Start here
Know exactly where AI pays back — before you spend
A short, evidence-led engagement that maps where AI will actually reduce cost or recover time in your business, ranks the options by return and risk, and gives you a sequenced plan you can budget against.
The problem
Why this keeps going wrong
Almost every mid-sized business we meet is in the same position. There is real pressure to "do something about AI" — from the board, from competitors, sometimes from customers. There is no shortage of vendors willing to sell something. What is missing is any basis for deciding between them.
So the decision gets made on the wrong criteria: whoever demoed most convincingly, whatever a competitor announced on LinkedIn, or whichever tool the most enthusiastic person in the business already likes. Six months later there is spend, a pilot nobody uses, and no way to tell whether the next investment should be larger or smaller.
The fix is not more research. It is applying the same discipline you would apply to any other capital decision: what is the process, what does it cost today, what would it cost automated, what is the risk, and in what order should we do it.
You will recognise this if
- You have been asked for an AI strategy and do not know where to start.
- You are evaluating vendors with no way to compare their claims.
- A pilot has stalled and nobody can say why, or what it returned.
- Different parts of the business are buying different tools independently.
- You suspect you are spending on AI already but cannot total it.
How we do it
The engagement, week by week
No discovery phase that never ends. Every stage produces something you can hold.
-
Find out what is actually happening
We interview across functions, not just IT, and we ask about tools people use rather than tools that are approved. This is deliberately non-punitive — the objective is an accurate picture, and you only get one if nobody is worried about admitting things. In parallel we look at spend, systems and where data actually lives.
You get: A current-state map: tools in use, owners, data touched, spend to date.
-
Quantify the processes, not the technology
For each candidate area we establish volume and cost as they stand today: how many calls, how many hours, how many errors, what each one costs. Most businesses have never measured this, and the measurement alone frequently changes which opportunity looks best.
You get: A baseline you can measure future change against.
-
Score every opportunity on the same scale
Each use case is scored on time or cost recovered, implementation effort, dependency risk, and governance risk. Scoring on one scale is what makes the comparison honest — and it is usually where a favourite idea drops down the list and an unglamorous one rises.
You get: A ranked opportunity register with the scoring shown, not hidden.
-
Sequence it, cost it, and say what to avoid
We turn the register into a phased plan with indicative cost and effort, dependencies made explicit, and a specific list of things we recommend you do not do this year and why. Then we present it to whoever has to approve it, and take the questions.
You get: A sequenced roadmap and a six-page board summary.
What you get
Deliverables, not a workshop
-
Current-state assessment
Every AI tool in use across the business, who owns it, and what data it touches.
-
Opportunity register
Each candidate use case scored on time recovered, cost, effort and risk.
-
Sequenced roadmap
What to build first, second and third, with indicative cost and effort per phase.
-
Board-ready summary
Six pages in commercial language, not a technical specification.
Included
- Interviews across every function that touches the candidate processes
- Discovery of unsanctioned AI tool usage
- Baseline measurement of current volume and cost
- Scoring and ranking of every candidate use case
- Indicative cost and effort estimates per phase
- A presentation to your board or leadership team, with the questions taken live
Not included
- Building anything — this engagement ends with a decision, not a deployment
- Vendor procurement or contract negotiation on your behalf
- A full data warehouse or systems architecture review
- Change management delivery, though we will flag where it will be needed
If you need any of these, we will say so and point you at someone who does them well.
How you judge us
What we measure
Agreed at kick-off, reported against at the end. If we cannot move these, the engagement did not work and we would rather both of us know that early.
- Number of AI tools discovered versus number previously known to management
- Baseline hours and cost established per candidate process
- Use cases ranked, with the top three costed to ±30%
- A decision made by leadership — the engagement has failed if it produces only a document
In context
What this looks like in your sector
FAQ
AI Readiness & Roadmap: your questions
Is three weeks really enough?
For a mid-sized business, yes — and the constraint is deliberate. A longer engagement produces a thicker document, not a better decision. If we find during week one that your situation genuinely needs more, we will tell you then rather than at the end.
What if the honest answer is that we should not invest yet?
Then that is what the report says, and we would rather deliver it than sell you a build. It happens perhaps one engagement in six, usually because the underlying data or process is not ready. Knowing that in three weeks is considerably cheaper than finding out in nine months.
Do we need to have data ready first?
No. Assessing the state of your data is part of the work. If a use case is blocked by data quality, that dependency appears in the roadmap as a prerequisite with its own cost.
Who needs to be involved from our side?
A sponsor with authority, and perhaps six to ten hours in total across the people who actually run the processes. We work around your operation rather than pausing it.
Will you then quote to build what you recommended?
If you want us to, and you are free not to. The roadmap is written so that another firm could execute it — which is the only way it can be trusted as advice rather than as a sales document.
Start with a conversation, not a proposal
Thirty minutes. We will tell you whether this is the right place to start for you — including if it is not.