a native declarative 'if this event then this action' rules engine inside the product. Missing for 10: evidence of a built-in no-code automation/rules builder, conditional logic across multiple triggers, and any UI for defining action chains without custom webhook-consumer code.

   • [claimed-docs] “The Webhooks API allows you to subscribe to events in Productboard and receive real-time notifications when data changes.”
   • [claimed-docs] “Subscribe to entity change events and receive real-time notifications at your endpoint.”
   • [claimed-docs] “Creates a new webhook subscription to be actively notified on each change in the specified entities.”
   • [claimed-docs] “Permanently deletes a webhook subscription. No further notifications will be sent after deletion.”
   • [claimed-docs] “Create and update features, subfeatures, initiatives, releases, objectives, key results, companies, users, feedback, documents, and more dir…”

   proof ↗⚑ flag

3. ai-native userSchedule recurring jobs or workflows

   weight 2 · round drawn
   Canny—none0/10

   Canny's evidence covers feedback capture, integrations, and an MCP server for querying/acting on feedback via LLM tools, but there is no mention of scheduling recurring jobs, cron-like automation, or persistent scheduled workflows. Autopilot automates capture but this is event-triggered, not schedule-based recurring automation.

   ⚑ flag

   Productboard—none0/10

   Evidence shows Productboard has webhooks (event-driven push), a REST API, and an MCP server for agent-driven actions, but nothing describes scheduling recurring jobs, cron-like automation, or timed/repeating workflows. The axis applies (automation platforms commonly offer this) but no supporting evidence exists.

   ⚑ flag

Changelog

1. product-managerServe a fully branded public changelog page on a custom domain

   weight 2 · round to Canny
   Canny~partialclaimed4/10

   Canny explicitly offers a Changelog product for announcing shipped features (canny-docs-8, canny-docs-14, canny-docs-26), confirming the axis applies, but none of the evidence describes custom-domain hosting or brand customization (logo, colors, CSS) for the changelog page. Missing for 10: explicit documentation of custom domain support, branding/theming options, and any independent confirmation of a fully white-labeled changelog experience.

   • [claimed-docs] “Announce what you shipped, show customers you heard them, and turn closed feedback into retention and upgrades.”
   • [claimed-docs] “Product Changelog Software and Release Notes Tool”
   • [claimed-docs] “Product Changelog Software and Release Notes Tool | Canny”

   proof ↗⚑ flag

   Productboard—none0/10

   No evidence in the pack mentions a public changelog feature, custom domain hosting, or branding for a changelog page; all evidence concerns API, webhooks, MCP server, and integrations. This is a plausible feature for a product management tool (Productboard historically has a Portal/changelog-adjacent feature area), so the axis applies, but nothing in the evidence supports it.

   ⚑ flag

2. product-managerLet customers subscribe to email updates when new changelog entries are published

   weight 1 · round drawn
   Canny—none0/10

   Evidence confirms Canny has a Changelog feature for announcing shipped updates (canny-docs-8, canny-docs-14, canny-docs-26), but none of the evidence describes a customer-facing email subscription mechanism for new changelog entries. Missing for 10: any documentation of an email subscribe/opt-in feature tied to changelog publishing.

   • [claimed-docs] “Announce what you shipped, show customers you heard them, and turn closed feedback into retention and upgrades.”
   • [claimed-docs] “Product Changelog Software and Release Notes Tool”
   • [claimed-docs] “Product Changelog Software and Release Notes Tool | Canny”

   proof ↗⚑ flag

   Productboard—none0/10

   The evidence pack covers Productboard's API, webhooks, MCP server, and integrations, but contains no mention of a customer-facing changelog feature or an email subscription mechanism for changelog updates. This is a plausible feature for a product-management tool, so the axis applies, but no evidence supports it.

   ⚑ flag

3. developerShowcase new updates within my product using an embeddable changelog widget

   weight 2 · round to Canny
   Canny✓fullclaimed8/10

   Canny explicitly offers a Changelog product for announcing shipped updates plus documented widget embed guides for web and mobile, directly matching the story. Missing for 10: independent/hands-on corroboration of the widget's ease-of-use and no evidence of changelog-specific widget customization details beyond general docs.

   • [claimed-docs] “Announce what you shipped, show customers you heard them, and turn closed feedback into retention and upgrades.”
   • [claimed-docs] “Product Changelog Software and Release Notes Tool”
   • [claimed-docs] “Embed a Canny widget into your website”
   • [claimed-docs] “Embed a Canny widget into your mobile app”
   • [claimed-docs] “Product Changelog Software and Release Notes Tool | Canny”

   proof ↗⚑ flag

   Productboard—none0/10

   Productboard's evidence pack covers REST API, webhooks, MCP server, and integrations for feedback/roadmap data, but there is no mention of a public-facing embeddable changelog widget for showcasing product updates to end users. This is a plausible feature for a product management tool (e.g., similar to 'What's New' widgets), but no evidence shows it exists.

   ⚑ flag

4. product-managerPublish a changelog announcing shipped features to close the loop with customers

   weight 3 · round to Canny
   Canny✓fullclaimed7/10

   Canny explicitly offers a Changelog product ('Announce what you shipped, show customers you heard them, and turn closed feedback into retention and upgrades') as a dedicated feature area, directly matching the story of closing the loop with customers on shipped features. Missing for 10: no evidence of workflow automation linking closed feedback items to changelog entries, no screenshots/hands-on detail on publishing flow, and no independent/user corroboration beyond marketing copy.

   • [claimed-docs] “Announce what you shipped, show customers you heard them, and turn closed feedback into retention and upgrades.”
   • [claimed-docs] “Product Changelog Software and Release Notes Tool”
   • [claimed-docs] “Product Changelog Software and Release Notes Tool | Canny”

   proof ↗⚑ flag

   Productboard—none0/10

   While Productboard emphasizes closing the loop between specification, delivery, and outcome (productboard-docs-23) and offers rich APIs/integrations for pushing data to delivery tools, there is no evidence of a dedicated changelog/release-notes publishing feature for announcing shipped features to customers. Missing for 10: any mention of a customer-facing changelog or release-notes publishing tool, templates, or distribution channel (email, in-app, public page).

   ⚑ flag

Lifecycle automation

1. product-managerBuild automated workflows for SLAs, trial extensions, and discount offers triggered by feedback events

   weight 2 · round drawn
   Canny—none0/10

   Canny's evidence covers feedback capture, integrations (Jira, Linear, GitHub, etc.), changelog, and an MCP server for querying/acting on feedback data, but there is no mention of automated workflow builders for SLAs, trial extensions, or discount offers triggered by feedback events. Missing for 10: any SLA automation, trial-extension logic, discount-offer triggers, or general conditional workflow/rules engine tied to feedback events.

   ⚑ flag

   Productboard—none0/10

   Productboard offers generic Webhooks/REST APIs for feedback events, but there is no evidence of built-in or documented workflow automation for SLAs, trial extensions, or discount offers — these are outside Productboard's feedback/roadmap product scope and no such workflow builder or triggers are described.

   • [claimed-docs] “The Webhooks API allows you to subscribe to events in Productboard and receive real-time notifications when data changes.”
   • [claimed-docs] “Create, search, and manage feedback, documents, customers, and product data through Productboard's REST API.”

   proof ↗⚑ flag

2. product-managerAutomatically notify customers when the status of their requested feature changes

   weight 2 · round drawn
   Canny~partialclaimed4/10

   Canny's changelog feature is described as a way to 'show customers you heard them' when shipping features (canny-docs-8), which implies some notification workflow, but the evidence never explicitly confirms an automated per-voter email/notification triggered specifically when a status changes on their requested feature — the closest matches are Discord alerts for team feedback (canny-intdir-1) and changelog publishing, not confirmed automatic customer-facing status-change alerts. Missing for 10: explicit documentation of automatic subscriber/voter notifications tied to individual status transitions, and any independent/hands-on confirmation of this workflow.

