Fractional CTO

You are looking for a fractional CTO, and the seat you need is almost certainly narrower than that title. "CTO" is the phrase people reach for when the real complaint is that nobody senior owns the technology decisions — but it bundles three different jobs that separate cleanly, and an engagement scoped to the whole title inherits all three and owns none of them well. Facet fills three fractional executive seats rather than four, and this page is here to tell you plainly which one you are describing, what it costs, and how to test whoever you hire — us included. We can be that specific because it is a decision we had to make about our own company before we ever sold it.

Why we sell three seats and not a CTO

Facet runs on Agent Loops. Our engineering, content, research, quality review and reporting all move through agents under human command, and we published the full account of how that works.

That had a consequence we did not anticipate, and it is the reason this page exists. When the recurring, checkable parts of executive work moved onto loops — the vendor and renewal register, the drift and threshold monitoring, the chasing of decisions to evidence, the reporting assembled against a baseline — what was left behind was only the judgment. And the judgment did not come back as one job. It came back as three, with almost nothing in common:

Somebody has to own the systems the company runs on — the portfolio, the vendors, the spend, the risk register. That is a Chief Information Officer's work, whatever the org chart calls it.
Somebody has to decide where AI goes first, second and never — and, harder, what gets shut down when it does not pay. That is an allocation job with a technical subject.
Somebody has to be answerable for whether the numbers are trustworthy — because every automation you build inherits whatever quality your data already has.

A single seat called "CTO" can hold all three on a good week and none of them on a busy one, which is the failure mode we were trying to design out. So we sell them separately. If you are here because of AI — a roadmap, build-versus-buy, sequencing, governance — the seat you are describing is the fractional Chief AI Officer, and everything below is about that seat wearing the title you searched for.

Humans stay in command: Agents carry the recurring half of the seat — the registers, the monitoring, the evidence behind a decision — and they stop at the decision itself. Your executive owns every decision, and a named Facet engineer owns every change that reaches a production system.
We name the failure modes: Hallucination, scope creep, runaway token cost and lost context are why an AI portfolio needs stopping rules at all. A seat that will not name them is selling you optimism.
The delivery sits behind the seat: When a decision needs a build, Facet's Agentic Engineering team builds it and AI Managed Services keeps it running. Strategy and implementation under one accountable relationship, which is what stops a good decision from queueing behind a team with no capacity to act on it.

Why hire a fractional CTO at all

Because the decisions in front of you are senior decisions, and getting them wrong is expensive in ways that do not surface for months. A fractional arrangement gives you that seniority now, aimed at a defined set of calls, without the search, the equity or the permanence.

You are buying decisions, not a desk: Strip back the calendar and an executive seat concentrates into a handful of calls a year that move real money. Which platform becomes your system of record. Whether to rebuild the thing that keeps breaking. Which process gets automated first.
Seniority does not scale by hours: Someone who has made the same call in nine other companies makes it faster and better than someone making it for the first time. You are paying for pattern recognition, not attendance.
You can start long before the org chart justifies a hire: Companies between roughly 25 and 500 people need executive-grade judgment years before a full-time officer is defensible.
And this is the part worth being cheerful about: the decisions you have been deferring are not waiting on budget, headcount or a transformation programme. They are waiting on somebody with the standing to make them. That is available to you in week one, and closing the biggest one is what lets the next ten be made properly instead of provisionally.

First step, and it costs you nothing: list the technology decisions you have deferred for more than a quarter. If they are mostly about systems that keep breaking, you want a Chief Information Officer. If they are mostly about where AI goes, you want a Chief AI Officer. If they are mostly about numbers two teams disagree on, you want a Chief Data Officer. That list is the job description, and it is usually unambiguous once written down.

What the seat actually covers

What the loop actually does

The historical weakness of part-time leadership was continuity: the executive is in the room two days a month and the organization moves at the speed of those two days. These passes are what changes on the other twenty-eight, and they go to agents for the usual reason — continuous, checkable, and the first thing to lapse when a seat is part-time.

The system, contract and renewal register, kept current: Every tool, owner and renewal date tracked continuously rather than reconstructed the week before a review. It is the register that decays fastest when nobody owns it, and the one every spend decision rests on.
Threshold and drift monitoring: Spend, uptime, exception rates and data freshness watched against the baseline set in the first thirty days, so a slip pages a person the day it happens rather than appearing in a quarterly pack.
Decision follow-through, chased to evidence: Every commitment made in a session recorded, assigned, and tracked until there is proof it shipped. It fails for a mechanical reason — nobody was chasing it — which is precisely why a loop is the right fix.
First-draft documentation: Policies, runbooks and vendor summaries drafted from your live systems, so your executive edits a document rather than starting one.
The opening inventory: What AI is already running inside your company, sanctioned or not, and what each instance touches. Agents read the systems; this is consistently the most surprising deliverable of the first month.

