Charla

@charla list all knowledge base articles tagged with 'billing' and summarize which topics we're missing based on recent customer questions

Found 12 knowledge base articles tagged 'billing' in your Charla property. Topics covered include payment methods (4 articles), invoice generation (3), subscription changes (2), refund policies (2), and tax documentation (1). Cross-referencing against your last 30 conversations shows recurring questions about payment plan upgrades (8 mentions), international billing (5 mentions), and proration logic (4 mentions) — none of which have dedicated articles yet. Consider drafting content for these gaps to reduce support volume.

This example combines the list-articles tool with AI synthesis to identify documentation gaps. Requires read access to both knowledge base and conversation history. The MCP returns article metadata (titles, tags, IDs) but not full content unless you fetch individual articles, so the AI infers coverage from titles and summaries.

@charla create a new knowledge base article titled 'How to Update Your Payment Method' in the Billing category, mark it as published, and include step-by-step instructions

Created knowledge base article 'How to Update Your Payment Method' (ID: kb_9x4j2m, slug: how-to-update-your-payment-method) in the Billing category. Article is now published and visible to customers. Charla will automatically include this content in chatbot training within the next sync cycle, typically 15–30 minutes. You can edit the full content or adjust visibility settings directly in the Charla dashboard.

This write operation demonstrates article creation with category assignment and publish status. The MCP generates the URL slug automatically from the title. Be aware that publishing immediately makes the article customer-facing; default to draft status if you need internal review first. The AI can draft article content, but you must paste it into the article body separately or via a follow-up update call.

@charla pull the last 20 conversations from contacts tagged 'enterprise' and draft a summary of their most common feature requests

Retrieved 20 conversations from 14 enterprise-tagged contacts in your Charla property. Most common feature requests: SSO integration (mentioned in 9 conversations, 6 unique contacts), custom reporting dashboards (7 conversations, 5 contacts), API rate limit increases (5 conversations, 4 contacts), and white-label branding options (4 conversations, 3 contacts). Two contacts explicitly asked about SLA guarantees for enterprise plans. This data reflects conversations from the past 45 days; older threads may contain additional context.

This read-heavy example uses the list-conversations tool filtered by contact tags, then applies AI reasoning to extract themes. Charla returns full message history per conversation, so this can hit token limits with verbose threads. The MCP paginates results; you may need multiple calls to cover all enterprise contacts if your property has more than one page worth.

When Charla MCP speeds up context transfer between support reps

A 6-person support team runs two 8-hour shifts covering US and APAC time zones. The morning handoff is a 15-minute Slack thread where the overnight rep summarizes active conversations and flags anything urgent. With Charla MCP, the incoming rep asks Switchy to pull the last 12 hours of conversations, filter by unresolved status, and summarize each thread with the customer's last question. The MCP's List Conversations tool returns message history in one call, so the AI can build the handoff doc in under 30 seconds. This works cleanly if your Charla property has under 200 active conversations per shift; above that, pagination slows the query and you're better off using Charla's native filters. If your team already writes handoff notes in Charla itself, skip the MCP and just read those. But if handoff lives in Slack or a doc, Charla MCP turns conversation export into a one-prompt task.

Using Charla MCP to find missing help articles from real tickets

A 3-person product team at a B2B SaaS company ships a feature every two weeks and maintains a 40-article knowledge base in Charla. After each release, the support lead reviews the week's conversations to spot repeated questions that aren't covered in the docs. Manually, this means opening 20+ conversation threads and cross-checking titles against the KB index. With Charla MCP, the lead asks Switchy to list the last 50 conversations, extract the customer questions, then check which topics don't match any existing knowledge base article title or category. The MCP's List Knowledge Base Articles and List Conversations tools run in parallel, so the gap report takes under a minute. This scenario breaks down if your KB has more than 100 articles, because title-matching gets noisy and you need semantic search instead. If you're already tagging conversations with KB article IDs in Charla, the MCP won't add much. But for small teams doing monthly content audits, this turns a 2-hour manual task into a 5-minute prompt.

When Charla MCP pulls conversation history for user interviews

A 4-person design team at a fintech startup runs quarterly user interviews with 8-10 customers. The researcher spends an hour before each call reading the customer's support history to understand their pain points and feature requests. With Charla MCP, the researcher asks Switchy to pull all conversations for a specific contact UUID, summarize the top 3 issues they've raised, and flag any unresolved threads. The Get Contact and List Conversations tools return the full message history, so the AI can build a one-page brief with timestamps and direct quotes. This works well if each customer has 5-20 conversations on record; below that, there's not enough signal, and above 50 conversations the summary gets too generic. If your team already uses Charla's contact notes feature, you might not need the MCP. But if interview prep is scattered across Slack DMs and Google Docs, Charla MCP centralizes the pull in one workspace prompt.