Your software is the thing
holding your business back.
You know you should be doing more with AI. But it's hard to bolt anything onto a system that was old five years ago. So I'd suggest rebuilding it — the software your business actually runs on, tailored to how you work, with the AI built in from the start.
Usually for around what you already pay in licence fees.
Based in the West Midlands · I work with a handful of clients at a time · You deal with me start to finish
Most businesses don't need AI added on. They need the old system gone.
Nearly every firm I talk to is paying a monthly licence for software that doesn't quite fit — so the team fills the gaps with spreadsheets, double entry and one person who knows how it really works. You can usually point to every one of those frustrations yourself. They've just never been solvable on the system you're on, so you live with them.
Replacing it is cheaper than owners expect, and it's the point where those frustrations can be removed, allowing AI to do something genuinely useful.
What I actually do
Three things, and the third is the one that makes the first two possible.
I replace the system
The software that runs your business was built for a generic version of your industry, so your team works around it. I build you something that fits your processes like a glove — and design the waste out on the way: the re-keying, the mouse clicks, the information your system already holds being copied from one screen to another. Between us we end up with a proper replacement you can be proud of.
The AI is built in, not bolted on
Because it’s your build, AI isn’t an integration you pay extra for. Bookings taken by phone. Emails read and answered in draft. Quotes assembled from what you already hold. It works to the rules you set, and your team approves anything that goes out.
For around what you already pay
This is the part people don’t expect. Modern AI development tools cut the build cost dramatically, so it can be rolled into a monthly fee — often close to what you’re already handing your current software vendor. You redirect a bill you’re paying anyway.
You don't have to replace everything at once, either. Plenty of jobs start as one AI solution sitting alongside what you've got, and grow from there. How a job runs →
What the AI does once it's in there
Not a chatbot in the corner — the donkey work, done, using the information your business already holds. Your team stops behaving like human robots and gets their time back for the work that actually needs a person.
Answers the phone and takes the booking
A phone agent that handles the routine calls — takes the booking, checks availability, confirms the details — and passes anything unusual straight to a person.
Reads the inbox and drafts the reply
Enquiries, requests for information, chasers. It pulls the details together, drafts the response to your rules, and leaves it for someone on your team to check and send.
Builds the quote from what you already hold
Prices from your own rate cards and rules rather than inventing a number, and assembles the quote from the information already sitting in your system.
Takes out the clicks
The fields your team retypes, the forms they fill from data the system already has, the same five screens every job. If the information exists, it shouldn’t need typing again.
Reads tenders and drafts the response
Pulls every question and the marking criteria out of an ITT, drafts answers grounded in your previous responses, and checks them against the criteria before submission.
Prepares the paperwork
Order confirmations, job sheets, service reports, certificates and handover packs — drafted from the job itself, ready for someone to sign off.
Fewer errors, fewer forgotten follow-ups, and the repetitive parts of the business turned into something that just runs. None of that is possible with off-the-shelf software — and a system built around you stopped being expensive a while ago.
Nothing has to change overnight
The two things that stop most owners starting: the risk of the switch, and not knowing whether it'll really work for them. Here's how both go away.
Run the old and the new side by side
Your current system doesn't get switched off on day one. The new one runs alongside it and you move across in stages — a process at a time, at your pace. You keep operating with no downtime, and your people get trained in stages, on processes that are simpler than the ones they're using today.
See it working before you commit
If it's something small — a couple of days' work — I'll build a working proof of concept on your own job, free of charge. Seeing it run on your business is what makes it real, and it shows you the kind of fit off-the-shelf software can't give you. No cost, and no obligation afterwards.
Work in progress, and work in use
Validera
LiveA recommendation tracker for their internal audit practice — findings followed through to closure in one place, instead of across spreadsheets and email chains.
Chase Plant Hire
In buildReplacing their off-the-shelf plant hire software with a bespoke system built around how they actually run the yard, with the AI in it from day one.
Tender response agent
In buildReads an ITT, extracts every question and the marking scheme, and drafts responses grounded in the answers you’ve already had marked well.
Why not just build it yourself?
It's a fair question, and one I'd ask. The tools that make this affordable are available to anyone now, and you could pick them up.
But you're already being pulled away from running the business. Doing it yourself means finding the time to map the process properly, learn the tools, build it, test it — and then keep maintaining it, on top of the day job.
And a working demo isn't a working system. If your customers are wary of AI making mistakes — and in a regulated industry they're right to be — what you actually need is the controls, the validation, the audit trail and a sensible way of working around it. That part comes from experience rather than tooling.
I build these things continually, so I already know the pitfalls, the patterns and the pieces that don't need reinventing every time. That usually means I can get you there faster, and for less, than building it from scratch — which is why it's worth getting a quote even if you fully intend to do it yourself. You might be surprised how affordable something properly designed turns out to be.
And if you're not ready to build anything at all, I'm happy to be a sounding board.
Why me, rather than anyone else
I'm Adrian Chapman. I run Chapman AI from Aldridge in the West Midlands, and I work with businesses across the UK — manufacturing, engineering, professional services, the kind of firms where the software was chosen years ago and everyone has quietly built workarounds since.
I trained in electronics and designed hardware systems, then wrote the firmware that ran on them. From there I worked on machine learning, and later deep learning, embedding those algorithms directly onto the hardware itself — and I was automating work back in the 2000s with Python and classical computing, long before any of it got called AI. Most people building AI solutions today, software developers included, have only started using it since ChatGPT arrived. I've had years to learn where these systems genuinely pay off and — just as usefully — where they fall over.
I've worked inside service organisations streamlining how they run: bringing in new processes, taking out waste, and getting people working in better ways. So I don't simply rebuild what you already have. While I'm designing the software I'm looking at the process with you — what can be simplified, and what only exists because the old system demanded it. Replacing your system is the one real opportunity you get to redesign the way you work. Improving the underlying processes is valuable in itself, and adding AI on top of that is a real multiplier — it's the difference between adding some AI and building something genuinely remarkable for your business.
I take on a few clients at a time, so what I build is shaped around how you actually work rather than pulled off a shelf. There's no account manager and no handover to a delivery team — you talk to the person building it, from the first conversation to the last release. And if you're nearby, I'll come to you.
Safe enough for your real work
The reason your team can use this on the everyday work they'd never put into a public AI tool.
Your data stays yours
Your choice where it lives: hosted in the cloud, or on your own server on site. The strongest AI models run off-site, so that step is processed elsewhere — and it can be kept to the UK or EU if that matters to you. What never happens, wherever it runs, is your data being used to train anyone’s model.
A person stays in control
The AI proposes; your team approves. Nothing goes out — no quote, no email, no booking confirmation — without sign-off, and every step is recorded.
Never invents a number
Prices and answers come from your own rates, records and rules. If it doesn’t know, it says so rather than guessing.
What are you paying for software right now?
That's the conversation. Tell me what you're running, what it costs you a month, and where your team works around it — and I'll tell you honestly whether replacing it makes sense. Let's have a chat, and you can bounce your ideas off me.