How Real Minds AI Compares
Four kinds of provider will answer your brief, and they are good at different things. Here is how to tell them apart, including where we are the wrong pick.
The four kinds of provider
Most briefs in this market get answered by one of four types. None of them is a bad choice. They are built for different problems, and the mismatch is what costs money.
Absorbed boutiques inside large firms. Serious engineering depth and enterprise governance. Best when the programme is large, the compliance load is heavy, and continuity across years matters more than speed. The people who pitch are often not the people who deliver.
General AI consultancies. Broad, tool-agnostic, comfortable across vendors. Best when you have not chosen a direction yet and want options laid out. The work usually arrives as a strategy before it arrives as a system.
Training providers and industry bodies. Cheap per seat, good at general AI literacy at scale. Best when you need many people to understand the basics. Rarely built to then change how the work actually runs.
Small specialist practices. Narrow, opinionated, founder-delivered. Best when you want one thing working soon and the people you met are the people who build it. Not built for a three-year programme across a large enterprise.
We are the fourth kind.
What we are, specifically
Founder-delivered. Tracy and Dennis do the work. There is no bench, no handover to a junior after the sale, and no account manager between you and the person building the thing.
A specialist Claude practice, not tool-agnostic. We work in one AI family and we know it deeply. That is a genuine trade: if your organisation has already standardised on something else, a tool-agnostic firm serves you better than we will.
Productised, with prices published. Every engagement has a name, a scope and a number on a public page. There is no discovery call required before we will tell you what something costs.
We run our own operation on it. Our own business is built on the tools we teach, so what we bring is what we use daily rather than what we read about.
We leave. The work is finished when your team can run it without us. That is the opposite of a retainer model, and it is deliberate.
Questions worth asking any provider
Ask us these too. The answers separate the four types faster than a capability deck.
Who actually does the work? Ask for the names, and whether they are in the room today.
What does it cost, before a discovery call? A provider who cannot answer this has not productised anything.
What do we own at the end? Check whether the thing runs in your tenancy, on your data, without a licence to keep paying.
What happens if you stop working with us? Ask what breaks. If the honest answer is everything, that is a dependency, not a solution.
Show me something you built, running. Not a case study. The thing itself.
When we are the wrong choice
This is the part most comparison pages leave out, so it is the part worth reading.
You need a large multi-year programme. Two founders cannot staff that, and pretending otherwise would fail slowly and expensively. Go to a firm with a bench.
You are committed to a different AI vendor. Our depth is in one family. A tool-agnostic consultancy will serve you better and we will say so.
You want a strategy document. We are not good at producing decks that describe work instead of doing it. If the board needs a written position first, a general consultancy is the better buy.
You want someone to own it forever. We build so we can leave. If you would rather outsource the capability permanently, a managed service fits better than we do.
We would rather lose a job than take one we are wrong for. It is cheaper for both of us.
More on what we turn down · AI Services by Engagement Stage · AI diagnostic scope
Tell us where it hurts
A free first conversation. If we are not the right fit, we will say which of the four types is.
Tell us where it hurts