AI Consultancy
Find out where AI actually pays before you spend anything building it
Most AI budgets get spent on the most visible process rather than the most expensive one. We map how the work really flows, put a cost against it, and tell you which parts are worth automating — and which aren't.
The problem
The loudest complaint is rarely the biggest leak
Nearly every team we meet knows something is slow. Almost none of them have ever put a number against it. Without that number there is no way to compare one fix to another, so the decision gets made on instinct — and instinct buys tools. A subscription feels like progress, but a tool dropped on top of a broken process just adds another place for the work to queue up.
The other trap is volume of noise. The process people moan about is usually the annoying one, not the expensive one: fifteen minutes of irritation a day beats four hours of quiet re-typing that nobody complains about because it has always been done that way. Costing the work in hours and pounds re-orders the list almost every time, and the thing at the top is normally not what anyone expected.
What this includes
What we actually do in the audit
Process mapping with the people doing the work
We run the sessions with the people who actually run the process, not just the leadership view of it. What comes out is the real flow, including the workarounds nobody documented.
Opportunity scoring on effort, cost and risk
Every candidate gets scored on what it costs you today, what it would take to fix, and what could go wrong if it went wrong. That gives one ranked list instead of six competing opinions.
A simple ROI model per opportunity
For each opportunity we model hours saved, revenue recovered or errors avoided, using your own numbers. No blended industry averages — figures you can check against your own tools.
Tooling and stack review
Plenty of businesses already pay for capability they never switched on. We review what you own across CRM, helpdesk and productivity licences before recommending anything new.
A 90-day roadmap
The output is sequenced, not a wish list: what to do first, what follows it, and what to leave alone this quarter. Each item carries an expected outcome and an owner.
Readiness view of data, permissions and ownership
Most stalled AI projects stall on access, not ambition. We flag where data is too messy, where permissions will be a blocker, and where no one internally owns the outcome yet.
How it runs
One week, four stages, no theatre
- 01
Discovery call (30 minutes)
A straight conversation about where the time and money go. We agree which one or two workflows are worth looking at properly.
- 02
Working sessions (2–3 sessions across one week)
Short, focused sessions with the people who run the process, mapping each step, handover and delay as it really happens.
- 03
Analysis and modelling (3–4 days)
We cost the current state, score the opportunities and build the ROI model. This is where the ranking gets decided.
- 04
Findings and roadmap walkthrough (90 minutes)
We walk your team through the findings live, argue the trade-offs, then hand over the roadmap in writing.
What you end up with
What you end up with
A prioritised list with numbers attached
Every opportunity carries a cost today and an expected return, so the ranking is arguable on evidence rather than opinion.
A clear 'do this first' recommendation
One recommendation, stated plainly, with the reasoning behind why it beat the alternatives.
An honest list of what not to build
The items we think are not worth automating yet, marked as such, with the reason they failed the test.
A roadmap you can act on either way
It is written so another team could pick it up. You are free to build it with us, in-house, or not at all.
Next step
Find out where the money is actually leaking
Thirty minutes, no pitch. You will leave with an opinion on where AI pays in your business, whether or not you work with us.
If the audit doesn't find a credible use case, we'll tell you honestly and we won't push a build.