AI consulting
AI consulting that starts with your data.
Most AI projects don't fail on the model. They fail on the data underneath it. I help mid-market teams get AI-ready first, then put AI to work where it actually pays off.
Definition
What is AI consulting?
AI consulting is independent, senior help deciding where artificial intelligence fits your business, getting your data and processes ready for it, and putting it into production without the hype or the false starts. Done well, it has less to do with the model and more to do with everything around it: the data the model reads, the guardrails it runs under, and the people who have to trust its answers.
For most mid-market companies the honest first step isn't a copilot — it's the foundation underneath one. That's the work I lead, and it's why this practice starts with your data, not with a demo.
Why data first
Before you plug in AI, fix the foundation.
A model trained on everyone else's internet is a commodity. Your advantage is your data — and if it's duplicated, undocumented, and nobody trusts it, an LLM will make that worse with total confidence. I do the inverse: get the foundation trustworthy, then put AI on top of it.
Skip the foundation
- Three dashboards, three different numbers for the same KPI.
- No lineage, so nobody can say where a number came from.
- An LLM confidently hallucinates on a metric no one ever defined.
- Adoption stalls the first time the answer is visibly wrong.
Build it first
- One definition per metric, one source of truth.
- A catalog and lineage an LLM can actually query.
- Governance that decides what AI is allowed to see.
- The team trusts the answers, so they keep using them.
What it includes
What AI consulting looks like here
Four kinds of work, sequenced to your situation. Most teams start with readiness and adoption; the rest follows from what the diagnostic finds.
01
AI readiness assessment
A clear-eyed read of your data, tooling and processes against what AI actually needs. You get a map of what's ready, what's missing, and what to fix before you spend a cent on models.
02
Data foundation & governance
The unglamorous part that decides everything else: catalog, lineage, data contracts, and the access rules that keep an LLM away from data it shouldn't touch.
03
AI training & adoption
Hands-on tracks that get your team using AI day-to-day with judgment instead of hype — prompts, workflows and guardrails tied to real tasks, not a demo reel.
04
Use-case selection & rollout
Picking the two or three places AI pays off first, building the narrow version that works, and shipping it without betting the company on one moonshot.
Who it's for
Built for mid-market teams, not the Fortune 500.
The big firms sell AI-strategy decks to companies with a hundred-person data team to execute them. I work with the companies in between: 50 to 500 people, a small data team or none, and real pressure to do something with AI this year.
- Mid-market companies (50–500 people) without a large data team.
- Leaders who were asked for an “AI strategy” and suspect the data isn't ready.
- Teams burned by a pilot that never made it to production.
- Brazilian and US companies — remote, or in person in São Paulo and Belo Horizonte.
How we work
Four steps, no consulting theater.
01
Diagnose
A readiness assessment: what data you have, what AI needs, and the gap between the two.
02
Prioritize
We pick the few use cases worth doing first and the foundation work they depend on.
03
Build
I build alongside your team — the foundation first, then the narrow AI that earns its place.
04
Handoff & train
Documentation, guardrails and training, so your team runs it without me.
FAQ
Questions I get a lot
Keep reading
More on AI and data
Next step
Not sure if it's an AI problem or a data problem?
That's exactly what the first call is for. 30 minutes, no commitment — you tell me where you are, and I'll tell you honestly whether AI is the right next move or whether the foundation comes first.