   • [claimed-docs] “Announce what you shipped, show customers you heard them, and turn closed feedback into retention and upgrades.”
   • [claimed-docs] “Get instant notifications in Discord about your feedback.”

   proof ↗⚑ flag

   Productboard~partialclaimed4/10

   Productboard exposes a Webhooks API that fires real-time notifications when entity data (e.g., feature status) changes, and feedback entries can be linked to customers, so a PM could build an automation to notify requesters. However, there is no evidence of a native, built-in 'notify customer when feature status changes' feature — it would require custom integration work using the webhook/API to reach customers. Missing for 10: a documented native customer-notification feature (e.g., automated customer emails on status change), evidence of end-to-end customer messaging rather than just internal event triggers, and any hands-on confirmation this workflow is used in practice.

   • [claimed-docs] “The Webhooks API allows you to subscribe to events in Productboard and receive real-time notifications when data changes.”
   • [claimed-docs] “Subscribe to entity change events and receive real-time notifications at your endpoint.”
   • [claimed-docs] “Creates a new webhook subscription to be actively notified on each change in the specified entities.”
   • [claimed-docs] “Associate feedback with users and companies, and connect it to features, components, and other product hierarchy entities.”
   • [claimed-docs] “Notes can be linked to features to quantify demand and inform prioritization.”

   proof ↗⚑ flag

3. product-managerSegment product release announcements by user attributes to send targeted release notes

   weight 2 · round drawn
   Canny—none0/10

   Canny's changelog feature (canny-docs-8, canny-docs-14) covers publishing release notes, and separate docs mention custom fields and company identification (canny-docs-20, canny-docs-23), but no evidence ties these together to show audience segmentation of changelog/release announcements by user attributes for targeted delivery.

   • [claimed-docs] “Announce what you shipped, show customers you heard them, and turn closed feedback into retention and upgrades.”
   • [claimed-docs] “Product Changelog Software and Release Notes Tool”
   • [claimed-docs] “Custom Fields | Canny”
   • [claimed-docs] “Install Companies | Canny”

   proof ↗⚑ flag

   Productboard—none0/10

   Evidence shows release grouping and general product data management, but nothing about segmenting or sending targeted release-note announcements based on user attributes. Productboard's docs cover feedback, integrations, MCP server, and API endpoints, none of which address customer-facing release note personalization or segmentation.

   ⚑ flag

Centralized repository

1. product-managerAutomatically capture customer feedback mentioned in sales calls, support tickets, and chat conversations

   weight 3 · round to Canny
   Canny✓fullclaimed8/10

   Canny's Autopilot feature explicitly automates capturing feedback from sales calls (Gong, Fireflies, Fathom transcripts), support tickets (Intercom, support platforms), and chat (Slack, Discord), aggregating them into Canny as a unified feedback layer. This directly matches the story's capture-intake theme across multiple channel types with named integrations. Missing for 10: independent/hands-on verification of Autopilot's accuracy or reliability in practice, and clearer detail on chat conversation capture beyond Slack/Discord notifications.

   • [claimed-docs] “Automatically capture feedback from Gong, Intercom, Slack, and dozens of other sources, or directly from customers on your Portal.”
   • [claimed-docs] “Autopilot captures feedback from support, sales, and review platforms automatically. Native integrations keep the rest of your stack in sync…”
   • [claimed-docs] “Autopilot pulls feature requests out of your Fathom call transcripts automatically.”
   • [claimed-docs] “Capture product feedback from your Fireflies sales call transcripts.”
   • [claimed-docs] “Send new conversations to Autopilot to capture feedback you haven't tracked yet.”

   proof ↗⚑ flag

   Productboard~partialclaimed5/10

   Productboard's API/webhooks let feedback be routed in from CRMs, internal tools, and other sources, and it offers a single searchable feedback repository with automatic insight detection, but there's no direct evidence of native, purpose-built connectors that auto-capture from sales calls, support tickets, or chat tools (e.g. Gong, Zendesk, Intercom) — only generic API-based ingestion is documented. missing for 10: named native integrations for sales-call/support-ticket/chat capture, evidence of automatic (not developer-built) capture pipelines, and independent confirmation of these integrations working in practice.

   • [claimed-docs] “Create, search, and manage feedback, documents, customers, and product data through Productboard's REST API.”
   • [claimed-docs] “Developers and technical teams use it to route feedback from proprietary CRMs, internal tools, and other non-native sources into Productboar…”
   • [claimed-docs] “Associate feedback with users and companies, and connect it to features, components, and other product hierarchy entities.”
   • [claimed-docs] “Access one searchable repository for all product requests and feedback from colleagues and customers.”
   • [claimed-docs] “Productboard automatically detects insights within feedback that can be linked to related feature ideas.”

   proof ↗⚑ flag

2. product-managerAccess one searchable repository for all product requests and feedback from colleagues and customers

   weight 3 · round to Productboard
   Canny✓fullclaimed7/10

   Canny's core product is a centralized feedback repository: it captures requests directly from customers via Portal/widgets and automatically ingests feedback from Gong, Intercom, Slack, Fathom, Fireflies, and other tools via Autopilot, consolidating everything with statuses, syncing, and MCP-based query access for PMs to interrogate the whole feedback base. Missing for 10: explicit documentation of a search feature/UI within the repository, and independent/hands-on corroboration beyond vendor docs.

   • [claimed-docs] “Automatically capture feedback from Gong, Intercom, Slack, and dozens of other sources, or directly from customers on your Portal.”
   • [claimed-docs] “Autopilot captures feedback from support, sales, and review platforms automatically. Native integrations keep the rest of your stack in sync…”
   • [claimed-docs] “Autopilot pulls feature requests out of your Fathom call transcripts automatically.”
   • [claimed-docs] “Capture product feedback from your Fireflies sales call transcripts.”
   • [claimed-docs] “Ask your LLM what customers asked for and understand the context behind each request. Dig into top themes, ARR impact across ideas, and quer…”
   • [claimed-docs] “Connect Canny to ChatGPT, Claude, and Cursor. Ask questions about customer feedback, take action on it, and ground your AI workflows in real…”
   • [claimed-docs] “Embed a Canny widget into your website”
   • [claimed-docs] “Embed a Canny widget into your mobile app”

   proof ↗⚑ flag

   Productboard✓fullclaimed9/10

   Productboard's own docs explicitly state the exact capability ("Access one searchable repository for all product requests and feedback from colleagues and customers"), reinforced by feedback API features supporting associating feedback with users/companies and searching across it. missing for 10: independent/hands-on corroboration beyond vendor docs.

   • [claimed-docs] “Access one searchable repository for all product requests and feedback from colleagues and customers.”
   • [claimed-docs] “Associate feedback with users and companies, and connect it to features, components, and other product hierarchy entities.”
   • [claimed-docs] “Productboard automatically detects insights within feedback that can be linked to related feature ideas.”

   proof ↗⚑ flag

3. product-managerLink individual feedback items to the specific customers who submitted them

   weight 2 · round drawn
   Canny✓fullclaimed8/10

   Canny's JS SDK identifies users so feedback is tied to their existing account (canny-docs-2), and Install Companies docs support linking feedback to customer/company records (canny-docs-23). SSO and secure identify further ensure feedback is authenticated to real customers (canny-docs-19, canny-docs-21, canny-docs-27, canny-docs-28). Missing for 10: independent/hands-on corroboration beyond vendor docs, and explicit UI screenshots showing a feedback item's linked customer profile.

