An independent AI advisor for decisions that are expensive to reverse
Investing in AI, buying a company built on it, or putting a budget behind it? I'm an independent AI advisor: I give you a technical judgement on what's in front of you, with no vendor to recommend and no build work to sell you afterwards.
Dr Colin Kelly
PhD in NLP (Cambridge)
Former Head of Applied AI Research
Fully independent. No vendor affiliation, no implementation upsell
What an independent AI advisor does
You know how you get a survey before you buy a house? An independent AI advisor does that for an AI decision. I look at the model, the data, the engineering and the claims made for them, and tell you what's real, what it will cost to keep working, and whether the advantage will last.
The report leaves you with the same four options a survey does: go ahead, re-price, set conditions, or walk away. What it doesn't do is turn into a quote for the repairs, because I don't do the repairs. That one rule is what makes the advice independent, and it's the reason you can put the findings in front of a board or an investment committee without anyone asking what I stood to gain.
Most of this work is AI due diligence for investors and acquirers. The same judgement is available to leaders who own an AI decision and have to take it to a non-technical board.
Why independence matters
Plenty of people can tell you about AI. Far fewer can tell you about it without a product, a partnership or a follow-on project shaping the answer. Independence here is structural, and it comes down to four commitments.
No build work on anything I assess
If I've assessed a company for you, I don't take build or remediation work from it afterwards. The findings can't become a pitch for my own next engagement.
No vendor affiliation
I'm not a partner, reseller or referrer for any AI vendor, platform or model provider. When I compare options, nobody is paying me for the answer.
The person you brief does the work
I read the code, run the interviews and write the report myself. There is no junior team behind the senior face on the first call.
Paid by you, and only you
The fee comes from the person making the decision, so the advice answers to that decision and nothing else. I'll tell you quickly if it's not a fit.
Who hires an independent AI advisor
Investors and acquirers
Pricing a deal where the AI is a large part of the value. You need to know whether the AI is worth the price, and whether the advantage will last, before the money moves.
Boards and leaders approving an AI budget
You own the AI decision and have to take it to people who aren't technical. You need a read you can defend in the room, from someone with no stake in the outcome.
Founders preparing for a raise or a sale
The claim in the deck will be tested by the other side. An independent assessment before the process opens means you find the problems first.
Executives choosing a vendor or platform
Supplier demos are convincing. You want someone who has built these systems to tell you what will hold up in production and what the switching costs look like.
What I do, and what I don't
The lead service is AI due diligence: a two-week technical assessment for investors, acquirers and owners, with a written report that gives a severity rating on each finding. It also runs sell-side, for founders who want the AI verified before a buyer's diligence starts.
For leaders who own an AI decision and have to carry it to a board, AI leadership advisory is the same judgement on a longer timeline: what's worth doing, in what order, and what it will cost. For a small number of companies at a time I take a fractional CTO role, and the companies I do that for are not ones I've assessed for an investor.
What I don't do is build. I'll design, assess and advise on an AI system, and I'll help you choose who builds it, but I don't build the thing I've just surveyed.
Who you're dealing with
Independence only counts if the judgement behind it is sound. I've built AI systems and led the teams that build them, which is how I know which claims to test first and where production systems break.
- PhD in Natural Language Processing, University of Cambridge, and a BA in Mathematics and Computer Science from Oxford
- Over fifteen years building and assessing AI systems, across IBM, PA Consulting and two VC-backed AI scale-ups
- Former Head of Applied AI Research, leading the team delivering production AI capabilities
- Identified over €50m in AI value at IBM across telecoms, financial services and automotive
- Eighteen months as fractional CTO taking an AI product from concept to public beta
- AI strategy and implementation roadmap for a major global foundation
Case Studies
Fractional CTO Building Patent-Pending AI Product from Zero to Beta
How we took a ghostwriting firm from copy-pasting into ChatGPT to a patent-pending AI product in beta with enterprise users.
20 months
Engagement duration
5 hires
Team recruited
85%
AI rendering speedup
AI Strategy and Implementation Roadmap for a Major Global Foundation
How we helped a large communications division turn AI momentum into an implementation plan — in six weeks.
6 weeks
Engagement duration
11
Stakeholder interviews
2 delivered
Technical blueprints
AI Leadership Advisory: Transforming a Communications Firm's AI Capability
How we turned fragmented AI experimentation into shared frameworks, tested playbooks, and a concrete product roadmap over two months.
2 months
Engagement duration
3 half-day
Workshops delivered
Full team
Team members trained
Frequently Asked Questions
How is an independent AI adviser different from a large consultancy?
Two things. A large firm usually has an implementation practice, a partner programme or both, so the advice sits next to something it would like to sell you. I have neither, and I take no build work on anything I assess. The second difference is who does the work: the person you brief is the person who reads the code, runs the interviews and writes the report.
Do you recommend specific AI vendors or tools?
I'll tell you which options fit your situation, what each one will cost you over time and where it breaks. I don't take referral fees or partner commissions, so the answer doesn't change depending on who you pick.
Will you build what you recommend?
No. On anything I've assessed, I take no build or remediation work, so none of the recommendations are work for me. Where I act as a design partner or fractional CTO, it's for a company I haven't assessed on someone else's behalf. You can be the buyer's surveyor or the architect on a house, and I don't do both on the same one.
What does an engagement look like?
It starts with a conversation about the decision you're making and what would change it. AI due diligence runs as a two-week assessment with a written report and a live debrief, priced from $15,000. Advisory for leaders is an ongoing relationship, scaled up during hiring or major technical decisions and down in between.
Do you make the investment recommendation?
I don't, because that decision depends on factors beyond the technology. What you get is a technical verdict on each claim, a severity rating on each finding, and what each one means for your thesis or your budget.
Do you work under NDA?
Yes. Engagements are confidential and I work under your NDA.
Weighing up a large firm against an independent? Read Big 4 consultant vs independent AI advisor.