One agent. Four weeks. In production.
$12,500 fixed. One AI agent shipped to your stack in four weeks, scoped from your diagnostic. Production observability included. We build one agent that actually runs, not five half-built ones.
A working agent, not a demo.
One production-ready agent
Picked from the diagnostic roadmap. Built to your stack and your data. Observable, replayable, documented.
Eval suite
Concrete pass/fail tests covering golden-path and edge cases. Run on every change. CI-friendly.
Observability + cost telemetry
Every invocation tracked. Token spend, latency, failure modes. You see what the agent is doing in real time.
Operator runbook
How to monitor, intervene, kill, and roll back. Written for whoever owns operations after we leave.
Handoff session
Your team and ours, on a call, walking the agent end-to-end. Recorded. Q&A until they can run it solo.
Four-week shape.
Week 1
Scope lock + data access
Final scope agreement, environment access, eval set defined.
Week 2
Build
Build the agent against eval set. Daily progress notes.
Week 3
Hardening
Edge cases, failure modes, cost optimization, observability wiring.
Week 4
Production + handoff
Deploy to your environment, runbook delivered, handoff session, support window opens.
$6,250 to start. $6,250 on delivery.
Stripe Invoice. Net-zero on receipt. We start the moment first invoice clears. Final invoice releases on successful handoff session.
Start with the diagnosticSprint scope is locked from the diagnostic. We won't scope a sprint without one.