   • [claimed-docs] “Install our JavaScript SDK to identify your users. This way, when they leave feedback on Canny, it'll be tied to their existing user account…”
   • [claimed-docs] “Install Companies | Canny”
   • [claimed-docs] “How to Implement Single Sign-On”
   • [claimed-docs] “Secure Identify Requests Install Docs”

   proof ↗⚑ flag

   Productboard✓fullclaimed8/10

   Productboard's API and product docs explicitly support associating feedback with customers/companies, with dedicated attribution fields and relationship endpoints (create/list/delete), plus UI-level customer attribution on feedback capture. missing for 10: no independent/hands-on corroboration of the linking UX beyond vendor docs, and no explicit detail on how attribution surfaces in the feedback list/search UI.

   • [claimed-docs] “Associate feedback with users and companies, and connect it to features, components, and other product hierarchy entities.”
   • [claimed-docs] “Create, read, update, delete, and search customer feedback, complete with text content, tags, owners, source metadata, and customer attribut…”
   • [claimed-docs] “Associate feedback with users and companies, and connect it to features, components, and other product hierarchy entities. Dedicated endpoin…”
   • [claimed-docs] “Access one searchable repository for all product requests and feedback from colleagues and customers.”

   proof ↗⚑ flag

4. product-managerReply to customers directly from within the feedback record to close the loop

   weight 2 · round to Canny
   Canny~partialclaimed5/10

   Evidence confirms replying to customers is possible via the MCP integration ("reply to customers" listed as an action from LLM tools), implying the underlying feedback record supports replies, but there is no direct documentation of the native in-app reply-to-customer workflow within the feedback record itself. missing for 10: direct product docs/screenshots of replying to a commenter within the Canny feedback record UI, evidence of email/notification triggered to the customer, and independent/hands-on confirmation of this closing-the-loop workflow.

   • [claimed-docs] “Move ideas through your workflow without leaving your LLM. Update statuses, link deals, organize requests, and reply to customers.”
   • [claimed-docs] “Connect Canny to ChatGPT, Claude, and Cursor. Ask questions about customer feedback, take action on it, and ground your AI workflows in real…”

   proof ↗⚑ flag

   Productboard—none0/10

   Evidence covers feedback capture, tagging, linking to features, and API/webhook management, but nothing describes a customer-facing reply mechanism from within a feedback record. The 'close the loop' mention (docs-23) refers to specification-to-shipped tracking, not customer communication, so this specific capability is unevidenced.

   ⚑ flag

Embeddable widget

1. product-managerCustomize the feedback widget's label and colors to match my brand

   weight 1 · round to Canny
   Canny~partialprobed5/10

   Canny provides docs for embedding a web/mobile feedback widget (canny-docs-17, canny-docs-18) and its dev docs index mentions 'widget customization' (canny-probe-1), implying some branding controls, but no evidence explicitly confirms label text or color customization options. missing for 10: explicit documentation of label/text customization, explicit color/theme customization settings, and independent confirmation of brand-matching capability.

   • [claimed-docs] “Embed a Canny widget into your website”
   • [claimed-docs] “Embed a Canny widget into your mobile app”
   • [probe] “PROBE llms.txt: HTTP 200 at https://developers.canny.io/llms.txt # Canny Developer Documentation > Developer resources for integrating with…”

   proof ↗⚑ flag

   Productboard—none0/10

   No evidence in the pack addresses a feedback widget for capturing customer input, let alone branding customization (label/color) for it; all evidence covers API, MCP, and third-party integrations.

   ⚑ flag

2. developerEmbed a feedback widget on my website with a single script tag and no build step

   weight 3 · round to Canny
   Canny~partialclaimed6/10

   Canny documents a website widget install page (canny-docs-17) confirming embeddable widget support, but the evidence doesn't specify a single-script-tag, no-build-step method — docs also mention SDK install, SSO, and security/identify requests that suggest additional setup steps beyond a trivial snippet. missing for 10: actual snippet/code sample showing single script tag, confirmation that no build step or bundler is required, independent/hands-on verification of ease of embed.

   • [claimed-docs] “Embed a Canny widget into your website”
   • [claimed-docs] “Install our JavaScript SDK to identify your users. This way, when they leave feedback on Canny, it'll be tied to their existing user account…”
   • [claimed-docs] “Install Docs | Canny”

   proof ↗⚑ flag

   Productboard—none0/10

   No evidence of an embeddable feedback widget with a script tag; the evidence pack only covers Productboard's REST API, webhooks, MCP server, and internal feedback repository features, none of which describe a client-side embed widget.

   ⚑ flag

3. developerIdentify logged-in users when they submit feedback so it ties to their existing account

   weight 2 · round to Canny
   Canny✓fullclaimed9/10

   Canny's JS SDK explicitly identifies users so feedback ties to their existing account (canny-docs-2), backed by dedicated SSO and secure identify-request docs (canny-docs-19, canny-docs-21/28) plus widget/mobile install docs (canny-docs-17,18). Missing for 10: independent/hands-on verification beyond vendor docs.

   • [claimed-docs] “Install our JavaScript SDK to identify your users. This way, when they leave feedback on Canny, it'll be tied to their existing user account…”
   • [claimed-docs] “How to Implement Single Sign-On”
   • [claimed-docs] “Secure Identify Requests Install Docs”
   • [claimed-docs] “Embed a Canny widget into your website”
   • [claimed-docs] “Embed a Canny widget into your mobile app”

   proof ↗⚑ flag

   Productboard~partialclaimed5/10

   Productboard's API explicitly supports associating feedback with existing users and companies, and docs mention routing feedback from CRMs/internal tools into user records (productboard-docs-4, productboard-docs-2), which implies developers can tie submitted feedback to an existing account. However, there is no evidence of a specific mechanism for identifying a 'logged-in' end user at submission time (e.g., SSO/JWT-based identification in a feedback widget or portal). Missing for 10: documentation of end-user authentication/identification flow for feedback submission, widget/portal-specific user linking, and independent confirmation of this behavior.

   • [claimed-docs] “Associate feedback with users and companies, and connect it to features, components, and other product hierarchy entities.”
   • [claimed-docs] “Developers and technical teams use it to route feedback from proprietary CRMs, internal tools, and other non-native sources into Productboar…”

   proof ↗⚑ flag

Public portal

1. developerAdd a link in my product that directs users to the feedback portal

   weight 1 · round to Canny
   Canny✓fullclaimed9/10

   Canny's docs explicitly instruct developers to add a link to the Canny feedback site in their app, plus provide widget embed options for web and mobile as alternatives/complements to a plain link. Missing for 10: independent/hands-on confirmation beyond first-party docs.

   • [claimed-docs] “In your application, add a link to your Canny feedback site. Put the link somewhere it's easy to find.”
   • [claimed-docs] “Embed a Canny widget into your website”
   • [claimed-docs] “Embed a Canny widget into your mobile app”
   • [claimed-docs] “Install Docs | Canny”

   proof ↗⚑ flag

   Productboard—none0/10

   The evidence pack covers Productboard's API, webhooks, MCP server, and third-party integrations, but contains no mention of a customer-facing feedback Portal or an embeddable link/widget that developers could add to their product to direct users to submit feedback. This capability is plausible for a product-feedback tool but is simply unevidenced here.

   ⚑ flag

2. product-managerHost a public feedback portal where customers submit and vote on ideas

   weight 3 · round to Canny
   Canny✓fullclaimed8/10

   Canny's core product is a public feedback portal embeddable via widget (web/mobile) where customers submit and vote on ideas, with SSO/identify integration and a changelog to close the loop; this is well documented across install and widget docs. Missing for 10: no independent/hands-on third-party review confirming the voting/portal UX in practice, and no explicit customer-facing voting mechanic quote (only 'submit feedback on your Portal').

