Customer Success · Analyze & Recommend
Turn repeat support tickets into knowledge base articles automatically
Weekly, Doe reads all resolved tickets in Freshdesk, identifies recurring questions that don't have knowledge base articles, and drafts the missing articles in Notion. Also flags existing articles that are outdated based on how tickets were actually resolved.
Works acrossFreshdeskNotionSlack
What you get.
Recurring support questions identified, missing knowledge base articles drafted in Notion, and outdated articles flagged every week from resolved Freshdesk tickets. Doe turns your best agent responses into self-service documentation so customers find answers before they open tickets.
Your support team is answering the same question for the 50th time
A customer asks "how do I export my data?" An agent writes a thoughtful, detailed response, resolves the ticket, and moves on. Next week, a different customer asks the same thing. The next agent writes a slightly different version of the same answer. And again. And again. Everyone on the team knows a knowledge base article would prevent this, but writing documentation always loses to answering the next ticket in the queue. The backlog never shrinks, and the KB never grows.
The other side of the problem is articles that exist but are wrong. The product shipped a new export flow three months ago, but the KB article still describes the old one. Agents know the article is outdated and work around it, writing the correct steps from scratch every time. Customers who find the article on their own follow the wrong instructions, get confused, and open tickets, creating more work, not less. The documentation that was supposed to reduce volume is actively generating it.
What changes.
- 01KB article productionBefore · Written when someone has spare time (rarely)With Doe · Draft articles generated weekly from actual ticket resolutions
- 02Time to identify gapsBefore · Months. Gaps are felt, not measuredWith Doe · Flagged weekly with ticket volume and impact data
- 03Ticket deflectionBefore · Unmeasured. KB coverage grows slowly or not at allWith Doe · Quantified per topic with estimated tickets preventable
- 04Article freshnessBefore · Outdated articles discovered by frustrated customersWith Doe · Stale articles flagged when agent answers diverge from published content
How Doe finds knowledge gaps and drafts articles
- 01Reads all resolved tickets from the past weekFreshdeskDoe ingested 214 resolved tickets: questions, agent responses, resolution steps, tags, and customer segments
- 02Clusters tickets by topic and identifies missing KB coverageDoeDoe found 5 recurring topics with 8+ tickets each and no corresponding article: data export formatting, SSO setup for Okta, API rate limits, team permissions, and billing plan changes. 47 tickets this week alone
- 03Drafts the missing articles using the best agent responses as sourceNotion5 draft articles created with step-by-step instructions, common edge cases from ticket history, and links to related tickets for context
- 04Cross-references existing articles against how agents actually resolved ticketsDoeDoe flagged 3 outdated articles: the data export article references a deprecated UI, the integrations article is missing the new Slack connector, and the pricing page links to an old plan structure
- 05Delivers the weekly gap analysis to the support ops channelSlack5 new articles drafted with Notion links, 3 outdated articles flagged with specific discrepancies, and an estimate that 47 tickets could have been deflected
- 06RecurringEvery Monday at 8:00 AMEvery Monday, Doe reads the previous week's resolved tickets, clusters them by topic, identifies recurring questions without KB coverage, and drafts the missing articles in Notion. Your support team starts the week with new documentation ready for review and a clear picture of what is driving ticket volume. Gap analysis and draft articles posted to #support-ops in Slack.
Up and running in under ten minutes.
- 01Connect your toolsOne-click OAuth for each integration. No API keys, no engineering.
- 02Describe what you need“Read last week's resolved Freshdesk tickets, find any question that came up 5+ times without a matching KB article, and draft the missing articles in Notion using the best agent response as the starting point.”
- 03It runs on scheduleEvery Monday, the gap analysis and draft articles land in your support ops channel.
Before you delegate.
- 01Does it work with Zendesk or Intercom instead of Freshdesk?Yes. The same task connects to Zendesk, Intercom, or any ticketing system with an API. Freshdesk is shown here as the example, but Doe reads tickets from whichever platform your support team uses.
- 02Can it match our knowledge base writing style?Yes. Doe learns your documentation style from existing KB articles, including tone, formatting, heading structure, and level of technical detail. Draft articles follow your conventions so the review process is about accuracy, not rewriting.
- 03How does it determine what is "outdated"?Doe compares how agents actually resolved tickets against what the corresponding KB article says. When agents consistently provide different steps, different URLs, or different feature names than the published article, it flags the article as outdated and shows the specific discrepancies.
- 04Does it handle multiple product lines?Yes. Doe groups tickets by product line or feature area based on your tags and routing rules. Articles are drafted in the appropriate section of your knowledge base, and gap analysis is broken down by product so each team sees what matters to them.
- 05Can it suggest improvements for existing articles, or does it only flag them?Yes. When an existing article is flagged as outdated, Doe drafts updated sections using the correct resolution steps from recent tickets. Your documentation team reviews the suggested edits rather than researching the changes from scratch.