What is an AI discovery workshop?
It is the step before building. Instead of guessing what an AI project will take, you spend two focused days finding out. What are we building? For whom? What does it connect to? What will it cost, and is it worth it?
It is not a strategy deck. It is not a free sales pitch either. It is paid, which keeps both sides honest. We have to give you real answers, and you have to bring real information.
Why pay for discovery?
Because the alternative is paying for the same thinking later, hidden inside a build that went the wrong way. A fixed-price discovery puts the cost of finding out in one known place, before you've committed to the build.
It also protects the things that matter most in AI work: scope, data and who owns what. If any of those turn out to be a problem, you learn it for the price of two days, not two months.
What happens across the two days?
The work is done with the founders, Tim Morris and Jins Kadwood, not handed to a junior team. It falls into two halves: understanding the work, then shaping the answer.
Understanding the work
- The problem. What is slow, costly or error-prone today, and who feels it?
- The workflow. We walk through how the work really gets done, step by step, in the tools people use.
- The data. Where the information lives, who owns it and how clean it is.
- The risks. Privacy, record keeping and any rules your industry has to follow.
Shaping the answer
- Options. What could be built, and where AI is the right tool or not.
- Requirements. What each build must do, written so you can check it.
- The plan. The order of the work in two-week sprints, with a test week after each.
- The costing. What the build, and running it, will cost.
About four hours of the two days are face to face. The rest is us working on your problem.
What should you bring to an AI discovery?
You don't need to be ready. You do need to turn up with the real thing.
- A problem in one sentence. "Our quotes take two days to prepare" beats "we want AI".
- The people who do the work. The person who does the task knows more than the person who manages it.
- Examples. Real documents, spreadsheets, emails or screenshots, with anything private marked or removed.
- Your tools list. The systems involved, and who holds the logins.
- A decision maker. Someone who can say go or no-go at the end.
- Your constraints. Deadlines, rules, and the things you've already tried.
Client time through a build is about two hours a week each, approvals only. Discovery asks for more up front, so that the build can ask for less.
How do you get the most out of the two days?
- Show the ugly examples. The odd case that breaks the process teaches us more than the tidy one.
- Say what you've already tried. A failed tool or a past project saves us from repeating it.
- Keep the room small. The people who do the work and one who decides. A crowd slows it down.
- Be blunt about the budget and the deadline. A plan that ignores either is no use to you.
- Don't tidy first. You don't need clean data or a finished brief. Finding out what is messy is part of the job.
Discovery doesn't build the product. It makes sure the build, if you go ahead, starts with the right scope and a price you know.
What do you get at the end of an AI discovery?
| You get | What it is for |
|---|---|
| Requirements for each build | A written description of what it must do, so you can check the finished work against it |
| A refined plan | The order of the work, and what is in and out of each sprint |
| Costing | What each part will cost, with every line priced on its own so you can remove any you don't need |
| A fixed build price | A known cost for the build before it starts |
| A clear recommendation | Including when the answer is "don't build this yet" |
How does the go or no-go decision work?
At the end you decide. Go means we start the first two-week sprint at the price we gave you. No-go means you stop there, having spent only the discovery fee.
A no-go is a good result when the numbers don't stack up, or when the data isn't ready. We would rather say so after two days than in month three. The same applies to a "not yet". Sometimes the best advice is to fix the data first, which you can read about in how to prepare your business for AI.
What is an AI readiness assessment, and is discovery the same thing?
An AI readiness assessment asks whether your business is ready for AI: data, systems, people, rules. Discovery covers all of that, but it is tied to a specific thing you want to build and it ends in a price. A readiness assessment on its own can end in a report and nothing else.
If you aren't sure whether you're ready, discovery is the quickest honest way to find out. See the full path on how we work, and how to move a pilot into production once the build is done. If you are comparing partners, read how to choose an AI implementation partner.