   • [claimed-docs] “Embed a Canny widget into your website”
   • [claimed-docs] “Embed a Canny widget into your mobile app”
   • [claimed-docs] “In your application, add a link to your Canny feedback site. Put the link somewhere it's easy to find.”
   • [claimed-docs] “Automatically capture feedback from Gong, Intercom, Slack, and dozens of other sources, or directly from customers on your Portal.”
   • [claimed-docs] “Announce what you shipped, show customers you heard them, and turn closed feedback into retention and upgrades.”
   • [claimed-docs] “How to Implement Single Sign-On”

   proof ↗⚑ flag

   Productboard—none0/10

   The evidence pack covers Productboard's API, webhooks, MCP server, and internal feedback repository, but contains no mention of a public-facing customer portal where customers submit and vote on ideas. Missing for 10: any documentation of a public portal, idea submission form, or voting mechanism for customers.

   ⚑ flag

Convert to work

1. ai-native userHave AI turn a plain-language feedback description into a structured set of proposed tasks

   weight 3 · round drawn
   Canny~partialclaimed4/10

   Canny's MCP lets an LLM query and act on feedback, including drafting PRDs and organizing/linking requests, which implies AI could turn feedback into structured tasks, but there's no explicit feature or workflow described that ingests a plain-language feedback description and outputs a structured task list. Missing for 10: a dedicated 'feedback-to-task' generation feature, hands-on example of the described workflow, and evidence of structured task output format.

   • [claimed-docs] “Use Canny as the data layer for your AI workflows. Draft PRDs from real requests and build the features with Claude, generate reports for le…”
   • [claimed-docs] “Move ideas through your workflow without leaving your LLM. Update statuses, link deals, organize requests, and reply to customers.”
   • [claimed-docs] “Connect Canny to ChatGPT, Claude, and Cursor. Ask questions about customer feedback, take action on it, and ground your AI workflows in real…”
   • [claimed-docs] “Ask your LLM what customers asked for and understand the context behind each request. Dig into top themes, ARR impact across ideas, and quer…”

   proof ↗⚑ flag

   Productboard~partialclaimed4/10

   Productboard's auto-insight detection links feedback to feature ideas (docs-16) and its MCP server lets AI agents create features/initiatives from context (docs-10–13), which is adjacent to turning feedback into structured work items, but there is no explicit documentation of AI parsing a plain-language feedback description into a distinct, structured set of proposed tasks. missing for 10: explicit feature/doc showing AI decomposing a single feedback item into multiple structured task proposals, evidence of task granularity (subtasks, acceptance criteria), independent/hands-on confirmation of this specific workflow.

   • [claimed-docs] “Productboard automatically detects insights within feedback that can be linked to related feature ideas.”
   • [claimed-docs] “Now you can use any agent to add new features, reschedule an initiative, or start building off a product specification.”
   • [claimed-docs] “Update custom drivers and other data fields on features and initiatives (like value, risk, impact, confidence, and effort) to automate prior…”
   • [claimed-docs] “Access your latest product specifications from coding agents and AI prototyping tools to initiate the product delivery process.”

   proof ↗⚑ flag

2. product-managerReview and approve the scope of a task before any work begins on it

   weight 2 · round to Productboard
   Canny—none0/10

   Canny's evidence covers feedback collection, roadmap status workflows, and syncing to task trackers (Jira, Linear, GitHub, etc.), but there is no documented feature for a PM to formally review/approve a task's scope before work starts — Canny hands off to the external tracker once linked, and no approval gate is described.

   ⚑ flag

   Productboard~partialclaimed5/10

   Productboard's docs mention gating delivery on 'fully-validated features' and using specs to 'initiate the product delivery process', implying a review step before work begins, but there's no explicit evidence of a formal approval/sign-off workflow, status gate, or scope-lock mechanism a PM would use to approve a task. missing for 10: explicit approval/status workflow documentation, evidence of a sign-off step or scope lock before handoff to engineering, independent corroboration of this gating in practice.

   • [claimed-docs] “Ensure developers are working on what really matters by sharing customer context and only sending fully-validated features into delivery.”
   • [claimed-docs] “Access your latest product specifications from coding agents and AI prototyping tools to initiate the product delivery process.”
   • [claimed-docs] “Access one searchable repository for all product requests and feedback from colleagues and customers.”

   proof ↗⚑ flag

3. product-managerTurn a customer feedback submission directly into an actionable task or project intention

   weight 3 · round drawn
   Canny✓fullclaimed8/10

   Canny lets PMs directly convert feedback into actionable work: integrations sync feedback to Jira, Linear, GitHub, ClickUp, and Asana as tasks/issues with status sync, and the MCP server allows updating statuses, linking deals, organizing requests, and even drafting PRDs from real requests directly from an LLM workflow. This covers the feedback-to-task pipeline end-to-end with both native integrations and AI-assisted workflows. Missing for 10: independent/hands-on corroboration beyond vendor docs, and more detail on native in-app task/project creation without third-party tool integration.

   • [claimed-docs] “Send feedback to Jira, link it to existing issues, and sync statuses. Only for Jira Cloud.”
   • [claimed-docs] “Send feedback to Linear, link it to existing issues, and sync statuses.”
   • [claimed-docs] “Send feedback to GitHub, link it to existing issues, and sync statuses.”
   • [claimed-docs] “Send feedback to ClickUp, link it to existing tasks, and sync statuses.”
   • [claimed-docs] “Send feedback to Asana, link it to existing tasks, and sync statuses.”
   • [claimed-docs] “Move ideas through your workflow without leaving your LLM. Update statuses, link deals, organize requests, and reply to customers.”
   • [claimed-docs] “Use Canny as the data layer for your AI workflows. Draft PRDs from real requests and build the features with Claude, generate reports for le…”

   proof ↗⚑ flag

   Productboard✓fullclaimed8/10

   Productboard has a core feedback repository that automatically detects insights linkable to feature ideas, notes can be linked to features to quantify demand, and features can be pushed into delivery/tracked—directly supporting turning feedback into actionable tasks/initiatives. This is reinforced by API/MCP capabilities to create feedback, link it to features/initiatives, and push prioritized features into delivery. Missing for 10: independent/hands-on verification of the feedback-to-feature workflow UI and any case-study evidence of PM usage in practice.

   • [claimed-docs] “Productboard automatically detects insights within feedback that can be linked to related feature ideas.”
   • [claimed-docs] “Notes can be linked to features to quantify demand and inform prioritization.”
   • [claimed-docs] “Ensure developers are working on what really matters by sharing customer context and only sending fully-validated features into delivery.”
   • [claimed-docs] “Push prioritized features into delivery, track their status, and sync key data.”
   • [claimed-docs] “Access one searchable repository for all product requests and feedback from colleagues and customers.”
   • [claimed-docs] “Associate feedback with users and companies, and connect it to features, components, and other product hierarchy entities.”
   • [claimed-docs] “Create, read, update, delete, and search customer feedback, complete with text content, tags, owners, source metadata, and customer attribut…”

   proof ↗⚑ flag

4. developerPull an approved product specification into the delivery workflow to start implementation

   weight 2 · round to Productboard
   Canny~partialclaimed6/10

   Canny lets teams sync feedback/roadmap items into delivery tools like Jira, Linear, GitHub, ClickUp and Asana (linking issues, syncing statuses), and its MCP integration explicitly supports drafting PRDs from real requests and building features with Claude, which maps to pulling an approved spec into implementation. However, this relies on third-party issue trackers rather than a native 'approved spec' handoff step, and there's no dedicated documentation of an approval gate or spec format being pulled into a dev workflow. missing for 10: a native concept of an 'approved' spec/status gate, first-party workflow docs showing the full approve-to-implement handoff, and independent/hands-on confirmation beyond vendor pages.

