For executives and investors with a decision to make
Before the next pilot.
A clearer decision.
Before funding another AI pilot or commissioning more software, make the decision explicit: which business capability needs to change, what would justify the investment, and what must work beyond the prototype?
I bring technology operating experience, software engineering and ongoing doctoral research to that question. The review below is a defined starting point, not a commitment to a large transformation programme. For a chair, founder or nomination team considering a board or ongoing advisory remit, there is a separate path.
Start with the decision in front of you.
CEOs and operating leaders
Which AI investment deserves to go further?
You have a promising pilot, an unclear business case, or a workflow that needs to change. The next decision is what to fund, buy, build or stop. A particular focus is document-intensive service businesses, where review work, exceptions and integration shape the economics.
Review a business decision →Investors and portfolio leaders
Where does the opportunity survive contact with operations?
You need to distinguish a plausible AI story from a capability a business can use and maintain. Start with one company, one workflow and a material decision before considering a wider portfolio programme.
Examine a portfolio-company decision →For chairs and nomination teams: board and strategic advisory has a separate remit and conflict check.
One scoped starting engagement
AI Value and Execution Review
A focused review of one business decision, with a recommendation you can examine and a next step you can test.
A useful fit
For a leadership team with a pending investment or operating decision, a named decision owner, and access to the people and evidence needed to examine it.
What this is not
Not a general AI introduction, an open-ended transformation programme, a compliance certification, or a promise to justify a preferred supplier.
What you take away
01
A decision memo
A recommendation and comparison of buy, build, partner and no-change options, with the evidence supporting each.
02
An assumptions and risks record
What remains uncertain, what the recommendation depends on, who owns the decision, and the operating dependencies that could change it.
03
An investment gate
A bounded next test, its baseline and acceptance criteria, and the conditions for continuing, changing direction or stopping.
What we need to examine it
- The decision, its owner and timing, plus the current workflow and alternatives already considered.
- A baseline for cost, quality or service, including human review, exceptions and failure costs where relevant.
- Agreed access to stakeholders, relevant system constraints and a bounded evidence sample, subject to confidentiality and data-handling terms.
How the work is scoped
Agree the question
Check fit and conflicts, then agree scope, fee, timetable, access and confidentiality in writing before work begins.
Examine the evidence
Work through the workflow, economics, data, evaluation and operating constraints with the relevant people. Make missing evidence explicit.
Make the decision usable
Discuss the findings with the decision owner, record the trade-offs and hand over the memo and next-test criteria. Delivery is a separate decision.
A review can recommend buying an established product, using another provider, or not building. It is not an implementation contract or a guarantee of financial, technical or regulatory outcomes. Engagements remain subject to fit, availability and existing responsibilities.
An operating perspective
Experience behind the judgment.
Computacenter Romania
Managing Director
Leading the Romanian technology organization and its professional-services centre of excellence.
Public role sourceMETRO.digital Romania
Managing Director
Software and technology organization serving the wholesale business.
Public role sourceaccesa.eu
Executive leadership
Operations leadership followed by co-CEO and country managing-director responsibilities.
Public role sourceYonder
Managing Director
Software services leadership focused on Western European and North American software-product companies; earlier roles included engineering, project management, and technology leadership.
Public role sourceThese are roles, not endorsements or a claim that the organizations are advisory clients. See the sourced professional record and résumé, including the historical executive-board role and ongoing doctorate.
Methods you can examine
How I approach the problem.
The public research and engineering below show how I test assumptions and examine evidence. They are not client case studies or claims of business returns.
Research-informed AI judgment →
Understanding how data, model design, and evaluation shape the strengths and limitations of AI systems.
Engineering judgment →
Diagnosing complex system failures, making assumptions testable, and verifying that changes preserve intended behavior.
Evidence and accountability →
Making sources, uncertainty, and the limits of AI-assisted recommendations visible to the people responsible for a decision.
A useful first message
What are you deciding?
Send a short, non-confidential outline of your organization and role, the decision and its timing, what you have already tried, and the main constraint. We can establish whether a scoped review is appropriate before discussing access or an engagement.
Discuss the review on LinkedIn →Please do not send sensitive company data in the first message. This is a personal site, not an offer on behalf of my employers. Outside engagements require a fit, availability and conflict check; an enquiry does not imply acceptance.