AedeumBuilt For Life
Studio demoLive

Kestrel AI Assistant

Retail AI

Kestrel AI assistant for retail

Retail AI assistant for product questions, order status, and lead capture after hours, with clear boundaries and human handoff when needed.

Use case

Retail support

Coverage

24/7 on-site

Guardrails

Scoped knowledge base

Handoff

Human escalation path

Engagement

Studio demonstration · AI Integration tier

AI Assistant package reference

Deliverables

  • Product FAQ trained on catalogue data
  • Order status lookup integration pattern
  • Lead capture and enquiry routing
  • Tone and boundary configuration
  • Analytics on resolution vs escalation

Stack & systems

Claude APIRAG pipelineNext.js embedWebhook routing

Project narrative

Context

A retailer needed product answers and order lookups outside business hours without hiring overnight staff. Generic chatbots either hallucinated or felt off-brand within minutes.

Approach

We scoped knowledge tightly: catalogue, policies, and order-status patterns only. The assistant defers to humans on edge cases and logs what it could not resolve so the team improves coverage over time.

What we delivered

An on-site assistant that handles routine questions, captures leads, and escalates with context intact. Same stack and discipline we deploy on paid client AI work.

Why it matters

Shows founders what a bounded, useful retail AI looks like before they commit to a larger transformation programme.