   • [claimed-docs] “Use Canny as the data layer for your AI workflows. Draft PRDs from real requests and build the features with Claude, generate reports for le…”
   • [claimed-docs] “Send feedback to Jira, link it to existing issues, and sync statuses. Only for Jira Cloud.”
   • [claimed-docs] “Send feedback to Linear, link it to existing issues, and sync statuses.”
   • [claimed-docs] “Send feedback to GitHub, link it to existing issues, and sync statuses.”
   • [claimed-docs] “Send feedback to ClickUp, link it to existing tasks, and sync statuses.”
   • [claimed-docs] “Send feedback to Asana, link it to existing tasks, and sync statuses.”
   • [claimed-docs] “Move ideas through your workflow without leaving your LLM. Update statuses, link deals, organize requests, and reply to customers.”

   proof ↗⚑ flag

   Productboard✓fullclaimed8/10

   Productboard's MCP server explicitly lets coding agents and AI prototyping tools pull approved specs directly into the delivery workflow ('Access your latest product specifications from coding agents... to initiate the product delivery process', 'Pull a spec, talk through the plan with your agent, and build'), plus GitHub/Jira push-to-dev integrations and a REST API for programmatic access. Missing for 10: independent/hands-on developer confirmation that the workflow works smoothly in practice beyond vendor docs.

   • [claimed-docs] “Now you can use any agent to add new features, reschedule an initiative, or start building off a product specification.”
   • [claimed-docs] “Access your latest product specifications from coding agents and AI prototyping tools to initiate the product delivery process.”
   • [claimed-docs] “The Productboard MCP server lets your coding agent read product specs straight from Productboard — and write back to them.”
   • [claimed-docs] “Pull a spec, talk through the plan with your agent, and build. No setup beyond the connection.”
   • [claimed-docs] “Push prioritized features into delivery, track their status, and sync key data.”
   • [claimed-docs] “Push features ready for development into GitHub as issues.”
   • [claimed-docs] “Ensure developers are working on what really matters by sharing customer context and only sending fully-validated features into delivery.”

   proof ↗⚑ flag

5. product-managerWatch the real-time status of each in-progress task, including what it's blocked on and its last update

   weight 2 · round to Canny
   Canny~partialclaimed3/10

   Canny syncs statuses with tools like Jira, Linear, GitHub, ClickUp, and Asana, so a PM could see a high-level status label update via those integrations, but there is no evidence of a dedicated real-time view showing why a task is blocked or a detailed 'last update' timeline. missing for 10: explicit blocked-reason field, real-time update timestamps/activity feed, dedicated in-progress task dashboard.

   • [claimed-docs] “Send feedback to Jira, link it to existing issues, and sync statuses. Only for Jira Cloud.”
   • [claimed-docs] “Send feedback to Linear, link it to existing issues, and sync statuses.”
   • [claimed-docs] “Send feedback to GitHub, link it to existing issues, and sync statuses.”
   • [claimed-docs] “Send feedback to ClickUp, link it to existing tasks, and sync statuses.”
   • [claimed-docs] “Send feedback to Asana, link it to existing tasks, and sync statuses.”

   proof ↗⚑ flag

   Productboard—none0/10

   Productboard's evidence covers pushing prioritized features into delivery tools and syncing status/key data (e.g., docs-20, docs-23), but there is no evidence of a real-time view showing what a specific in-progress task is blocked on or its last update timestamp — that granularity lives in delivery tools like Jira that Productboard integrates with, not in Productboard itself.

   • [claimed-docs] “Push prioritized features into delivery, track their status, and sync key data.”
   • [claimed-docs] “Close the loop between what you specified, what you shipped, and whether it actually worked.”
   • [claimed-docs] “Productboard Integrations | Jira, Azure DevOps, Zendesk & More”
   • [claimed-docs] “Push features ready for development into GitHub as issues.”

   proof ↗⚑ flag

Tool connections

1. product-managerConnect feedback capture to support, sales, and chat tools I already use

   weight 3 · round to Canny
   Canny✓fullclaimed8/10

   Canny documents native integrations explicitly aimed at support, sales, and chat sources (Gong, Intercom, Slack, Fathom, Fireflies) via Autopilot, plus a broader integrations directory and MCP access from chat-based AI tools, directly matching the PM's need to connect feedback capture to existing tools. Missing for 10: independent/hands-on verification of these integrations working in practice and more detail on setup/configuration beyond marketing copy.

   • [claimed-docs] “Automatically capture feedback from Gong, Intercom, Slack, and dozens of other sources, or directly from customers on your Portal.”
   • [claimed-docs] “Autopilot captures feedback from support, sales, and review platforms automatically. Native integrations keep the rest of your stack in sync…”
   • [claimed-docs] “Autopilot pulls feature requests out of your Fathom call transcripts automatically.”
   • [claimed-docs] “Capture product feedback from your Fireflies sales call transcripts.”
   • [claimed-docs] “Get instant notifications in Discord about your feedback.”
   • [claimed-docs] “Integrations | Canny”

   proof ↗⚑ flag

   Productboard~partialclaimed5/10

   Productboard's API and webhook capabilities (docs-1, docs-2, docs-5) let feedback be routed in from CRMs and other external systems, and custom integrations can be built into the UI, which technically enables connecting support/sales/chat data sources. However, the evidence pack contains no mention of specific pre-built native connectors for common support, sales, or chat tools (e.g., named integrations), only generic REST/webhook/custom-integration capability. Missing for 10: named native integrations with support/sales/chat platforms, evidence of a marketplace/app directory, and independent confirmation of ease-of-setup for non-technical PMs.

   • [claimed-docs] “Create, search, and manage feedback, documents, customers, and product data through Productboard's REST API.”
   • [claimed-docs] “Developers and technical teams use it to route feedback from proprietary CRMs, internal tools, and other non-native sources into Productboar…”
   • [claimed-docs] “The Webhooks API allows you to subscribe to events in Productboard and receive real-time notifications when data changes.”
   • [claimed-docs] “Use this to build custom integrations that appear and operate inside the Productboard UI.”

   proof ↗⚑ flag

2. developerBuild custom integrations that appear and operate directly inside the product's own interface

   weight 1 · round to Productboard
   Canny—none0/10

   Canny's developer docs cover embedding Canny's own widget into a customer's app, its API/SDK, and native integrations that push feedback out to third-party tools (Jira, Linear, GitHub, etc.) or connect via MCP to external AI tools — but there is no evidence of a platform, app marketplace, or extension framework letting developers build custom integrations that run/appear directly inside Canny's own interface (dashboard, portal, or admin UI).

   ⚑ flag

   Productboard~partialclaimed6/10

   First-party docs explicitly state you can 'build custom integrations that appear and operate inside the Productboard UI' (docs-7) and describe an API/webhooks platform for extending Productboard (intdir-10, docs-5, docs-31), which supports the core claim. However, there's no detail on an actual UI-extension SDK, embed points, or hands-on/independent confirmation of what such an in-UI integration looks like in practice. Missing for 10: concrete UI-extension developer docs/SDK, third-party example integrations that embed in the Productboard UI, and independent corroboration beyond vendor marketing copy.

