Fractional Head of AI: you used to have an IT guy, now you need an AI guy.
For decades you had someone who owned your IT infrastructure. AI is infrastructure now too, and it needs the same thing: one person who owns it.
A Fractional Head of AI is a part-time executive who owns the AI inside your company. I set up and run your AI infrastructure, the Claude brain, memory, Skills, and connectors, train your team to use it, and keep improving it every month. You get senior AI leadership without hiring a full-time executive.
The IT guy became infrastructure. So does the AI guy.
For decades, every company had an IT guy. Someone set up the computers, kept the printers running, managed the servers, and fixed it when it broke. You did not think of it as optional. It was infrastructure. AI is infrastructure now too, and it needs the same thing: one person who owns it, keeps it running, and makes it better. That is the AI guy. That is what I do.
What the AI guy actually runs
It is not hardware. Your AI infrastructure is four things, and I own all of them:
The brain.
A CLAUDE.md file that knows your business: your voice, your pricing, your process, your team.
The memory.
The context Claude carries across every conversation and every person, so nothing restarts.
The Skills.
The workflows your team runs every day, built once and reused, in your voice.
The connectors.
The links to your email, calendar, files, and the tools your work already lives in.
I set them up, I keep them running, I evolve them as your business changes, and I sharpen them as the models improve.
How your team actually learns it
Adoption is what makes or breaks a rollout, so it starts with the people, not the tool.
First, the executive Claude workshop.
I sit with the leadership team, show them a real, useful output in the first twenty minutes, and we set the rules: what AI is for here, what data goes in, what never does. The leaders leave bought in, because they used it, not because they saw a slide.
Then, department by department.
Sales, operations, finance, HR. Each team learns Claude on the work that actually eats their week, and I build the first Skills with them live. The cautious people become the guardrails experts, not the holdouts.
The full method, department by department, lives on the training and adoption page.
This is the real job
An IT guy did not set up your network and disappear. Neither do I. Every month I add new Skills as new workflows show up, keep the brain current as the business changes, and migrate the whole setup each time a new Claude model ships, so your team never thinks about versions. The system keeps getting better. That is the Plan de Evolución, and it is the point.
Fractional, not full-time
You do not need to hire a full-time Head of AI to get one. A full-time AI director runs about $200K to $300K a year, plus benefits, plus the months it takes to find one (the benchmark a mentor at a $14B advisory firm shared with me). I do the same job fractionally, I start the same week, and I cost a fraction of that. The engagement runs $6,000 to $10,000 per month. You get the senior AI leadership without the headcount. See full pricing.
The build is the start
Everything I build is yours: the brain, the Skills, the connectors, the account. The build is the start. After it, your AI guy keeps the whole thing evolving, and every month the system does more than the month before.
Five questions to ask before you hire anyone
Most fractional AI firms are advisory only. These questions separate the ones who build from the ones who advise.
Do they build, or just advise?
Most fractional AI firms deliver strategy and governance recommendations, then leave. If you need someone who will actually build the data layer, deploy the automations, and train the team, ask for proof: real examples, real systems, real results.
Who actually does the work?
Some firms say fractional but send a rotating team of three to five, or delegate to junior staff. Ask directly whether the person you meet will do the work, and how many people you will deal with. The answer tells you the engagement quality.
Can they show real mid-market implementations?
Enterprise case studies do not translate to mid-market. A 50-person company has different constraints than a Fortune 500. Ask for examples in your size range, your industry, and your budget.
What happens if you stop?
A good fractional AI leader builds systems you own: open infrastructure, no proprietary platform, documentation a different engineer can read. Ask whether another engineer could pick it up tomorrow. If the answer is no, it was built wrong.
What does governance look like?
AI without governance is shadow AI with a budget. Ask about the policy framework, shadow-AI monitoring, the approved-tool list, and compliance documentation. If the answer is vague, keep looking.
How the AI guy works in practice
Frequently Asked Questions
Your AI guy, part-time. I set up and run your company's AI infrastructure, the Claude brain, memory, Skills, and connectors, train your team, and keep it improving every month, without you hiring a full-time AI executive.
A fraction of a full-time AI director, who runs about $200K to $300K a year. I price by phase, with a starter engagement first, then a monthly Evolution Plan. You will know the number before we start. (Full pricing on /pricing.)
Most companies between $10M and $1B do not need a full-time AI executive yet, but they do need someone who owns AI. Fractional gives you the senior leadership without the salary and the year-long search.
The Claude brain (CLAUDE.md), the memory, the Skills, and the connectors, plus the governance and the migration to each new model.
Usually a few focused days a month, plus async availability with same-day response. The engagement is output-based, not time-based. You pay for governance, new capabilities, and results, not for hours logged.
No. The role is complementary. Your CTO owns the broader technology stack. I own the AI-specific layer: the data it reads, the automations, the governance policy, the team training, and AI vendor evaluation. In practice we collaborate. I design, your IT team implements what touches core infrastructure.
Any industry where operations run on scattered data, manual work eats hours, and AI adoption has stalled or never started. Professional services, construction, financial services, and distribution are where I have the deepest case studies. Mid-market companies of roughly 20 to 200 employees see the most value.
You own everything that was built: open infrastructure on your own accounts, full documentation, a clean handoff another engineer can pick up. The AI use policy, approved-tool list, and governance framework stay in place. The monthly Evolution Plan is the part that ends, not the system.
The engagement runs under an NDA and a data-handling agreement from day one, and the systems are built on infrastructure your company owns. Building your governance framework is a core deliverable: data classification, an approved-tool list, access controls, and shadow-AI monitoring.
A freelance developer builds what you spec. I figure out what needs to be built, build it, train your team, and stay to govern it. I do not wait for a ticket. I identify the problem, propose the fix, deliver it, then measure whether it worked and adjust.
No. Everything runs on your accounts, documented so another engineer can maintain it. Nothing depends on me to keep running. That is the design from day one, not an afterthought.
Book a 15-minute call
Tell me where your team is using AI today. I will tell you what your AI guy would set up first, and what it would cost.