Fixed price • 2–4 weeks • £15–25k
A diagnosis tells you where the constraint sits. The PoC Sprint proves the fix is real: working software in your context, built with your team, handed off cleanly.
I have shipped seven production AI products of my own. The sprint is the same muscle applied to your constraint: fast, honest, and designed from day one to be handed over, not to make me indispensable.
Proving it works is step one. Getting it used is the Embed Sprint — the next rung, and the one that decides whether any of this survives contact with your Monday morning.
Find where the constraint actually sits. Not where the org chart says it should be.
Build the thing that proves the pathway exists. Working software, not a slide.
Get it into live use with your people, on real cases, until it is load-bearing without me in the room.
Every engagement has a defined end. Extension is your choice, never my pitch.
I leave, and it keeps running. If you want a consultant who moves in for eighteen months, I am the wrong person — the whole point is that you don't need me.
From the field
A consultancy whose product is technical reports had directors as the bottleneck: every document queued for senior review. Two targeted AI experiments moved the constraint.
Lever: Tech — placed at the constraint, not sprayed across the firm
A ten-year-old software business asked for AI to automate its purchase-order workflow. The diagnosis said no — half the orders were raised after the fact. Fix the process; put AI where it pays.
Lever: Process first, then tech
Details anonymised. Clients are never named without permission.
Most PoC Sprints follow a diagnosis: the AI Maturity Diagnosis or the Pattern Diagnostic. That is deliberate. Building before diagnosing is how pilots end up in the graveyard. If you already know exactly what needs proving, we can start with a conversation instead. And if nothing is broken yet — you're starting an AI investment and want the first proof built properly — the sprint works just as well as a first move as it does as a recovery.
One distinction worth knowing: this sprint proves a technical pathway exists. If what has to change is how decisions and learning flow, that is the Operating Model Redesign Sprint — plateaued rollouts often need both, in that order.