You know you need to do something about AI. You've known for a while.

But where do you start? And who do you trust — when everyone selling it sounds exactly the same?

A 30-minute conversation What Merlin does

Your people haven't waited.

They're already using it — you just don't know which tools, what it's costing you, or what's exposed.

That's not a failure. It's where nearly every business we walk into is right now.

Do all your team use AI? Do you know which ones? And how comfortable are you with that?The first thing we ask.

The choice was never AI or no AI. It's AI you decided on, or AI that happened to you.

The conversations we keep having

None of them start with technology.

01Nobody decided this

“I know my people are using AI. I don't know which ones, I don't know what they're pasting into it, and I've never actually told them what's alright and what isn't. Nobody's done anything wrong — I just haven't given them anything to follow. And the longer I leave it, the harder it is to raise without sounding like I'm accusing someone.”

Company and customer information leaves the business through personal accounts nobody has any visibility of
Every person develops their own habits, so quality, tone and accuracy vary across the team
The exposure grows as these tools stop answering questions and start taking actions
Raising it late reads as a crackdown, which pushes usage further underground rather than bringing it into the open

02We tried it and nothing changed

“We bought the licences. We ran a session. Everyone came out saying it was interesting. Six weeks later I'd be surprised if three people are still using it. There's a line in the budget for something nobody opens.”

Money spent with nothing to show for it, which makes the next attempt much harder to fund internally
The team now believes AI "doesn't really work here" — an expensive belief to reverse
Licences renew automatically whether or not anyone signs in
Competitors who got it right compound their advantage while you start again

03We're not a tech business

“We're not a technology company. We make things and we look after customers. I keep reading that AI changes everything and I can't picture what it changes here. My problem isn't the shiny front end — it's four people in the back office doing the same job the same way they did it ten years ago.”

Productivity plateaus while labour cost keeps climbing
The only way to add capacity is to hire, because the process can't absorb more volume
Knowledge stays in individual heads and walks out of the building when they leave
The gap shows up as a margin problem long before anyone calls it a technology problem

04Reputation risk

“I've been asked what our AI plan is twice this year — once by the board, once by a client. Both times I said something vague. I can't speak on it with any authority, I don't know enough to tell who's talking sense, and I'm aware that being the person who didn't move on this is going to look a particular way in two years.”

Decisions get deferred, or made on the advice of whichever vendor was loudest
Fear of looking uninformed leads to buying the wrong thing, at the wrong scale, at the wrong time
Nobody internally will challenge a decision the leadership doesn't understand well enough to defend
The exposure is personal as well as commercial, which is why it keeps getting postponed

05The people problem

“If I stand up and announce an AI programme, half my team will hear 'redundancies'. The good ones will start looking. And the ones already using it quietly will go quieter, because now it's a thing with a policy attached.”

Adoption stalls, because nobody wants to be seen being a beginner at something that might replace them
The most capable people leave first — they're the ones with options
Usage goes underground, which is the exact outcome the programme was meant to prevent
The morale cost lands immediately; any productivity benefit arrives months later

06I can't measure any of it

“Say we do this. How would I know it worked? I can't put 'people feel a bit faster' in a board pack. I need to know what it cost, what came back, and who's actually using it — and every conversation I've had about AI so far has been about the technology, not about that.”

The investment can't be justified, defended or repeated without evidence
No way to separate productive use from expensive experimenting
Renewal decisions get made on anecdote and the loudest person in the room
Impossible to know which parts to expand, so nothing gets expanded

The third answer

So we built Merlin.

Most businesses answer this by buying licences and hoping, or by banning it and pretending. One gets you nothing; the other moves it onto personal phones where you can't see it at all.

Merlin is the third answer. We take each person in your business, work out what they actually do all week, and build AI into that specific job — one real workflow at a time, in their real work, with you able to see it happening. It isn't another app for you to roll out; the rolling out is ours. It's the decision, made and executed, about how this business uses AI.

One person. One job they hate. One tool built for it.

What it does

Someone finally decides.

Forty private decisions nobody has seen become one, made deliberately, that everyone is brought into. It isn't a technical block on anything — it's the agreed way of working, and people actually following it.

Your people get faster at the work they already do.

Each person builds something for a job they genuinely dread, and gets time back in the first week. Then it compounds, because each thing they build makes the next one easier.

And they stop being frightened of it.

Half a team think this is how they get replaced, so they won't touch it in front of anyone. Merlin is private, unhurried and unembarrassing by design. People come out of it more valuable, and knowing it.

You can see what it's worth.

What each person built, what they're getting better at, and where the next worthwhile piece of work sits. We measure it and show you the same view we're looking at — we don't promise you a number.

What's involved

The rolling out is ours.

1
You say yes. We email an agreement with a link — two minutes, from your phone.
Within the hour
2
A short onboarding form. Mostly one question: who's driving this internally?
Within 24 hours
3
Half an hour with your IT people. We talk to them directly.
First week
4
Kick-off with the team. One hour. Everyone leaves set up.
Week 1–2
5
Each person spends 5–7 minutes with Merlin, and their pathway starts.
Straight after

You will never be in a room with your IT person being asked whether you're comfortable giving us access to your systems. We have that conversation, with them, directly — and we're not asking for admin access to anything that touches your infrastructure.

What changes, and when

Week 1 from kick-off

The first working tool, built for someone's own job, saving real time.

Month 1

Most of the team have one. The patterns show up, and what's worth sharing surfaces.

Month 3

People spot things to automate without being asked. That's the change you were actually buying.

How it's priced

A one-off implementation, then a small monthly amount per person. You buy your own Claude seats.

Oversight without surveillance

You see what they're building. Never what they typed.

Nothing in normal use surfaces what anyone typed. Your manager view shows what each person built and what they're getting better at — the same view we're looking at. Any exception is a deliberate act by an owner, and it's logged — the same as email, or any other business system.

Merlin never connects to your systems or reads your data — it coaches, it doesn't reach in.

Why Claude, and not ChatGPT or Gemini?

Honestly? Because it's the one we'd put our name to — not the one we're paid to sell.

Because we picked one and went all the way in.

We could tell you we're platform-neutral. Plenty in this market do — and what it usually means is a programme spread thinly across three tools and tuned to none of them. Every module in Merlin, every workflow we build with your people, is built on Claude and tested on Claude. That's why it works on a Tuesday rather than in a demo.

It also happens to be the tool whose manager view matches what we promise your staff: you see what your people are building, not what they wrote. That's the difference between staff adopting this and staff hiding it.

And it sits alongside what you already run. Nothing gets ripped out.

For your team

It's just you and Merlin.

Your manager sees that you're making progress — never your mistakes, never what you typed.

The page written for the people who'll use it

Merlin

Everyone's work companion.

The next step is a half-hour conversation.

You'll know by the end of it whether this is right for your business — and if it isn't, we'll tell you that rather than sell you something that won't stick.

You'd be early, and I'd rather say that than dress it up.