   • [claimed-docs] “Use this to build custom integrations that appear and operate inside the Productboard UI.”
   • [claimed-docs] “Extend Productboard's capabilities, connect it with other tools, update product data programmatically, develop your own integrations, and pr…”
   • [claimed-docs] “The Webhooks API allows you to subscribe to events in Productboard and receive real-time notifications when data changes.”
   • [claimed-docs] “API v2 introduces a new concept: configuration endpoints that describe the shape of data in your workspace.”
   • [claimed-docs] “You can retrieve the ID of any entity from within Productboard without having to call an API endpoint.”

   proof ↗⚑ flag

3. product-managerConnect common third-party tools to my feedback workspace in just a few clicks

   weight 2 · round to Canny
   Canny✓fullclaimed8/10

   Canny documents a native integrations directory covering Jira, Linear, GitHub, ClickUp, Asana, Slack, Discord, Intercom, Gong, Fathom, Fireflies and others, described as easy native connections that sync statuses and capture feedback automatically, plus a low-friction MCP connection to ChatGPT/Claude/Cursor with 'no API keys, no engineering.' Missing for 10: independent hands-on confirmation of the actual click-count/setup friction and no evidence of a broader app marketplace beyond the listed integrations.

   • [claimed-docs] “Automatically capture feedback from Gong, Intercom, Slack, and dozens of other sources, or directly from customers on your Portal.”
   • [claimed-docs] “Autopilot captures feedback from support, sales, and review platforms automatically. Native integrations keep the rest of your stack in sync…”
   • [claimed-docs] “Get instant notifications in Discord about your feedback.”
   • [claimed-docs] “Autopilot pulls feature requests out of your Fathom call transcripts automatically.”
   • [claimed-docs] “Capture product feedback from your Fireflies sales call transcripts.”
   • [claimed-docs] “Send feedback to Jira, link it to existing issues, and sync statuses. Only for Jira Cloud.”
   • [claimed-docs] “Send feedback to Linear, link it to existing issues, and sync statuses.”
   • [claimed-docs] “Send feedback to GitHub, link it to existing issues, and sync statuses.”
   • [claimed-docs] “Send feedback to ClickUp, link it to existing tasks, and sync statuses.”
   • [claimed-docs] “Send feedback to Asana, link it to existing tasks, and sync statuses.”
   • [claimed-docs] “Add Canny's MCP server to ChatGPT, Claude, or Cursor with a secure sign-in. No API keys, no engineering.”
   • [claimed-docs] “Integrations | Canny”

   proof ↗⚑ flag

   Productboard~partialprobed5/10

   Evidence focuses on API/webhooks/MCP server for building custom integrations, which requires developer effort rather than a few-click connection for a PM. There's no mention of a pre-built integrations marketplace or one-click connectors (e.g., Slack, Jira, Salesforce) that a non-technical PM could enable directly. Missing for 10: evidence of a native integrations directory/marketplace with one-click setup, no-code connector examples, and PM-facing (not developer-facing) setup flow.

   • [claimed-docs] “Create, search, and manage feedback, documents, customers, and product data through Productboard's REST API.”
   • [claimed-docs] “Developers and technical teams use it to route feedback from proprietary CRMs, internal tools, and other non-native sources into Productboar…”
   • [claimed-docs] “Use this to build custom integrations that appear and operate inside the Productboard UI.”
   • [probe] “official MCP server documented at https://www.productboard.com/integrations/mcp-server/”

   proof ↗⚑ flag

Navigation

1. developerJump directly to any feedback item, task, or initiative by typing its short name

   weight 1 · round drawn
   Canny—none0/10

   No evidence describes a search/navigation feature for jumping directly to a feedback item, task, or initiative by typing its short name (e.g., a quick-open/command-palette shortcut). The evidence pack covers MCP integrations, widget install, changelog, and third-party sync tools, none of which address this navigation UX.

   ⚑ flag

   Productboard—none0/10

   No evidence describes a quick-open or 'jump to' navigation feature where a developer types a short name/ID to instantly open a feedback item, task, or initiative. Evidence only covers retrieving entity IDs via the API/UI (docs-46) and general search repositories (docs-15), which is a different capability.

   ⚑ flag

Setup

1. product-managerUse a rich text editor with custom components to format changelog and help content

   weight 1 · round drawn
   Canny—none0/10

   Evidence covers Canny's changelog feature, MCP integration, feedback capture and integrations, but nothing describes a rich text editor with custom components for authoring changelog/help content. Missing for 10: any mention of an editor, rich text formatting capabilities, custom components/blocks, or help-content authoring tools.

   ⚑ flag

   Productboard—none0/10

   The evidence pack mentions 'editor documents' and document management via the API, but nothing describes a rich text editor with custom components, nor any changelog or help-content authoring feature for product managers. Missing for 10: any documentation of a WYSIWYG/rich text editor, custom embeddable components, or changelog/help-center publishing capability.

   ⚑ flag

2. product-managerSet up a public-facing feedback portal without engineering help

   weight 2 · round to Canny
   Canny~partialclaimed6/10

   Canny's core product is a hosted feedback portal, and docs describe simply linking to 'your Canny feedback site' (canny-docs-1) plus a changelog tool (canny-docs-8/14) that require no code to use. However, deeper setup steps described (SDK install, widget embedding, SSO, custom fields) are aimed at developers, and there's no explicit documentation walking a non-technical PM through portal creation without engineering. Missing for 10: explicit no-code onboarding wizard documentation, independent PM testimonials confirming self-serve setup without engineering.

   • [claimed-docs] “In your application, add a link to your Canny feedback site. Put the link somewhere it's easy to find.”
   • [claimed-docs] “Announce what you shipped, show customers you heard them, and turn closed feedback into retention and upgrades.”
   • [claimed-docs] “Product Changelog Software and Release Notes Tool”
   • [claimed-docs] “Embed a Canny widget into your website”

   proof ↗⚑ flag

   Productboard—none0/10

   The evidence pack covers APIs, webhooks, MCP server, and integrations, but contains no mention of a self-service public feedback portal feature that a PM could configure without engineering help.

   ⚑ flag

1. ai-native userDo everything through the API that I can do in the UI

   weight 2 · round to Productboard
   Canny~partialprobed4/10

   Canny publishes a public API reference and an MCP server that lets AI tools query and act on feedback data (statuses, replies, linking), suggesting meaningful API coverage, but no evidence establishes full parity between the API/MCP surface and everything available in the Canny UI (e.g., admin settings, roadmap customization, changelog publishing, integrations config). Missing for 10: explicit documentation or independent confirmation of complete UI-API parity, and an OpenAPI spec (probe found only 404s) that would demonstrate full endpoint coverage.

   • [claimed-docs] “Canny API Reference”
   • [claimed-docs] “Connect Canny to ChatGPT, Claude, and Cursor. Ask questions about customer feedback, take action on it, and ground your AI workflows in real…”
   • [claimed-docs] “Move ideas through your workflow without leaving your LLM. Update statuses, link deals, organize requests, and reply to customers.”
   • [probe] “PROBE openapi: all candidate paths 404 (https://developers.canny.io/openapi.json, https://developers.canny.io/swagger.json, https://develope…”
   • [probe] “official MCP server documented at https://canny.io/features/mcp”

   proof ↗⚑ flag

   Productboard~partialprobed7/10

   Productboard's REST API (v2) and MCP server cover most core objects—features, initiatives, releases, objectives, key results, feedback, documents, teams, companies, users, webhooks, and custom fields—letting AI agents create/update/query nearly all major workspace entities also editable in the UI. However, there's no discoverable OpenAPI/swagger spec (probe returned 404s), no vendor claim of 100% UI-API parity, and some UI-only workflows (e.g., insights auto-detection, Spark synthesis agent, deep configuration/admin settings) aren't clearly exposed via API. Missing for 10: a published complete OpenAPI schema, explicit parity statement, and independent confirmation that all UI actions (not just major entities) are API-accessible.

