Fractional AI Leadership

Senior technical judgement, a few days a month, for as long as it's useful. We set direction, grow the people you already have, and keep the roadmap honest when the hype cycle turns.

Most teams don't need a Head of AI yet

They need someone senior in the room for the handful of decisions that are expensive to get wrong, and nothing at all for the rest. Hiring full time for that is a large bet placed at the moment you have the least information, and the market rate reflects how many other people are placing it.

The alternative is to rent the judgement and keep the execution in house. We show up regularly enough to have real context, stay out of the way otherwise, and work steadily towards not being needed.

What that looks like in practice

  1. Own the technical direction

    Build or buy, which model, how much to spend, what to stop doing. Written down with the reasoning attached, so decisions can be revisited on evidence rather than relitigated on memory.

  2. Review the hard calls

    Architecture reviews, evaluation design, vendor contracts, and the occasional unpopular opinion that this feature will not survive contact with users.

  3. Grow your engineers

    Pairing, code review, and reading groups aimed at the people who will still be there long after we've gone. The goal is a team that stops needing us.

  4. Keep score in public

    A shared view of what was promised, what shipped and what it actually did. Uncomfortable at first, then quickly the most useful document in the company.

  5. Translate upward

    Board updates and investor questions answered in plain language, with the uncertainty stated rather than smoothed away.

in the room. not on the org chart.

Vericode · São Paulo · 2025–present

Vericode is a São Paulo IT company with two decades in quality assurance for large Brazilian companies. They had the domain and the customers to put AI into their products. They did not have that expertise in house.

We have been their fractional AI lead since the start of 2025: technical direction, growing the team's AI knowledge, and sitting with their engineers through design and build.

One of the products that came out of that is dott.ai, which simulates a stock exchange or payments environment and runs the simulation so new participants can certify against it. The simulator is Vericode's. The AI on top of it was designed and built with their team.

Want a second opinion before you hire?

Plenty of these conversations end with us saying you should hire someone permanent. That's a fine outcome too.