Few bets
Three at most to start. Everything else goes on a cut list, with the reason.
Fractional Head of AI
Your Head of AI,
2 to 4 days a month.
I put AI to work inside your business. I come in, learn how the work gets done, pick the few bets worth making and stay until your team uses them every week and they show up in your numbers. Then I hand it back to your team.
What changes
Every bet gets
an owner, a number and a date.
I call it the Rule of One. A bet that cannot get all three goes on the cut list.
One owner per bet
A named person on your team who answers for the result. Never a committee.
One number, in money
Revenue added, cost removed, cash freed or hours that changed payroll, a hire or a sale. Agreed in writing before work starts.
One date
No test runs past 90 days without a verdict: kill, fix or scale. The date goes in the calendar on day one.
The engagement
Start small.
Stay as long as it pays.
It starts with a free 30-minute call, with no pitch: you leave with one decision. Then three steps, each with a defined scope and written deliverables you keep.
AI Decision Audit
- An inventory of every AI tool, subscription and experiment in the company, including what your team uses on its own
- A read on the data you already have
- A kill, fix or scale verdict on each one
- Three ranked bets and a cut list
- A 90-day plan with a name next to each item
- A memo you read in ten minutes
- A 60-minute working session with your team
Good fit when you pay for AI tools, people use them their own way and nobody can say what it all brings back. From you: your list of tools and subscriptions, short interviews with the people who use AI most and one hour of your time.
AI Implementation Sprint
- One or two bets put to work inside your operation
- An owner on your team for each one, named on day one
- A baseline and a number in money, agreed before the build
- Your team actually using it, with usage counted
- A one-page AI policy
- The AI Bill at days 30, 60 and 90
- A runbook your team can follow
Good fit when you know the bets and need them running in the daily work.
Fractional Head of AI
- A monthly review of the AI Bill
- Tool and vendor decisions
- The next bet, chosen with you
- Adoption with your team
- Direction of the build, with your team or Tecela's builders
- A named successor inside your company
- An exit plan
Good fit when AI needs an owner every month and a full-time hire is not the answer yet.
Start the Sprint within 30 days of the Audit and the Audit fee is credited toward it.
Independent advice
I advise. Tecela also builds.
So the rules are written down.
Without rules, that is a conflict. With them, you get an adviser who builds and still tells you to buy.
- Two agreementsAdvice and build are separate engagement letters. You can sign one without the other.
- Buy before buildMy default recommendation is to buy what already exists. Tecela builds what you cannot buy.
- A second option in writingEvery Tecela build proposal comes with an alternative: another vendor, your own team or a product on the market.
- You own what we makeCode, documentation and data are yours from the first day.
- Take the plan anywhereYou can put the plan out to bid and hand it to any builder, Tecela or not.
- Documents that work without meEach step ends in a document you can use with or without me.
Who it is for
For founders
who run the business.
Founders and owners of companies with 10 to 200 people in the United States and Europe, who decide on their own or with a partner. You pay for AI tools, your team uses them in different ways, and you want someone who comes in, learns the process and makes AI work inside it.
A Head of AI a few days a month, without a full-time hire. Remote and distributed teams are a natural fit: I work remote, inside your business hours. Your industry does not change the method: every business has work to speed up, people to own it and numbers to prove it.
I take two retainers at a time, plus one audit or sprint. When I give you a start date, it holds.
Signs you need one
If three of these sound familiar,
AI needs an owner.
- You pay for AI tools that nobody opens.
- Three people use ChatGPT in three different ways, with company data and no policy.
- You bought an automation from an agency and it stopped working.
- The team says AI makes them faster, and nobody can show where that went in the numbers.
- A demo looked great six months ago. Nobody uses it today.
- You do not know where to start or which tools are worth paying for.
- Your developer or your agency runs the tools, and nobody decides where AI should make money.
- You have thought about hiring an AI person full-time and are not sure there is a full-time job yet.
How we measure
One number per bet.
Written down first.
If a bet cannot be measured in money, it does not make the list.
Baseline before the build
I record where the number stands today, so the improvement is a difference and not an opinion.
The owner signs the number
The person who answers for the result agrees the number and the date.
The Cash Test
A saved hour only counts when it changed payroll, a hire or a sale, and someone outside the team confirms it. Until then, it is a good story.
Adoption, measured
Who uses it and how often, against the seats you pay for. Usage is a number too.
The AI Bill, every month
One page: what you pay for AI and what came back, by bet, with an owner on every line.
A cut list that gets reviewed
What we stopped and why. Cutting is part of the result.
Questions
What founders ask
before they start.
My developer or my agency already handles AI.
They own the tools, and they should keep them. Someone also has to own where AI makes money for the business and when to stop paying for it. That is the job I take.
Isn't this just another consultant?
A consultant hands you a report and leaves. I go inside the work with your team, under a written mandate: a named owner for each bet, a decision log every month and an exit plan with a successor inside your company.
You don't know our business.
The Audit is built for that: it starts by reading your tools, your spend and your data before it recommends anything. I spent 15+ years inside multinationals, 7 of them in finance at Apple, so I look at the numbers before I look at the model.
What happens when you leave?
The handover is in the contract: runbooks, a trained owner and a named successor by month 3. I plan to be needed less every quarter. If you still need me after twelve months, we should be hiring someone full-time.
How do we contract?
With FR Consulting LLC, a US company: contract in English, invoice in USD, paid by ACH or card. Every deliverable in writing.
You also sell software. Isn't that a conflict?
It would be without rules. Advice and build are separate agreements, buying comes before building, any Tecela build comes with a second option in writing, you own everything we make, and you can take the plan to any builder.
We're too small for a Head of AI.
You do not need one on payroll. You need someone accountable for the AI result a few days a month, and a 10-person company has an AI bill too. The Audit is the smallest way to test it.
How do I know it will work?
You do not bet on the engagement. The Audit is two weeks, fixed scope, and ends in a decision memo you can use with or without me.
AI moves too fast to commit for months.
That is why step one is two weeks and the retainer starts at three months, then runs month to month. The decision log keeps what you learned.
Why not a big consulting firm?
One person from the first call to the work running in your operation, who runs what he recommends and builds what is needed. No pyramid between you and the work.
What about our data and confidentiality?
An NDA and a data processing agreement come before day one. I get access only to the accounts you scope, and your data stays in your environment.
Why not wait a year?
Your people already use AI with or without a policy. Waiting does not pause the spend or the risk; it only leaves both without an owner.
How much does it cost?
Fees are agreed in writing after the first conversation, before any work starts. The Audit has a fixed scope and a fixed fee, so you know what you get for what you pay.
A lighter first step
Not ready to talk?
Read Running on AI.
One AI decision a week from a founder who runs his own companies on AI agents, in plain words and real numbers. It is the easiest way to see how I think.
Follow Henrique on LinkedInNext step
Which bet should have
an owner first?
Tell me what is running today and what it is meant to change. I will tell you if an audit is the right first step.