Skip to content
ClelandCo

Vol. III — Fractional CAIO

Currently booking

The accountable owner of your AI function, one to two days a week.

Someone has to own it. One to two days a week.

An embedded Chief AI Officer who owns the roadmap and its results, sits in leadership meetings, makes the vendor and build-versus-buy calls, and implements where a handoff would kill the work. Executive ownership at a fraction of the seat.

Book a call about the fractional seat (opens in a new tab)

Briefing

A full-time Chief AI Officer is an executive hire: executive compensation, a search measured in quarters, and a job description most companies are still drafting. Plenty of organizations need that level of judgment long before they can justify or fill that role, and the gap is where pilots go to stall.

The failure is rarely technical. It is that no single person is answerable. The CTO adds AI to a full plate. A vendor sells a tool nobody owns the strategy for. A consultant delivers a roadmap and leaves. Six months later the license is shelfware, the board asks a question that gets answered the night before, and nobody can say which of the three funded pilots is worth continuing.

This is a standing seat, one to two days a week, with a mandate and a budget line. I own which use cases get funded and which get killed, what evaluation and governance look like, which vendors you buy from and which capabilities you build, and what the executive team hears about all of it. Anything with legal, financial, or headcount consequences comes to you with a recommendation and the reasoning attached.

I also build. Decision-only AI leadership produces good decisions that queue behind engineering capacity which never frees up, so implementation is in scope for the pieces where a handoff would kill the work — deployment, evals, monitoring, latency, and cost.

This is not consulting and it is not coaching. It is an executive function, held part-time by someone who has done the work.

Three ways forward

Three options. One recommendation.

01

Hire a full-time Chief AI Officer.

Executive compensation, benefits, and a search that runs for quarters, for a mandate most mid-market companies cannot yet fill a week with. Right eventually, expensive now.

02

Give it to the CTO and hope.

Your CTO owns infrastructure, uptime, and a roadmap that was already full. AI strategy becomes the thing that slips, and it slips invisibly until a board meeting makes it visible.

03

Rent the seat, not the salary.

Senior ownership one to two days a week, with decision rights, a budget line, and a hiring plan for the person who eventually replaces the arrangement.

The engagement ladder

Two rungs. One clear path.

Almost every engagement starts with the assessment. Signing a retainer before anyone has scoped the problem is how you end up with a vague retainer.

AI Adoption Assessment

The front door

$6,000

One time, 1 to 2 weeks

A written read on where AI is worth spending, what has to be true before anything ships, and what to do first. The same assessment starts both the advisory and fractional tiers; it ends with a straight recommendation about which, if either, is worth it — including when the answer is neither.

Fractional CAIO

The engagement

$10,000–$15,000/mo

50–70 hrs/mo · Six-month initial term, then month to month

An embedded Chief AI Officer who owns the roadmap and its results, sits in leadership meetings, makes the vendor and build-versus-buy calls, and implements the pieces where a handoff would kill the work.

Book a call about the fractional seat (opens in a new tab)

AI Adoption Assessment

$6,000

One-time

Scope the problem before anyone signs a standing seat.

Format

  • Interviews with the people building and the people funding, held separately, because they rarely describe the same roadmap
  • A use-case inventory scored by value and feasibility, with an explicit list of what not to do — including funded pilots that should be killed
  • A readiness and governance gap assessment covering data handling, evaluation, deployment, and monitoring, written so you could hand it to a customer or an auditor
  • Build-versus-buy guidance per use case with the criteria written down, plus a review of the vendor commitments already in flight
  • A prioritized roadmap and a straight recommendation on whether ongoing work is justified, including when the answer is no
Book a call about the fractional seat (opens in a new tab)

Usually the AI Adoption Assessment, $6,000 one time. Starting with a retainer before anyone has scoped the problem produces a vague retainer.

