Demo Scenario

AI customer support with handover to people

The agent answers from the knowledge base, checks orders in the CRM, creates tickets and hands anything beyond its authority to an operator.

This is a demo scenario: how we would build the solution. Names and data are illustrative — it is not a specific client’s story.

01Task

Task

Operators answer the same questions again and again, while complex cases wait in the common queue.

Constraints

  • Refunds above the limit are decided by a person only.
  • Answers only from the current knowledge base, with a source link.
02Solution

Solution

  1. 01

    RAG over the knowledge base with article versions.

  2. 02

    Tools: order status, ticket creation, handover to an operator with context.

  3. 03

    Action limits and a log of every answer.

03Architecture

Architecture

Architecture: AI customer support with handover to people
  1. Customer channel → Agent runtime
  2. Agent runtime → Knowledge base · RAG, CRM · orders, Helpdesk, Operator
  3. Knowledge base · RAG
  4. CRM · orders
  5. Helpdesk
  6. Operator
Integrations
Knowledge base · CRM · Helpdesk · Telegram / web chat
Stack
LLM · RAG · Vector search · PostgreSQL · Node.js
04What it demonstrates

What it demonstrates

Demonstrates: answers from the knowledge base with sources, actions in systems within limits, and handover to a person with full context.