Zendesk → Claude → Intercom: Knowledge Base Builder

Connect your Zendesk account and set up a scheduled query that pulls recently resolved tickets grouped by topic tags. Filter for tickets that were resolved with helpful answers and have high satisfaction ratings. Extract the original question, the resolution steps, and any internal notes from agents.

Organize the extracted ticket data in an Airtable base that groups similar issues by topic, tracks frequency counts, and flags high-impact question clusters that lack existing help articles. This prioritization step ensures Claude focuses on generating articles for the topics that will deflect the most tickets rather than writing about one-off edge cases.

Send clusters of related tickets to Claude with a prompt that synthesizes the common question pattern, writes a clear step-by-step solution, anticipates follow-up questions, and formats the article with proper headings, screenshot placeholders, and troubleshooting tips. Maintain a consistent tone matching your brand voice.

Push the generated articles to your Intercom Help Center as drafts for review. Assign them to the appropriate collection, add relevant tags, and set up a review workflow where a support lead approves before publishing. Track article performance to identify which generated content deflects the most tickets.

Post a weekly digest to the support team's Slack channel listing all newly published help articles, their source ticket clusters, and early deflection metrics. Include a feedback mechanism where agents can flag articles that need refinement, creating a continuous improvement loop that ensures the knowledge base stays accurate and useful.