This is right for you if…

The board has started asking about AI and the answers are assembled the night before. You have vendor spend growing and no measurement of what it returns. You are considering a standing seat and would rather scope the mandate than invent it in month two. This is the same assessment the advisory tier starts with; here it ends by saying whether a standing seat is justified yet.

Fractional CAIO

$10,000–$15,000/mo

50–70 hrs/mo

A seat in leadership, decision rights, and one person answerable for the results.

Format

  • Ownership of the AI roadmap against outcomes agreed in writing before anyone starts building.
  • A seat in leadership meetings and a standing monthly report to the exec team or board.
  • Governance, evaluation, and risk controls you can show a customer, an auditor, or an acquirer.
  • The AI Contingency Framework run end to end, with rehearsal — red teaming, failure injection, misuse simulation — as a standing gate before anything ships.
  • Vendor selection, contract review, and build-versus-buy decisions made, documented, and defended.
  • Hands-on implementation where a handoff would kill the work: deployment, evals, monitoring, latency, and cost.
  • A hiring plan and interview support for the permanent team, including the person who eventually replaces me.
Book a call about the fractional seat (opens in a new tab)

One to two days a week on your calendar, attendance at leadership meetings, and a written monthly report to the exec team or board.

This is right for you if…

Pilots keep stalling after the demo and no single person answers for it. You need someone who can tell a board what is true, kill the work that is not paying, and still be technical enough that the roadmap survives a conversation with your engineers.

Fit

Worth knowing before you call.

Signals this is your tier

Mid-market companies, Series A to C startups, and PE portfolio companies with real pressure to adopt AI, nobody senior accountable for it, and not enough scope to justify a full-time executive.

  • Pilots keep stalling after the demo and no single person answers for it.
  • The board is asking about AI and the answers are assembled the night before.
  • Vendor spend is growing and nobody is measuring what it returns.
  • A customer, auditor, or acquirer has started asking how the AI systems are governed.

Where this is the wrong call

  • Organizations that want a reporter rather than a decider. That is the advisory tier and it costs less.
  • Anyone who needs the title but will not grant the authority. Without a seat in leadership meetings and a budget line, the role cannot own anything.
  • Teams who want one project delivered. Scope that as an implementation engagement instead of a standing seat.
  • Companies that cannot name a business outcome for AI. Start with the assessment; it may tell you to wait.

What happens first

The opening phase, step by step.

One person is answerable for whether AI work produces business results: which use cases get funded, what ships, what gets killed, and what the board hears.

  • Days 1–30

    Assessment and roadmap.

    A full audit of current AI activity: what is working, what is burning budget, where the highest-return opportunities are, and what has to be true before anything ships. Ends in a prioritized roadmap with outcomes agreed in writing, not a list of features.

  • Days 31–60

    Governance, vendors, and the first build.

    Evaluation and risk controls you could show a customer, an auditor, or an acquirer. Vendor contracts reviewed and renegotiated where they deserve it. The first implementation moves, chosen because it is provable inside the window rather than because it demos well.

  • Days 61–90

    Proof, and the report that says so.

    The first outcome is measured against the definition of done agreed in writing at the end of the first thirty days. The board or executive team gets a written report with the numbers and the honest read, including which funded pilots should be killed.

  • Ongoing

    Operate, and plan your own exit.

    Monthly reporting, continuous vendor and roadmap ownership, and a hiring plan for the permanent team — including the interview loop for the person who takes the seat. The handover artifacts exist from month three, whether or not you extend.

Against the alternatives

Three tiers. One ladder.

DimensionAI Visibility LeadAI Strategic AdvisorFractional CAIO
Price$1,500–$2,500/mo$5,000–$8,000/mo$10,000–$15,000/mo
Time per month10–15 hrs/mo20–30 hrs/mo50–70 hrs/mo
Starts withAI Visibility Baseline, $1,000AI Adoption Assessment, $6,000AI Adoption Assessment, $6,000
Who owns the outcomeYou own the outcome. I measure it and close the gaps.You own the outcome. I pressure-test the plan.I own the outcome and report on it.
Decision rightsNone. I recommend and execute inside an agreed scope.Advisory. Your leadership decides.Full, inside an agreed mandate and budget.
Hands on keyboardYes. Profile, schema, and citation work is included.Review only. No implementation in this tier.Yes, on the pieces where a handoff would kill the work.
Best forGetting found and cited correctly by AI search.A capable team that needs senior judgment.An AI function with nobody accountable for it.

