Who I am, in plain terms
I'm Michael English — most people call me Mike. I founded IMPT.io in 2024 after twenty years inside large operators: Tesco, Dunnes Stores, and Oracle. I've sat on the buyer side of enterprise software contracts, on the vendor side selling them, and on the operator side where someone actually has to make the thing work on a Monday morning. That triangulation is why I build the way I build.
I live and work in Annerpark House in Clonmel, Co. Tipperary. The workshop is here. The thinking is done here. The Intelligence Brain — the product this page sits inside — was designed and prototyped here before it ever touched a customer environment.
Why I built the Intelligence Brain
By late 2024 I'd watched enough regulated firms — finance, legal, healthcare, professional services — try to bolt public LLMs onto their workflow and quietly back away from it. The reasons were always the same: data leaving the perimeter, no audit trail, no way to prove to a regulator what the model saw or said, and an inability to point at a specific document and say "this is where the answer came from."
The Intelligence Brain is my response to that. It's an on-premise organisational intelligence layer. Your documents, your policies, your historical decisions, your institutional memory — all of it sits inside your perimeter, indexed and queryable, and nothing leaves the building unless you explicitly send it. It's not a chatbot wrapper. It's the layer underneath, the part that should have been there before anyone started talking to a model.
The twenty years before IMPT
I won't bore you with the full CV, but the relevant parts are these:
- Tesco and Dunnes Stores — operations and systems work at scale. You learn quickly that anything that adds friction to a checkout, a stock count, or a back-office close gets rejected by the people who actually use it. That lesson sits underneath every design decision in the Brain.
- Oracle — enterprise software, the architecture of it, the selling of it, the implementation reality versus the slide deck. I saw what went wrong on big projects and why. Most of it wasn't technical. Most of it was that nobody mapped the actual work the software was supposed to support.
- The independent years — consulting and advisory work with mid-market firms in Ireland and the UK before founding IMPT. This is where I started to see the pattern that became the Brain.
The pattern, briefly: every regulated firm has an enormous amount of accumulated knowledge — case files, board packs, policy documents, email threads, decision logs — and almost none of it is retrievable in any practical sense. People rebuild the same answer from scratch every quarter because nobody can find what was already done.
The workshop in Clonmel
Annerpark House is a working address, not a marketing one. The Brain is built here. When I say on-premise I mean it — I run the same architecture I deploy to clients, on hardware I can put my hand on. The reason matters: if a client in a regulated sector wants to know whether the Brain will behave the way I claim it will, I need to have lived with it myself, on real workloads, before I put it in front of them.
Being based in Clonmel rather than Dublin or London is a choice. It keeps overheads honest, it keeps me close to the build, and it means when a client wants to come and see the thing run, there's an actual workshop to come to. Tipperary is two hours from Dublin and accessible from Cork and Shannon airports for international visitors.
How I work with clients
I don't run a sales team. If you're talking to someone about the Intelligence Brain, you're talking to me. That's deliberate for the current stage of the product — the deployments are bespoke enough that handing them to a salesperson would do the client a disservice.
A typical engagement looks like this:
- A scoping conversation — what's the actual problem, what data is in scope, what does "done" look like.
- A constrained pilot — narrow domain, real documents, measurable output. Usually four to eight weeks.
- A go/no-go decision based on the pilot, not on a demo.
- Production deployment on your infrastructure, with handover documentation and a maintenance arrangement that suits your IT function.
If at any point during the pilot it becomes clear the Brain isn't the right answer for your problem, I'll tell you. I'd rather lose the engagement than ship something that doesn't earn its keep.
What I won't do
I won't quote customer counts, logos, or revenue figures I can't substantiate. I won't promise a deployment timeline before I've seen your data estate. I won't claim the Brain does things it doesn't do — there are plenty of vendors making those claims and the market is already tired of them.
What I will do is show you the product running, explain honestly what it's good at and what it isn't, and let you make the call.
What to do next
If you want the full picture of the product itself — architecture, deployment model, what it actually does — the place to go is the Intelligence Brain product page. If you already know roughly what you need and want to see how it applies to your sector, the vertical pages under /intelligence-brain/ cover finance, legal, and professional services use cases in more depth.
Or write to me directly. I read everything that comes in.