Service · AI PoC & MVP Development

Prove it on your data before you commit a budget

A pilot is only useful if it can fail. We agree the pass mark first, run the idea against your real data, and write down whether it cleared the bar. A no is a cheap answer, and you keep it.

  • The pass mark is written and signed before any code is committed
  • Runs on your data, not a demo dataset
  • Fixed fee and fixed window, so a pilot cannot quietly become a project
  • You keep the code, the notes, and the result either way

What you get

Deliverables, not slideware

  • 01

    Pass mark definition

    The accuracy, cost, and speed the pilot must hit to be worth building, agreed before we start.

  • 02

    Working pilot

    A narrow but real implementation your team can use, running against your data.

  • 03

    Evidence pack

    The measurements, the failure cases, and the cost per run at your actual volume.

  • 04

    Go or no-go memo

    A recommendation with the scope and price of the full build if it is a go, and the reason if it is not.

pilot-result.md
  • 01pass mark: 90% correct routing, < $0.04 per item
  • 02measured: 93.1% correct across 1,240 items
  • 03cost per item: $0.021 at current volume
  • 04failure mode: multi-language threads (4.2%)
  • 05recommendation: GO — scope phase one at 6 weeks
Sample of the deliverable

How it runs

What actually happens, step by step

No discovery theatre. Each step ends with something you can read or use.

  1. 01

    Set the bar

    What has to be true for this to be worth building.

  2. 02

    Get the data

    A real sample, handled under your access rules.

  3. 03

    Build narrow

    One use case, done properly, in a fixed window.

  4. 04

    Measure

    Against the bar, including the cases where it fails.

  5. 05

    Decide

    Written memo: build, adjust, or stop.

Tangible

What lands in your hands

Named objects with a format and a week attached, so the handover is easy to picture.

  • Verdict memo

    Build, build with conditions, or do not build — with the reason attached.

    PDF · 9 pagesWeek 2
  • Data readiness read

    What exists, how complete it is, and what has to be fixed first.

    Findings tableWeek 1
  • Risk and blocker list

    Each risk with severity and the cheapest way to remove it.

    Ranked listWeek 2
Sample

Feasibility Verdict

Can it be built on your data — with the blocker and the cost to clear it.

9 pp
Download the sample (PDF)

This is the real template, with client figures replaced by representative ones. Read it before you talk to us — if the format is not useful to you, the engagement will not be either.

Before and after

What the change looks like in the week

Red is the cost you carry today. Green is what the system gives back.

Today

  • A vendor demo that never touched your data
  • A pilot with no end date and no definition of success
  • Nobody can say what the thing costs to run per month
  • The decision keeps getting deferred to the next meeting

After

  • A number you can check, from your own data
  • A fixed window with a written outcome at the end
  • Cost per run known before you scale it
  • A decision made on evidence, in weeks

Objections

Answered before you ask

What if the pilot fails?

Then you saved a build budget. You keep the evidence pack and the memo, and we say plainly that this one is not worth doing. That happens, and it is the point of the engagement.

How long does it take?

Typically two to four weeks from data access. The window is fixed at the start; if the scope has to grow, that is a new decision, not a silent extension.

Who owns the code?

You do, from day one, along with the notes and the measurements behind the result.

One next step, and it is a paid one on purpose

The AI Opportunity Diagnostic is a fixed-scope engagement. You leave with a ranked plan you can act on with us or without us.

Still sizing the problem? Write to us in your own words first. A person reads it.