   • [claimed-docs] “Create, search, and manage feedback, documents, customers, and product data through Productboard's REST API.”
   • [claimed-docs] “Associate feedback with users and companies, and connect it to features, components, and other product hierarchy entities.”
   • [claimed-docs] “You can create and update teams, retrieve team details, and manage team membership by adding or removing members.”
   • [claimed-docs] “The Teams API allows you to programmatically manage teams in your Productboard workspace.”
   • [claimed-docs] “Create and update features, subfeatures, initiatives, releases, objectives, key results, companies, users, feedback, documents, and more dir…”
   • [claimed-docs] “Create and manage editor documents and product document boards, with support for bulk operations, document templates, and thread and comment…”
   • [claimed-docs] “The Webhooks API allows you to subscribe to events in Productboard and receive real-time notifications when data changes.”
   • [claimed-docs] “Sync objectives between Productboard and your OKR tool (e.g., Lattice, Ally.io, Gtmhub) to keep strategy aligned.”
   • [claimed-docs] “If you have multiple teams working out of the same workspace, each using a different release system, you can manage each as a separate relea…”
   • [probe] “PROBE openapi: all candidate paths 404 (https://developer.productboard.com/openapi.json, https://developer.productboard.com/swagger.json, ht…”
   • [claimed-docs] “The Productboard MCP server lets your coding agent read product specs straight from Productboard — and write back to them.”

   proof ↗⚑ flag

2. ai-native userExport all of my data in open formats and leave

   weight 3 · round to Productboard
   Canny—none0/10

   While Canny provides an API reference and developer docs, there is no evidence of a bulk data export feature, open-format export (CSV/JSON dump of feedback, users, etc.), or any documented data-portability/account-closure workflow. The axis applies to a SaaS feedback platform, but no capability is evidenced.

   ⚑ flag

   Productboard~partialclaimed5/10

   Productboard's REST API v2 lets users programmatically read/list virtually all entities (feedback, features, documents, customers, teams) with cursor-based pagination, which effectively enables a full data export via JSON over HTTP — an open, non-proprietary format and mechanism to leave with your data intact. However, there is no documented one-click 'export all my data' feature, no explicit mention of bulk/GDPR-style data export or account-closure export guarantee, and no confirmation of formats beyond implied JSON via REST responses. Missing for 10: a dedicated bulk/full-account export tool, explicit open-format guarantees (e.g., CSV/JSON export docs), and evidence of a data-portability/account-deletion export flow.

   • [claimed-docs] “Create, search, and manage feedback, documents, customers, and product data through Productboard's REST API.”
   • [claimed-docs] “For large result sets, both the list and search endpoints support cursor-based pagination.”
   • [claimed-docs] “Create, read, update, delete, and search customer feedback, complete with text content, tags, owners, source metadata, and customer attribut…”
   • [claimed-docs] “List and search feedback with structured filters by owner, creator, tags, date range, and status. For large result sets, both the list and s…”
   • [claimed-docs] “Create and manage editor documents and product document boards, with support for bulk operations, document templates, and thread and comment…”
   • [claimed-docs] “The Productboard REST API uses cursor-based pagination for all endpoints that return a list of results.”

   proof ↗⚑ flag

3. ai-native userSelf-host the core product

   weight 3 · round drawn
   Canny—none0/10

   Canny is a hosted SaaS feedback platform with no evidence of a self-hostable/open-source core product; all docs describe cloud install/SDK/widget/MCP integrations, not self-hosting options.

   ⚑ flag

   Productboard—none0/10

   Productboard is a SaaS-only cloud product; no evidence anywhere in the pack mentions self-hosting, on-prem deployment, or a self-managed core product. Absence of evidence for an applicable capability means 'none'.

   ⚑ flag

Insights surfacing

1. ai-native userRely on AI to automatically detect insights within feedback and link them to related feature ideas

   weight 2 · round to Productboard
   Canny~partialclaimed5/10

   Canny's Autopilot automatically captures feedback from calls, support, and sales tools, and its MCP server lets an LLM be asked about 'top themes' and context behind requests, suggesting some AI-driven insight surfacing (canny-docs-4, canny-docs-11, canny-intdir-2/3). However, there is no evidence of the AI proactively detecting insights and automatically linking them to existing/related feature ideas within Canny itself — the linking features shown (Jira, Linear, GitHub, ClickUp, Asana) are manual issue-tracker syncs, not AI-detected relations between feedback and ideas. Missing for 10: evidence of automatic duplicate/related-idea detection, proof of unsupervised insight-to-idea linking (vs. on-demand LLM queries), and independent validation of this capability.

   • [claimed-docs] “Ask your LLM what customers asked for and understand the context behind each request. Dig into top themes, ARR impact across ideas, and quer…”
   • [claimed-docs] “Autopilot captures feedback from support, sales, and review platforms automatically. Native integrations keep the rest of your stack in sync…”
   • [claimed-docs] “Autopilot pulls feature requests out of your Fathom call transcripts automatically.”
   • [claimed-docs] “Capture product feedback from your Fireflies sales call transcripts.”
   • [claimed-docs] “Use Canny as the data layer for your AI workflows. Draft PRDs from real requests and build the features with Claude, generate reports for le…”

   proof ↗⚑ flag

   Productboard~partialclaimed6/10

   Productboard's marketing docs explicitly claim automatic insight detection linked to feature ideas, with citations for findings, but this is described only at a high marketing level with no technical documentation, independent validation, or hands-on evidence of accuracy/scope. Missing for 10: technical/API documentation of the insight-detection mechanism, independent or hands-on verification of accuracy, and detail on how linking to feature ideas actually works in practice.

   • [claimed-docs] “Productboard automatically detects insights within feedback that can be linked to related feature ideas.”
   • [claimed-docs] “Understand critical feedback trends and act fast on new opportunities. Productboard includes citations for every finding, so decisions are a…”
   • [claimed-docs] “Go from new idea to delivery-ready in a fraction of the time. Productboard's built-in agent handles all the information synthesis”

   proof ↗⚑ flag

Scoring

1. product-managerLink feedback requests to roadmap features, initiatives, and releases

   weight 3 · round to Productboard
   Canny~partialclaimed4/10

   Evidence shows Canny's MCP layer lets teams 'query your team's roadmap,' 'update statuses, link deals, organize requests,' and the changelog feature turns closed feedback into shipped announcements, implying some link between feedback and roadmap/release stages, but there is no explicit documentation of a native roadmap board where feedback items are directly attached to specific initiatives or releases. Missing for 10: dedicated docs on roadmap board structure, initiative/epic linking, and release-tagging of feedback items beyond generic issue-tracker sync (Jira/Linear/GitHub).

   • [claimed-docs] “Ask your LLM what customers asked for and understand the context behind each request. Dig into top themes, ARR impact across ideas, and quer…”
   • [claimed-docs] “Move ideas through your workflow without leaving your LLM. Update statuses, link deals, organize requests, and reply to customers.”
   • [claimed-docs] “Announce what you shipped, show customers you heard them, and turn closed feedback into retention and upgrades.”
   • [claimed-docs] “Send feedback to Jira, link it to existing issues, and sync statuses. Only for Jira Cloud.”
   • [claimed-docs] “Send feedback to Linear, link it to existing issues, and sync statuses.”

   proof ↗⚑ flag

   Productboard✓fullclaimed8/10

   Productboard's core product model explicitly links feedback/notes to features, and the API/MCP docs confirm associating feedback with features, adding features to initiatives, and moving initiatives to releases. This is a core, well-documented native capability, not just a workaround via API. missing for 10: independent/hands-on user corroboration of the linking workflow in the UI (all evidence is vendor documentation).

