Find the value, prove it, then build it
Every engagement runs the same way. We start by finding where the return is, we prove the best opportunity on your own data, and we only build what has earned its place. No open-ended consulting, no pilots that quietly never end.
Phase 01
Find the value
Phase 02
Prove it on your data
Phase 03
Build what proved out
Most AI work fails in the gap between an idea and a working system
We built our approach around avoiding both. The value comes first, the proof runs on your actual data with all its inconsistencies, and the production build is scoped only once you have used the thing and know what it is worth.
Before any of this
If you are not sure whether a workshop is the right starting point, take a free hour first. One conversation, no preparation, no obligation, and an honest answer about whether there is anything here worth doing. Plenty of people go straight to the workshop. Nobody is expected to.
Clarity before commitment
01 Find the value
Some businesses come to us with a specific problem they want solved. Others come knowing they should be doing something with AI and having no idea where to start. Both are fine starting points, and both begin the same way.
The Discovery Workshop is half a day with the people who actually do the work, not just the leadership team. We map where your data lives, what state it is in, and what questions the business cannot currently answer.
Map the data landscape. Which systems hold what, what condition it is in, what is duplicated and what is missing.
Surface the use cases. A typical workshop produces ten to twenty candidate opportunities, drawn from the frustrations your team lives with daily.
Rank them on value and urgency. Not everything worth doing is worth doing now. We score each one so the sequence is a business decision, not a technical one.
Cost the first step. You leave with a defined, priced next step for the opportunity that ranked highest.
We talk to the people doing the work, not just the people describing it
Every use case is tied to a business outcome, not a capability
We tell you which ones are not worth doing
The output is yours to act on, with us or without us
A working system, not a slide deck
02 Prove it on your data
This is where most of the risk sits, so this is where we put the effort. We build the top-ranked use case as a working prototype, hosted and loaded with your data, and we give your team access to use it on real decisions.
Then we compare. Your people run their weekly cycle the way they always have, and they run it through the prototype, and we look at where the two disagree. Every disagreement is either a gap in the data, a rule we have not captured yet, or a better answer. All three are worth knowing.
Build on your real data. Not a sample, not a cleaned-up extract. The messy version, because that is what production will face.
Encode your judgement. The rules your experienced people carry in their heads get written down, tested and refined until the system reflects how your business actually decides.
Run a real evaluation period. Your team uses it on live decisions, and the outputs get challenged.
Show the working. Every recommendation carries its reasoning and a confidence flag, so your people know where to check.
Fixed fee with defined success criteria, agreed before we start
Built on your stack, so your data, governance and running costs stay in your control
The prototype interface is deliberately disposable. We are testing the thinking, not the buttons
If it does not prove out, we say so
Production, and then compounding
03 Build what proved out
Once a use case has been used in anger and the business trusts it, we scope and fix-price the production build with real knowledge of what it takes: the refresh cadence, the running costs, the interface people actually want.
The important part happens after that. The connected, governed data foundation built for the first use case is the same foundation the next five run on. The second use case costs a fraction of the first. That is where the return compounds.
Productionise. Proper engineering, proper security, proper refresh.
Extend the foundation. Each new source connected makes every future use case cheaper.
Transfer the capability. Your team learns to run and question it, because a system nobody understands does not survive its first surprise.
Keep curating. Context is not a one-off build. It needs maintaining as your business changes.
Scoped from evidence, not from assumption
Fixed price, with the running costs stated up front
Your team is trained to run and question it
Each use case makes the next one cheaper
Start with the question, not the technology
A Discovery Workshop is half a day with the people who actually do the work. You leave with a ranked list of where AI and connected data will pay for themselves in your business, and a costed first step.
QUESTIONS ABOUT THE PROCESS
Before you book
The workshop is half a day, with a written output back inside two weeks. A prototype phase typically runs over several weeks including a real evaluation period. Production scope depends entirely on what proved out, and we will not estimate it before we know.
Nothing. Bring the people who do the work. Preparing tidy inputs would defeat the purpose, because we need to see the real state of things.
No. Each phase stands on its own and ends with something you own. Plenty of businesses take the workshop output and act on part of it themselves.
Then we have spent a small amount finding out, instead of a large amount discovering it at go-live. We will tell you plainly, and explain what we learned about why.