Engagement 01
AI Opportunity Scan
Two to three weeks. You end with a ranked catalog of the AI use cases worth building across the functions we cover, a build estimate against each one, and working software proving the best of them.
- Price
- $12,000
- Duration
- 2–3 weeks
- Your time
- ~10 interviews + a readout
Why start here
The cheapest way to find out whether the expensive thing is worth doing.
The costly mistake in AI is not picking a use case that fails. It is picking one at all, on the strength of a workshop, and then spending two quarters and a six-figure budget finding out.
The Scan takes that decision off guesswork. Instead of one bet, you get the whole field ranked, priced, and ordered by return — and the leading candidate built far enough that you can use it before you fund it.
It is deliberately the smallest commitment we offer. If it does not convince you, the correct next step is to stop, and you will have spent three weeks rather than three quarters learning that.
What the Scan is not
- A strategy deck with a maturity model in it
- A vendor or platform selection exercise
- A training programme or an innovation workshop
- A twelve-month roadmap you will never execute
Every deliverable is either a decision instrument or working software. Nothing in the package exists to justify the fee.
Week by week
What actually happens.
Three weeks, with a named output at the end of each one. You are never more than five working days from something you can act on.
Week 1
Interviews and the first roadmap
We interview the leaders who own the processes and the people who run them day to day — typically eight to twelve conversations. We map how the work moves today, where it stalls, and what each stall costs. By Friday of week one you have the prioritised roadmap in writing.
OutputProcess map · prioritised roadmap
Week 2
Scoring and estimating every candidate
Every viable use case across the functions we cover, written up and scored: annual value, feasibility, data readiness, and delivery risk. Each one carries a build estimate — weeks and a price band — so the list is a budget instrument, not a wish list. Most catalogs run to eight to fifteen candidates.
OutputAgents catalog with build estimates
Week 3
Prototype and readout
We build a rough working version of the strongest candidate — real software, shaped around your data, doing the actual task. It runs on de-identified samples or synthetic data your team approves; nothing connects to your production systems during a Scan. Then a 90-minute readout with your leadership team, and a one-page Build scope at a fixed price for whichever candidate you pick.
OutputWorking prototype · Build scope and price
Deliverables
Six things you keep.
All of them are yours, including if you decide not to work with us again.
The agents catalog
Every viable use case across the functions we cover, scored on annual value, feasibility, data readiness and risk, each with a build estimate in weeks and dollars. This is the artefact you will keep referring back to a year from now.
A process map
How the work moves through the areas we covered — including the handoffs and waiting time that nobody has written down before.
A prioritised roadmap, in week one
Sequenced by value and dependency. You do not wait until the end of the engagement to get the thing you can act on.
At least one working prototype
Rough, unhardened, and real. Built on the top candidate so your decision to fund a Build is made after using software rather than after reading an estimate.
A one-page Build scope with a fixed price
For the candidate you choose. Deliverables, acceptance criteria, timeline, price. You are free to take it to somebody else to deliver — the catalog is yours either way.
A 90-minute leadership readout
The findings, the ranking, the prototype, and a straight answer on which items are not worth building.
Everything above is fixed scope and fixed price. The only variable left is when it starts.
Check availabilityPrice
Fixed, and agreed before we start.
$12,000
Fixed. 2–3 weeks. Invoiced 50% on start, 50% on the readout.
- No hourly billing and no time-and-materials clause.
- Travel, if you want us on site, billed at cost.
- No obligation to commission the Build.
Compare it against the thing it prevents rather than against a day rate. A mid-market company that commits to the wrong first use case usually spends two quarters and a six-figure budget before anybody says so out loud.
Compare it against what you already spend, too. If your company pays more than this in a single month for software licences your people barely open, then three weeks to find out which of them an owned application could replace is not an expensive question to ask.
The Scan is priced to be a decision you can make without a steering committee.
Questions
Before you book.
Why does a three-week engagement cost this much?
Because of what it replaces. The alternative is committing a six-figure budget and two quarters to a use case chosen in a workshop, and finding out afterwards. The Scan is priced to be an easy decision for a company that spends more than this on software licences in a single month, and it is deliberately the smallest useful commitment we offer.
How much of my team’s time does it take?
Around eight to twelve interviews of 45 minutes, plus a 90-minute readout at the end. We take the load of writing everything up. If your leadership cannot find those hours in week one, the Scan will underperform and we will say so before we start rather than after.
How do you get access to our data that fast?
We don’t, and that is deliberate. Scan prototypes run on de-identified extracts or synthetic data that mirrors your schema — whatever your team is comfortable approving in week one. Nothing touches production, nothing needs your identity provider, and your security review stays off the critical path. Full environment access, SSO, and the formal review happen once — in the Build, under your policies, when there is something real to deploy.
Am I obliged to do the Build with you?
No. The catalog, the process map, the prototype and the roadmap are yours to keep and yours to take elsewhere. We would rather you commission the Build because the Scan was good than because you were locked in.
What if the catalog says nothing is worth building?
Then that is the finding, and we will write it down plainly. It happens when the underlying process is broken in a way AI cannot fix, or when the data does not exist yet. Knowing that after three weeks and a fixed fee is a good outcome — it is considerably cheaper than discovering it in month seven of a build.
Book the Scan
Bring us one process that costs you money.
On a 30-minute call we will tell you whether it is a genuine AI candidate, what a Scan would cover in your business, and when we could start. If your problem does not need a Scan, we will tell you that too.
No deck, no discovery questionnaire, no sales engineer on the line. You talk to the people who would do the work.