   • [claimed-docs] “Associate feedback with users and companies, and connect it to features, components, and other product hierarchy entities.”
   • [claimed-docs] “Add features to initiatives, move an initiative to a new release, or create key results beneath an objective.”
   • [claimed-docs] “Notes can be linked to features to quantify demand and inform prioritization.”
   • [claimed-docs] “Associate feedback with users and companies, and connect it to features, components, and other product hierarchy entities. Dedicated endpoin…”
   • [claimed-docs] “Productboard automatically detects insights within feedback that can be linked to related feature ideas.”
   • [claimed-docs] “Create and update features, subfeatures, initiatives, releases, objectives, key results, companies, users, feedback, documents, and more dir…”

   proof ↗⚑ flag

2. product-managerScore feedback and ideas using fields like value, risk, impact, confidence, and effort

   weight 3 · round to Productboard
   Canny~partialclaimed4/10

   Canny's Custom Fields feature (canny-docs-20) would let a PM add fields such as value, risk, impact, confidence, or effort to ideas, but the evidence pack contains no documentation of a purpose-built scoring/prioritization framework (e.g., RICE/ICE formulas, weighted score calculation, or a dedicated prioritization view). Missing for 10: evidence of built-in scoring formulas or automated priority-score calculation, evidence of a dedicated prioritization board/view using these fields.

   • [claimed-docs] “Custom Fields | Canny”

   proof ↗⚑ flag

   Productboard✓fullclaimed8/10

   Productboard's docs explicitly reference custom fields/drivers such as value, risk, impact, confidence, and effort on features/initiatives, and both the product data model and MCP server allow updating these to drive prioritization workflows. Missing for 10: no independent/hands-on corroboration of a dedicated scoring UI or formulaic scoring model, and no direct screenshot/first-party doc showing all five named fields together in a native scorecard.

   • [claimed-docs] “Update custom drivers and other data fields on features and initiatives (like value, risk, impact, confidence, and effort) to automate prior…”
   • [claimed-docs] “Associate feedback with users and companies, and connect it to features, components, and other product hierarchy entities.”
   • [claimed-docs] “Ensure developers are working on what really matters by sharing customer context and only sending fully-validated features into delivery.”

   proof ↗⚑ flag

3. product-managerOrganize requests into a hierarchy of features, components, and initiatives

   weight 2 · round to Productboard
   Canny—none0/10

   The evidence pack contains no mention of hierarchical organization of feedback into features, components, or initiatives (e.g., no reference to Canny's category/board structure or initiative grouping) — only integrations, MCP, changelog, and SDK install docs are present. Missing for 10: any docs on categories, tagging hierarchy, or initiative/epic grouping features.

   ⚑ flag

   Productboard✓fullclaimed8/10

   Productboard's core product model natively supports organizing feedback/requests into features, subfeatures, components, initiatives, releases, objectives and key results, with API and MCP evidence confirming these hierarchy entities and the ability to link feedback and features to them (e.g., docs-4, docs-19, docs-29, docs-12, docs-47). This directly matches the story of organizing requests into a feature/component/initiative hierarchy. Missing for 10: independent/hands-on corroboration beyond vendor docs and a clearer description of the component-level UI hierarchy itself (most detail is API/MCP-centric rather than native UI workflow).

   • [claimed-docs] “Associate feedback with users and companies, and connect it to features, components, and other product hierarchy entities.”
   • [claimed-docs] “Create and update features, subfeatures, initiatives, releases, objectives, key results, companies, users, feedback, documents, and more dir…”
   • [claimed-docs] “Associate feedback with users and companies, and connect it to features, components, and other product hierarchy entities. Dedicated endpoin…”
   • [claimed-docs] “Add features to initiatives, move an initiative to a new release, or create key results beneath an objective.”
   • [claimed-docs] “Notes can be linked to features to quantify demand and inform prioritization.”
   • [claimed-docs] “Access one searchable repository for all product requests and feedback from colleagues and customers.”

   proof ↗⚑ flag

4. product-managerReschedule initiatives and move them between releases as priorities change

   weight 2 · round to Productboard
   Canny—none0/10

   Evidence covers feedback capture, integrations, MCP/AI tooling, and changelog features but nothing about roadmap/release planning, initiative scheduling, or moving items between releases. Missing for 10: any mention of a roadmap tool, release/version grouping, or drag-and-drop scheduling of initiatives.

   ⚑ flag

   Productboard✓fullclaimed8/10

   Productboard's core roadmap model explicitly supports releases and initiatives, and both docs and MCP-server pages state you can 'reschedule an initiative' and 'move an initiative to a new release' directly, including via API/agent workflows. This is native product functionality (not just a third-party workaround), covering the exact story described. missing for 10: no independent/hands-on UI walkthrough or screenshot showing the drag-and-drop reschedule experience in the roadmap view, only docs/MCP references.

   • [claimed-docs] “Now you can use any agent to add new features, reschedule an initiative, or start building off a product specification.”
   • [claimed-docs] “Add features to initiatives, move an initiative to a new release, or create key results beneath an objective.”
   • [claimed-docs] “Create and update features, subfeatures, initiatives, releases, objectives, key results, companies, users, feedback, documents, and more dir…”
   • [claimed-docs] “If you have multiple teams working out of the same workspace, each using a different release system, you can manage each as a separate relea…”

   proof ↗⚑ flag

1. ai-native userChoose where my data is stored (region/residency)

   weight 2 · round drawn
   Canny—none0/10

   No evidence in the pack mentions data residency, regional storage options, or any control over where customer data is hosted; Canny's docs focus on install, SDKs, MCP, and integrations only.

   ⚑ flag

   Productboard—none0/10

   No evidence in the pack addresses data residency, region selection, or storage location controls; all evidence concerns API, MCP, and integration capabilities. Missing for 10: any mention of data residency options, EU/US hosting choice, or compliance documentation addressing storage location.

   ⚑ flag

2. ai-native userPrevent my data from being used to train AI models

   weight 3 · round drawn
   Canny—none0/10

   No evidence in the pack addresses AI training data opt-out, data usage policies for model training, or privacy controls preventing Canny from using customer data to train AI models; the pack only covers MCP feedback tooling and integrations. missing for 10: any privacy policy, ToS clause, or admin setting about excluding data from AI/ML training.

   ⚑ flag

   Productboard—none0/10

   No evidence in the pack addresses AI/model-training data usage, opt-out controls, or any privacy policy statement about training on customer data; all evidence concerns API/MCP/integration functionality, not data-training privacy posture.

   ⚑ flag

3. ai-native userControl data retention and deletion

   weight 2 · round drawn
   Canny—none0/10

   No evidence in the pack addresses data retention policies, deletion controls, or privacy/data-lifecycle management for user or feedback data; the docs cover SSO, install, widgets, MCP, and integrations but nothing about retention/deletion mechanisms.

   ⚑ flag

   Productboard—none0/10

   The evidence pack covers API/webhook/MCP capabilities extensively but contains no mention of data retention policies, deletion of stored customer/product data for privacy purposes, GDPR/right-to-be-forgotten controls, or data export/purge tooling. This axis is applicable to a SaaS platform storing customer feedback data, but no supporting evidence exists.

   ⚑ flag

4. ai-native userOpt out of telemetry and usage tracking

   weight 2 · round drawn
   Canny—none0/10

   No evidence pack items mention telemetry, usage tracking, analytics opt-out, or any privacy/data-collection controls for AI-native or MCP usage; the documentation focuses on feedback collection, integrations, and MCP feature capabilities rather than telemetry controls. missing for 10: any mention of telemetry/usage-data collection, an opt-out mechanism, or privacy settings related to AI tool usage.

   ⚑ flag

   Productboard—none0/10

   No evidence in the pack addresses telemetry opt-out or usage tracking controls for end users; the only related item is an Analytics API describing workspace usage data for admins, not a user-facing telemetry opt-out mechanism.

   ⚑ flag