Read AI is officially a Claude Connector. If you're one of Read AI's 5 million+ monthly active users and use Claude, your meeting intelligence, action items, key questions, and transcripts are now available directly inside Claude from the connector directory.
That alone is useful. But connectors are only raw material. What turns that raw material into outcomes are Skills: pre-built workflows that take your Read AI data, analyze it, and deliver something you'd actually act on. We built three of them, and each one runs end-to-end with a first-run setup wizard (meaning the first time you run it, it’ll configure the Skill alongside your guidance, then run automatically, unattended, and on a schedule).
Try them today, and go from install to working output in seconds.
A weekly product-feedback rollup. It pulls meetings from Read AI whose titles match patterns like "disco" or "demo," extracts product feedback into four categories (user needs, product quality, feature gaps, general feedback), optionally tags each item by persona and vertical, and tracks trends across an 8-week rolling window.
It publishes to three places: a local markdown snapshot, a wiki page (Confluence, Notion, SharePoint, or OneNote—just pick one), and a chat channel digest (Slack or Teams). The value is in trend tracking: this skill surfaces insights like "bidirectional Salesforce sync has been mentioned in 5 of the last 8 weeks" so the product team can see recurring signal versus noise.
A personal weekly planner. Every Friday morning it pulls your past week's meeting action items from Read AI plus DMs and @-mentions from Slack or Teams, uses LLM judgment to decide which items are actually yours to do, estimates durations in fixed buckets (15/30/60/90/120 min), groups them by topic across both sources, and books tentative and private calendar blocks on Google Calendar or Outlook for the upcoming workweek.
It enforces a daily cap (default 3 hours), never double-books, and has a clean undo command. Anything that doesn't fit gets listed in an "Overflow" section of the markdown digest.
A weekly coaching and deal-health report. It pulls all customer-facing calls from Read AI, identifies reps by internal domain (or CRM user list if HubSpot/Salesforce is connected), groups calls by account, then runs two LLM analysis passes: deal health per account (stakeholder breadth, sentiment trend, urgency, budget signals, etc.) and coaching per rep (talk-time ratio, question density, discovery quality, objection handling, etc. with specific transcript moments and "what to try instead" notes).
The output is a single self-contained HTML file with inline Chart.js visualizations, plus no network requests, opens fully offline, file mode 0600 for sensitivity.
All three skills are designed to run unattended on a schedule and are also safe to run interactively. Each has a first-run setup wizard, dry-run mode, and detailed run logging for rollback.
The Read AI MCP Claude Connector is available to all Read AI users.
Do you prefer ChatGPT or another MCP-compatible platform? Click here for more guidance.
%2C%20Color%3DOriginal.png)
