AI advisory board
Boards, investors and leadership teams are being asked to decide on AI, data and security questions that arrive faster than anyone has time to learn them. I am an independent technologist who has been the CTO and COO on the receiving end of those decisions, and I am setting up a monthly advisory service that keeps the AI and technology questions in check for you. You get a written baseline first, then a standing monthly session, and a specialist study whenever a question needs a proper answer.
Management says the AI plan is in hand, a vendor says the data is safe, and a competitor has just announced something. A board needs a second opinion that has no product to sell and no reason to flatter anybody, and that explains the answer in the language of the board.
Before an investment, whether the AI is a product. After it, whether the company is using AI where it pays and keeping its data and security in order. The kit and the monthly brief work as a common standard across several companies, and each one still gets its own answers.
A CEO is often the one who has to say yes or no to the AI proposals on the table, with a CTO whose time is already spoken for. A monthly conversation with someone who has no stake in the answer makes those calls quicker.
Risk, audit, data protection and technology committees each get a piece of the AI question. I can sit with one of them, or serve all of them from the same baseline, so they stop reaching different conclusions about the same system.
The kit is made for your organisation, from a base that stays the same: how decisions about AI are made and recorded, who owns what, what data may go to which model and where it is processed, the review and security practice around AI tools, and the questions a director should expect management to answer. It ends up as a short document the board adopts, which also makes it the first line in the minutes.
Before each session you get a short written brief on what changed in AI and technology that matters to you: new models and prices, regulation, incidents elsewhere, and what your own management reported. In the session we go through the register, what to do about the new items, and what you are choosing not to do and why. Between sessions, a chair or CEO can ask me a question and get a straight answer.
Some questions need two weeks rather than ten minutes. A study answers one of them in writing, with the evidence attached, in a form that can go into a board pack. Larger studies are done with the AI SWAT team when the work needs more hands.
The training takes your board secretary, CTO or a few directors through building an AI agent that supports your board meetings, using AI assistants to build it. The agent reads the board pack and the register, and prepares an AI risk and opportunity assessment for each meeting, with its sources shown so a person can check them. Because your own people build it, they know what it can do and where it is confidently wrong, and the data stays inside the setup you chose.
A board decision only holds when the people who carry it out can do so. I pair with your CTO, head of product or engineering lead on the practical side of the same questions: agent rules and review practice, how to evaluate AI features, what to build and what to buy. I also read what they report to the board, so the board and the people running the technology are working from the same picture.
Directors tend to carry AI as a vague worry with no name on it. After the baseline each risk has a name, an owner and a status, and an independent person has said which ones matter. The relaxed feeling that follows is earned, because it comes from having looked.
The decisions stay with the board. What the advice changes is the record: the board can show that it asked the right questions, got independent written advice and weighed it. That is a stronger position than a decision made on a vendor's slides, even if I cannot take on anybody's legal liability, and I will not pretend to.
A cautious board can lose as much by waiting as a hasty one by rushing. The register has a column for opportunities, reviewed every month, with the cost of not acting written next to the cost of acting. Some items stay on the list for a year and some are done by spring, and nobody has to remember why.
Both sides of the table
COO and CTO through two acquisitions, which means technical due diligence seen from the company being bought and from the side buying. I know what a board and an investor ask about technology when a lot of money depends on the answer.
Security in the pipeline
Security requirements built into the delivery pipeline, with a watertight audit trail as a first requirement, and an on-premises language-model proof of concept with the National Cyber Security Centre Finland. That is the habit of asking where the data goes before asking what the model can do.
Built end to end
A document-analysis product over nearly 18,000 pages of maritime rules, now with a commercial shipyard pilot, built in an agentic pipeline whose rules, gates and failures I know first-hand. The advice is based on running this work myself.
Worth saying plainly
I am an independent technologist, with no products, resale or vendor commissions. I am not a law firm: where a question needs a legal opinion, I prepare the technical facts and your counsel gives the opinion. I do not sit on your board.
We talk about what the board is facing and what is already in place. If another kind of help fits better, I will say so.
A fixed-scope engagement that produces the baseline document and the first risk and opportunity register. It is what you adopt, and you keep it if the monthly service stops.
A monthly fee for the brief, the session, the direct line and the register. Studies, training and pairing are quoted separately, and a few clients at a time is the most I take, so that each one gets real attention. The price is higher than a generalist consultant's, and I explain it on the first call.