AI Proofwork

We put AI to work in your business, and prove the results.

AI adoption consultancy, London

We help organisations find where AI belongs, build it into the way they work, and measure what it delivers against what was agreed at the start.

What you get at the end: a plain report that shows whether it worked.

Before work begins we define what success looks like in your terms. At the end you receive a plain report against those measures, including what did not work and what its limits are.

Illustrative of the measures we report. Yours are defined with you and measured on your own data.

Illustrated dial showing an agreed baseline, a target and a measured resultTargetBaselineMeasured result
Agreed baseline
Hours spent on the task today
Measured result
62%
Limits
Measured over the pilot period only
Example measures with their baseline, result, limits and target
MeasureAgreed baselineMeasured resultLimitsTarget
Time returned to the teamHours spent on the task today62%Measured over the pilot period only70
Outputs accepted without correctionCorrections needed before use, at kick-off84%On the agreed test set, not every case80
Weekly active use after handoverNobody using it, before handover91%First eight weeks after handover75
Decisions with human sign-offAgreed in the design: every consequential decision100%Automation never approves on its own100

What we do.

Eight things we help organisations with, from a first conversation about what AI could mean for you to using it every day. Our strength is not the technology. It is understanding your business well enough to know where AI belongs, and where it does not.

The problems we are asked to solve.

Most leaders are not asking whether to use AI. They are asking where it would actually help, and how to do it without creating risk. These are the questions we help organisations answer.

What would it be worth? Try the maths we use to qualify a workflow.

Released capacity only becomes money through avoided hiring or overtime, higher throughput, or another use of staff time. We say this before you have to.

Capacity released each month

66.7hours

Gross value at your cost per hour
£1,667
After a 70% realisation factor
£1,167
Less software and retained measurement
£667 a month

Worth assessing. Volume supports a pilot.

Illustrative. Uses a 70% realisation factor and £500 a month of software and retained measurement, the same assumptions as our assessment report. Every figure is replaced with yours during the assessment, and a no-build recommendation is a valid outcome.

How an engagement works, start to finish.

Engagements are staged so that each step earns the next. Every stage defines what success looks like before work begins and reports against it at the end.

Discover

We learn how your organisation works, what you are trying to achieve, and where the friction is. No slides, just questions.

What is recorded: how the work flows today

What it costs.

Each rung has a fixed fee or a from-price, a duration, and a named deliverable. What gets measured at each rung, and when it is reported, is in the section that follows.

Readiness assessment

£1,500

Fixed. Two weeks. A written assessment of one workflow or opportunity, a baseline, the risks, and a fixed-scope proposal or a no-build recommendation.

Assisted workflow pilot

From £4,500

Six to eight weeks. One workflow, one input channel, one output, a test set, staff training, human approval before any consequential action, and an evaluation report against the agreed measures.

Evaluation review

£1,500

Fixed. Two weeks. For organisations already experimenting with AI: one workflow assessed, failure modes documented, a prioritised improvement report.

Retained measurement

From £350

A month. Cancel monthly, renewed on the measured result. Scheduled health and usage review and minor changes within named service windows.

Fixed fees where scope is definable.

No charge for proposals.

Software and model costs paid directly by you.

Payment staged: on booking, on a testable demo, on acceptance.

How we know it worked.

Every engagement is measured against what we agreed at the start. No accuracy claims that have not been tested on your own data.

  • At kick-off

    We agree the baseline, the target and the method, in your terms and on your figures. That agreement is written down before any work starts.

    Week 0

  • Every Friday

    A three-line written update: done, next, blocked. First visible output by week two or three. A testable demo by week four or five.

    Weekly

  • At the end

    A plain report against the agreed measures: handling time before and after, error categories and review rate, known limitations, and a no-build recommendation where warranted.

    Week 6 to 8

Why AI Proofwork.

We are not a software vendor, a reseller or an outsourcer. We are the people you bring in to make sure adoption works, and to say so plainly when something will not.

  • Independent

    No reseller ties, no commission. Our advice on platforms and partners is about your fit, not our margin.

  • A person stays in charge

    AI drafts, your people decide. Nothing consequential happens without someone on your team saying yes.

  • Measured, not promised

    We agree what success looks like before we start, then report against it at the end, limits included.

  • No handoffs

    The person who scopes your work is the person who delivers it. Nobody learning on your time.

Let's talk about where AI belongs in your business.

An introductory conversation is without charge or obligation. Tell us where you are and what you are trying to achieve, and we will give you a straight view of what we would do first.

What happens next

  1. A conversation to understand your situation and ambitions
  2. A short written view of where we would start, and why
  3. A clear, staged proposal if you would like to go further
Talk to us