Back to projects

Discovery Engine

AI-powered self-service discovery tool that guides clients through structured conversations to generate requirements documents. Built with Next.js, Supabase, and Claude API.

A self-service discovery tool that guides clients through structured conversations and produces requirements documents. Replace hours of discovery meetings with an AI-guided exploration they can complete on their own time.

Live: discovery.paz.land

What It Does

Clients select a discovery type, answer questions in a guided conversation, and receive a structured spec document at the end. You get qualified leads with real requirements instead of vague “let’s chat” requests.

For consultants and agencies: Clone your discovery brain into a YAML file. Clients self-qualify while you sleep.

For product teams: Gather requirements before the first meeting. Clients arrive having already thought through their needs.

How It Works

  1. Create a bearing - Define your discovery methodology in YAML (phases, questions, completion criteria)
  2. Client starts a session - They choose a discovery type and begin the conversation
  3. AI guides the exploration - Claude asks your questions, probes deeper, decides when each phase is complete
  4. Output generated - Structured markdown document summarizing what was discovered

No skipping phases. No checkbox forms. Real conversation that adapts to their answers.