What is the Intelligence Brain, in plain terms?
The Intelligence Brain is an on-premise organisational intelligence layer. It sits inside your network, ingests the documents and data you already have — policies, procedures, contracts, board packs, regulatory correspondence, operational logs — and lets your people ask it questions in natural language. It also surfaces patterns, contradictions, and risks across that material that no human reader would catch by working through it page by page.
It is not a chatbot bolted onto a public model. It is not a SaaS product where your data leaves your premises. It is a piece of software I install, configure to your environment, and hand over with the keys.
Who is it actually for?
I built it for regulated and operationally complex firms in Ireland and the UK who cannot send their data to a third-party cloud and cannot afford to wait twelve months for a transformation programme to deliver something useful.
The clearest fits so far are:
- Credit unions and smaller financial institutions — where compliance load is heavy and headcount is tight.
- Legal and professional services firms — where matter files, precedents, and regulatory updates need to be searchable without leaving the firm.
- Healthcare and care providers — where patient and operational data cannot move off-site.
- Public sector and semi-state bodies — where procurement, sovereignty, and data residency rules close the door on most cloud AI products.
- Mid-market manufacturing and logistics — where decades of operational knowledge sit in PDFs, spreadsheets, and people's heads.
If you are a two-person startup that just wants to summarise meetings, this is not the right tool for you. There are cheaper things on the market.
How is it different from ChatGPT, Copilot, or Gemini?
Those are general-purpose models served from someone else's infrastructure. They are good at what they do. They are also unsuitable for a lot of the work I get called in for, for three reasons.
First, data residency. When you put a contract or a board pack into a public model, you are sending it to that vendor's infrastructure under their terms. For regulated firms in Ireland that is often a non-starter, regardless of what the marketing page says about enterprise tiers.
Second, context. A general model knows the internet. It does not know your policies, your committee minutes, your supplier history, or the way your organisation actually runs. The Intelligence Brain is trained against your corpus and grounded in it, so the answers reference real documents you can open and verify.
Third, control. I install it on hardware you own or in a tenancy you control. You decide who has access. You decide what gets ingested. You decide when, and whether, it gets updated. There is no vendor on the other end watching the traffic.
Where does the data live?
On your premises, or in a private tenancy you control. That is the default and, for most clients, the only acceptable option. The model weights, the document index, the query logs — none of it leaves the boundary you set.
For organisations that are happy with a sovereign Irish or UK cloud setup, that is also possible. What I will not do is quietly route your data through a US-based public API and call it on-premise. That distinction matters and it is the line I will not cross.
How long does it take to deploy?
The honest answer is: it depends on what state your documents are in. The software install is days, not months. The work is in understanding your corpus, deciding what should and should not be ingested, agreeing access controls, and running the first few rounds of testing with your actual people asking their actual questions.
A typical first deployment runs in three phases:
- Discovery — usually two to three weeks. I sit with your team, look at what you have, and write down what good would look like.
- Pilot — six to ten weeks. A working Intelligence Brain against a defined slice of your data, used by a small group of named users.
- Rollout — scope-dependent. Wider access, more sources, integration with the systems you already use.
I will not quote you a fixed timeline before discovery. Anyone who does is guessing.
What does it cost, and how is it priced?
I do not publish a price list because the right number genuinely depends on scale, infrastructure, and how much of the work you want me to do versus your own team. What I can tell you is the shape: there is a one-off implementation fee covering discovery, install, and pilot; and an ongoing support and improvement arrangement after that. Hardware, if you need it, is separate and bought direct.
I would rather have a thirty-minute call and give you a real number than put a misleading figure on a website.
Is it safe? What about hallucinations and bias?
No system that uses language models is hallucination-free, and I will not pretend otherwise. What the Intelligence Brain does is constrain the model to your own documents, cite the source for every answer, and make it easy to flag a wrong or outdated response back to whoever owns that document. Over time, that loop is what builds trust.
For governance, I publish access logs, query logs, and ingestion logs to whoever you nominate as the owner. Your auditors get to see what the system was asked, what it answered, and where the answer came from.
What to do next
If you want the full picture of how the Intelligence Brain is built and what it does, start with the Intelligence Brain overview. If you already know your sector and want to see how this lands in your world, the vertical pages — credit unions, legal, healthcare, public sector — go into the specifics. If you would rather just talk to me, the contact route is on the main hub page and the email comes to me directly.