The operator

Who you're working with.

I am Jeremy Cleland. Fourteen years in uniform, eleven of them in Special Operations as an 18D Special Forces medic, then a master's in Artificial Intelligence in 2025 and a second career building systems that have to work when someone is depending on them.

Special Operations medicine is a useful background for this work for an unglamorous reason: the job is making decisions with incomplete information, under a deadline, where being confidently wrong is the failure mode that costs the most. That is also the job of running an AI function in 2026.

I build as well as advise. The roadmaps I write are constrained by what I know I would have to implement, which tends to keep them honest about what is actually shippable this quarter.

Most AI advisors are strategists who have never shipped a model, or engineers who have never sat in a leadership meeting. The useful part is doing both.

Read the longer version

What this engagement does not include.

Full-time availability. The retainer is sized in hours per month. One to two days a week is a real commitment and it is not five; work that needs a full-time executive needs a full-time executive, and I will say so.

A title without the authority. Without a seat in leadership meetings and a budget line, the role cannot own anything and becomes an expensive reporter. If the mandate is not there, the advisory tier is the honest fit.

Net-new research. This is applied work on your systems and your data. Original research, unrelated platform rebuilds, and speculative model training are quoted separately rather than absorbed quietly into the retainer.

Permanence. The intended ending is that you hire a full-time CAIO and I help you interview them. An engagement that has no exit condition is a subscription, and you should be suspicious of anyone selling you one.

Questions

Asked and answered.

How many days a week is this, and what does it cost?
One to two days a week — 50 to 70 hours a month, which is the figure the agreement is written against. $10,000 to $15,000 per month, sized to where in that range the work lands. Six-month initial term, then month to month. The rate covers decisions and implementation from the same person, not a decision-maker who then bills separately to build.
What decisions do you actually own?
Which use cases get funded and which get killed, what evaluation and governance look like, which vendors you buy from and which capabilities you build, and what the exec team hears about all of it. Anything with legal, financial, or headcount consequences goes to you with a recommendation and the reasoning attached.
Do you build, or only decide?
Both, from the same person. Decision-only AI leadership produces good decisions that queue behind engineering capacity that never frees up. A roadmap written by someone who has to implement it tends to be honest about what is actually buildable.
How is this different from hiring an AI consultant?
A consultant delivers a deliverable and leaves. The deliverable is often good; the problem is that nobody owns what happens next, so the roadmap goes stale and the pilots stall. This is a standing seat with responsibility for outcomes, which is the part project-based consulting structurally cannot provide.
What happens when we hire a full-time CAIO?
That is the intended ending, and the engagement is built for it. The hiring plan, the interview loop, and the first 90 days of the role are part of the work. Handover is documentation, runbooks, the roadmap, and the open decisions, delivered whether or not you extend.
What happens when the AI fails?
That question is the method, not an afterthought. All five phases of the AI Contingency Framework run in this seat: mission analysis, intelligence preparation, courses of action, rehearsal, and sustainment. Rehearsal becomes a gate every system clears before it carries real work, and what it finds is reported to the exec team alongside what shipped.
What is out of scope?
Net-new research, unrelated platform rebuilds, and full-time availability. Retainers are sized in hours per month rather than vague access, and implementation is scoped to a named system with written acceptance criteria. Anything outside that is quoted separately rather than absorbed quietly.

Ninety days is long enough to prove it.

Start with the assessment. It will tell you what the AI function is worth to you, what it would take to run it properly, and whether a standing seat is justified — including when the answer is not yet.

Book a call about the fractional seat (opens in a new tab)