CHATBOTS

Intercom Onboarding Concierge with Confidence-Based Human Handoff

An Intercom chatbot guides each new user through your setup checklist, and the moment its answer confidence drops below threshold it escalates the conversation to a live agent…

CategoryChatbots
Enginesim
Difficultyintermediate
Triggerevent
Steps6
Setup~15 min

How it runs

The automated pipeline, trigger to output.

  • TriggerNew onboarding conversation in IntercomIntercomIntercom
  • ActionLoad user's current setup step from PostgresPostgreSQLPostgres
  • ActionGenerate step reply + confidence score with OpenAIOpenAI
  • LogicBranch on confidence threshold
  • ActionReply and advance step, or assign to human with contextIntercomIntercom
  • OutputPersist updated onboarding statePostgreSQLPostgres

What it does

This concierge answers a new user's onboarding questions inside Intercom Messenger, walks them through your setup steps in order, and self-monitors its own certainty. When the model's confidence in an answer falls below your threshold, the conversation is handed to a human teammate instead of guessing.

When to use it

Use it when onboarding questions are mostly repetitive ("how do I connect my data source?") but a meaningful minority need a real person. It deflects the easy volume while protecting trust on the hard cases.

How it works

  1. 1A new inbound Intercom conversation from a user still in onboarding triggers the flow.
  2. 2The user's current setup stage and message are read from Postgres.
  3. 3OpenAI generates a step-specific reply plus a 0-1 confidence score.
  4. 4A logic branch checks the score against your threshold.
  5. 5If confident, the bot replies in Intercom and advances the stored step.
  6. 6If not confident, the conversation is assigned to a human with the AI's draft and reasoning attached as an internal note.

Set it up

What you configure once, before turning it on.

  1. 1
    Connect IntercomConversations, contacts, articles.
  2. 2
    Connect OpenAIModels, embeddings, files.
  3. 3
    Connect PostgresAny Postgres URL — query, write, migrate.
  4. 4
    Set each agent's modelWe leave models unset so you pick the tier — fast + cheap, or top-quality.
  5. 5
    Tune it to your dataEdit the prompts, filters, and field mappings so it matches how your team works.
  6. 6
    Test, then turn it onRun once against a sample, confirm the output, then enable the trigger.

Run this workflow in your colony.

14-day trial. No DevOps. No Sales call. Provisioned in under a minute.