CHATBOTS
Onboarding coach with confidence-gated human handoff
Answers onboarding questions from docs, but when the retrieved docs don't confidently cover the question it escalates to a human in Slack instead of guessing.
How it runs
The automated pipeline, trigger to output.
- TriggerNew in-app message in IntercomIntercom
- ActionRetrieve relevant docs from CodaCoda
- ActionDraft answer and score doc coverageOpenAI
- LogicBranch on confidence score
- OutputAuto-reply in IntercomIntercom
- OutputEscalate weak answers to SlackSlack
What it does
This is the safe version of the doc-answering coach. It still drafts answers from your feature documentation, but it scores how well the docs actually cover the question first. High-confidence questions get an automatic reply; low-confidence ones are routed to your onboarding team in Slack with the question, the user, and the best partial match attached.
When to use it
Use it when wrong answers are worse than slow answers, for example during high-touch enterprise trials. It keeps the bot helpful on common questions while making sure edge cases reach a human before a bad reply erodes trust.
How it works
- 1A new in-app message arrives via the Intercom conversation trigger.
- 2The relevant feature docs are retrieved from Coda for the user's question.
- 3OpenAI evaluates coverage and drafts an answer plus a confidence score.
- 4A branch checks the score: confident answers are auto-sent, weak ones are flagged.
- 5Confident answers post back into Intercom; flagged ones go to the onboarding Slack channel with full context for a human to take over.
Set it up
What you configure once, before turning it on.
- 1Connect IntercomConversations, contacts, articles.
- 2Connect CodaDocs, packs, automations.
- 3Connect OpenAIModels, embeddings, files.
- 4Connect SlackChannels, DMs, threads, mentions.
- 5Set each agent's modelWe leave models unset so you pick the tier — fast + cheap, or top-quality.
- 6Tune it to your dataEdit the prompts, filters, and field mappings so it matches how your team works.
- 7Test, then turn it onRun once against a sample, confirm the output, then enable the trigger.
More Chatbots workflows
Discord WAF False-Positive Triage from Ray ID
When a user posts a Cloudflare ray ID in a Discord support channel, the bot looks up the blocked request's WAF event, classifies whether it looks like a genuine false positive.
Agentic WAF Investigator with Discord Approval Gate
An autonomous agent investigates a Discord-reported WAF block end to end: pulls Cloudflare logs, correlates with PagerDuty incidents, decides if an allowlist is safe.
Discord Onboarding Completion to CRM Sync
The moment a member finishes Discord onboarding, captures their submitted profile answers, creates or updates a contact in HubSpot.
Draft WAF Allowlist Proposal to Linear from Discord Report
Lets a support engineer confirm a Discord WAF report is a false positive, then auto-drafts a scoped Cloudflare allowlist rule proposal and files it as a Linear ticket…
Discord FAQ Free-Text Answer Grader
When a member submits a free-text answer in their onboarding thread, an agent grades it against the canonical FAQ knowledge base, replies with targeted feedback.
Discord FAQ-Gated Role Concierge
Greets each new Discord member in a private onboarding thread, walks them through required FAQ steps one at a time.
Run it inside a business
This workflow drops into a full company template. Import the org, and this is one of the playbooks its agents run.
Run this workflow in your colony.
14-day trial. No DevOps. No Sales call. Provisioned in under a minute.