DocsBot AI

@docsbot list all unanswered questions from our support bot over the last week and group them by topic

Here's a summary of 47 unanswered questions from your support bot (ID: bot_8x3k) over the past 7 days, grouped by topic: **Billing & Payments (18)**: Questions about invoice downloads, payment method updates, and subscription changes. **API Integration (15)**: Requests for webhook examples, authentication troubleshooting, and rate limit clarifications. **Feature Requests (9)**: Inquiries about custom domain support, white-labeling options, and multi-language capabilities. **Account Management (5)**: Password resets, team member permissions, and data export requests. This illustrative grouping shows how the bot's question log can surface knowledge gaps in your documentation.

This example demonstrates the List Questions tool combined with AI synthesis to identify documentation gaps. Requires valid bot ID and appropriate team permissions. The actual API returns raw question data; the AI performs the temporal filtering and topical grouping. Useful for content teams auditing support coverage.

@docsbot create a new bot for our developer documentation using the GitHub repo at github.com/ourcompany/api-docs

I've created a new bot named "Developer Docs Bot" (ID: bot_9m2p) within your team (team_4k7n). The bot is configured to index content from your GitHub repository at github.com/ourcompany/api-docs. Next steps: - The initial indexing will take 5-10 minutes depending on repository size - You can add additional sources (URLs, files, sitemaps) via the dashboard - Customize the bot's response style and behavior in the bot settings - Embed the bot on your docs site using the provided widget code This example shows the bot creation flow; actual setup requires valid team credentials and repository access permissions.

Showcases the Create Bot tool for provisioning new documentation assistants. Requires API key with team admin permissions and a valid team ID. The MCP handles bot creation but does not automatically trigger indexing—that's a separate background process. Consider rate limits if scripting multiple bot deployments.

@docsbot generate a support ticket from conversation ID conv_5x8a including the user's original question and the bot's suggested solution

I've generated a support ticket from conversation conv_5x8a: **Ticket Summary**: User unable to authenticate API requests after token refresh **Original Question**: "I'm getting 401 errors after refreshing my API token. The old token still works but the new one doesn't. Is there a propagation delay?" **Bot's Suggested Solution**: The bot recommended checking for whitespace in the copied token and verifying the Authorization header format, but the user confirmed the issue persists. **Recommended Action**: Escalate to engineering team to investigate token generation service. User ID: usr_3k9m, Account: Enterprise tier. This structured ticket is ready for your helpdesk system. The actual ticket format depends on your DocsBot configuration.

Demonstrates the Generate Conversation Ticket tool for escalating bot interactions to human support. Requires valid conversation ID and read access to chat logs. Particularly valuable for identifying when the bot reaches its knowledge boundaries. The ticket structure is configurable per team; this example shows a typical support handoff scenario.

When DocsBot wins for support teams under 10 people

A 6-person support team handling 200 tickets a week needs instant answers from product docs, help articles, and past resolved tickets. DocsBot is the right call here: you create one bot per product line, index your knowledge sources, and the team queries it in Switchy during live chats. The Generate Conversation Ticket tool closes the loop by turning chat threads into structured tickets without copy-paste. This works best when your knowledge base is under 5,000 pages and updates weekly, not hourly. If you're fielding 1,000+ tickets a week or your docs change multiple times daily, you'll hit rate limits and stale-answer friction. For small teams with stable docs, DocsBot cuts lookup time from 3 minutes to 15 seconds.

DocsBot for new-hire ramp in remote startups

A 12-person remote startup onboards 2-3 engineers a month and wants them self-sufficient by day three. You create a DocsBot indexed on your internal wiki, runbooks, and Slack archives, then share it in Switchy's onboarding channel. New hires ask setup questions in natural language; the bot surfaces the right doc section or past answer thread. The List Questions tool shows you which topics trip people up, so you can patch the wiki. This setup works when your onboarding corpus is under 2,000 pages and you're hiring fewer than 10 people a quarter. Beyond that scale, you need versioned docs and role-based bots, which DocsBot doesn't handle cleanly. For early-stage remote teams, it's the fastest way to stop repeating the same Slack answers.

When DocsBot fits pre-demo prep for small sales teams

A 4-person sales team preps for 15-20 demos a week and needs quick answers on feature availability, pricing tiers, and integration specs. You index your product docs and competitive battle cards into a DocsBot, then query it in Switchy while drafting proposals or prepping calls. The Get Bot Details and List Bot Sources tools let you audit what's indexed, so you catch outdated pricing before a prospect does. This works when your product has fewer than 50 SKUs and your docs update monthly. If you're in a high-velocity SaaS with weekly feature drops or complex enterprise pricing, DocsBot's refresh lag will burn you. For small B2B teams with stable product lines, it's faster than searching Notion or asking the product team on Slack.