Why I don't publish a price list
Every Intelligence Brain engagement I take on is different. Not slightly different — materially different. A 40-person broker in Limerick has a different data estate, different regulators looking over their shoulder, and a different appetite for change than a 200-person credit union network or a multi-site pharmacy group. Putting "€X per seat per month" on a webpage would be dishonest. It would either over-quote the small end and lose the work, or under-quote the larger end and leave me delivering at a loss.
So I don't do that. Instead I start every engagement with a free assessment, and I scope from there. What follows is how the pricing actually works, what drives it up and down, and what you can expect to pay roughly — so you can decide whether to bother having the conversation.
The free assessment
Before any money changes hands, I sit down with you — usually two sessions, sometimes three — and work out three things:
- What you actually have. The data, the systems, the documents, the people who hold knowledge in their heads. Most firms have far more than they think, and far less of it is usable than they hope.
- What you actually need. Not the AI feature list from a vendor deck. The specific tasks where an on-premise intelligence layer would shave hours, reduce risk, or stop something falling through the cracks.
- What's in the way. Regulatory exposure (Central Bank, GDPR, DORA, the EU AI Act), legacy systems, internal politics, skills gaps. I'd rather flag these on day one than discover them in month four.
You walk away from the assessment with a written scope and a fixed-price proposal. If you decide not to proceed, you've lost nothing and you've got a document you can use to brief somebody else. That happens occasionally and I'm fine with it.
What drives the price up or down
Once we're past the assessment, the engagement cost is shaped by a handful of real variables. In rough order of impact:
- Deployment model. Fully on-premise, in your own datacentre or server room, costs more to stand up than a hosted-but-isolated deployment. For most regulated firms the on-premise model is non-negotiable, and that's fine — I just want you to know it's the biggest single line item.
- Data integration scope. Pulling the brain into one system (say, your document management) is straightforward. Pulling it into seven systems with three different vintages of database underneath is not. The number and age of integrations matters more than almost anything else.
- Volume of unstructured content. Policies, board packs, meeting notes, client correspondence, regulatory submissions. Indexing this properly takes time, and the time scales with the volume.
- User count and roles. Not just how many people use it, but how many distinct roles need different views and permissions. A brokerage with five role types is more work than one with two.
- Compliance posture. Firms under direct Central Bank supervision, or operating under DORA, need a heavier audit and documentation wrap. That's labour, and it's reflected in the price.
Rough ranges, with the usual caveats
I'm reluctant to put numbers on a page because they get quoted back at me out of context. But people ask, so here's the honest shape of it.
For a small regulated firm — say 20 to 50 staff, one or two core systems, on-premise deployment — you should be budgeting in the low five figures for the initial rollout, plus an annual support and update fee. For a mid-sized firm with multiple systems and a heavier compliance load, that initial figure moves into the mid to high five figures. Larger multi-site or multi-entity rollouts are scoped individually and I won't pretend a generic range is useful for them.
What I will commit to: no per-seat licensing that punishes you for growing, no surprise charges for model updates within the support term, and no lock-in to a cloud vendor whose pricing you don't control.
What's included, and what isn't
A standard engagement includes the assessment, the deployment, integration with the systems agreed in scope, training for the users and at least one internal administrator, and a defined support window — usually twelve months — covering updates, fixes, and reasonable tuning.
It does not include hardware (you either own it or I'll specify what to buy), unrelated IT remediation work, or rebuilding source systems that the brain needs to read from. If your CRM is held together with duct tape, the brain won't fix the CRM. I'll tell you that during the assessment.
How payment is structured
I work on milestone-based invoicing, not upfront lump sums. Typically that means a deposit on signature, a payment at successful integration, and a final payment at user acceptance. Annual support is invoiced separately and renews on the anniversary. For firms that prefer to spread cost across financial years, I'm happy to discuss that — within reason and in writing.
What to do next
If you've read this far, the next step is the free assessment. There's no pitch deck and no salesperson. You'll be talking to me.
If you'd rather understand the product itself first, head back to the Intelligence Brain overview. If you're in financial services specifically and want to see how this lands in your sector, the financial services page is the right place to start.