Home / Insights / The two-week agent build: what actually fits i
// playbook

The two-week agent build: what actually fits in a fortnight

A working agent in weeks isn't a stunt — it's a scoping discipline. Here's what fits, and what doesn't.

2026-05-15 · 5 min read

Scope to one job

A fortnight is plenty to ship one agent that does one job well. It is not enough to 'transform the business'. The teams that ship fast are the ones that resist scope.

Pick the single highest-payback job — usually qualifying or routing inbound — and build that. Everything else is a later iteration.

Week one: a working prototype

The first week should end with the agent talking. Design the conversation, train it on your real content and tone, wire the one integration it needs, and get it answering on a staging page.

Seeing it work in week one is the point. You make decisions against a real thing, not a slide deck.

Week two: harden and ship

The second week is for the unglamorous work: edge cases, handoff to a human, capturing the lead or booking cleanly, and the analytics to know it's working.

Then it goes live on real traffic. A narrow agent in production beats a broad one in a backlog every time.

// keep reading

Related notes.

Let's find your first agent.

One call. We'll point at the agent that'll pay for itself fastest — and tell you straight if you don't need one yet.

Book a call →