What a person still decides

None of the above is the seat. The seat is a small number of calls that carry consequence, and what they share is that no amount of data settles them. For a company with no senior technology leader in the building, these five are the ones that have been going unmade.

Whether to keep building on what you already bought. The platform you chose three years ago may be the reason delivery is slow. Saying so out loud costs somebody internal their credibility, which is why it usually goes unsaid — and it is frequently the most valuable sentence of an engagement.
What to tell the agency or developer who built your current system. The renewal figures can show the relationship is poor value. Whether to end it turns on what moving would cost you, how much of how your business actually works lives only in their heads, and whether you can absorb that disruption this quarter. No report settles it.
What an AI system may do with customer data, unattended. Where the human signature is required, and who owns an exception now carrying real risk. Most companies are answering this by default rather than deliberately, which is a decision too — just not one anybody made.
Which tradeoff your board or investors hear. Security against speed, spend against capability, build against buy. Somebody has to choose which risk the company carries, argue it in front of people who can overrule them, and own it afterwards.
Whether you need a seat at all. Sometimes the honest answer is that what you need is one process in production and a build, not an executive governing a portfolio of one. We would rather say that than take the retainer.

The math on fractional versus full-time

Everyone compares the monthly fee against a salary. That is the least important number in this decision, and it is also the only one you can look up. Run the version below on your own figures — it is illustrative, the point is the shape of the arithmetic, and it takes about twenty minutes.

Two things fall out of that table almost regardless of what you put in it.

The loaded cost sits materially above the salary everyone quotes: Which means the fee comparison people lead with is being made against the wrong figure.
The months are the real expense: A full-time search puts your first genuine decision quarters away. A fractional engagement is deciding in week one, and it ends with an email rather than a severance conversation.

Which seat do you actually need, and what should a loop be carrying instead of a person? Book a Fractional Fit Review: two weeks, fixed fee, one working session with your leadership team. You leave with a written assessment of the decisions currently going unmade and what each is costing you, plus a one-page recommendation naming the seat that fits — including "none of them, hire full-time" or "none of them, you need one process built" when that is the honest answer.

Fractional CTO FAQ

What does a fractional CTO actually do?

A fractional CTO is a senior technology leader engaged part-time, on an ongoing basis, who owns a defined set of decisions and is accountable for them over time.

They own decisions, not a function's headcount: Architecture, platform, vendor and sequencing calls, made and defended.
They set the guardrails: As AI takes on real work, somebody has to own access, approvals and risk, matched to your compliance reality.
They translate for the business: Between what the technology can do and what the company needs, for your board, your team and your customers.
They are ongoing, which separates it from consulting: A consultant answers a question and leaves with the answer still on paper. A fractional executive holds the seat.
At Facet, the work is scoped to one of three seats: Chief Information Officer, Chief AI Officer or Chief Data Officer — named for the decisions rather than the title, so it can be held to them.

Is a fractional CTO the same as a fractional Chief AI Officer?

For most people arriving on this page, yes in substance and no in scope — and the distinction is worth two minutes because it changes what you can hold the seat to.

The overlap is nearly total when AI is the reason you are hiring: Roadmap ownership, build-versus-buy, sequencing and governance are the Chief AI Officer's job description.
The difference is what else the title implies: "CTO" also implies engineering management, systems ownership and product architecture. If you need those, you want a different seat or a full-time hire.
Naming it correctly is not pedantry: A seat scoped to a title gets measured on presence. A seat scoped to decisions gets measured on whether the decisions were made.
We have written about the timing question separately: when to hire a fractional CTO for AI covers the signals that mean now rather than next year.

When is a fractional CTO the wrong choice?

When the function is how you compete, when the job is mostly managing people day to day, or when you are mid-crisis.

Core functions belong inside: If software is the product you sell, that leadership should be permanent and invested.
People-heavy mandates need daily presence: Designing an org is fractional work. Being a team's manager for two years is not.
Crises need responders first: An active breach or a failing migration with a contractual deadline needs incident response this week, not a governance layer. We will say so rather than take the engagement.
And if a technology leader already has the mandate: What they need is budget and air cover, not a second executive. (More on where fractional advisory does and does not fit, particularly for earlier-stage companies.)

How does pricing work, and how many hours do we get?

A fixed monthly fee quoted before the engagement starts, never hourly, with a cadence set to your decision rhythm rather than an hour count.

We do not sell hours, because hours are the wrong unit: Buying a calendar rewards presence; buying an outcome rewards the decision getting made and the follow-through landing.
The fee changes only when scope changes and you have agreed it.
Build work is quoted separately: If a decision requires a migration or an automation, that implementation is scoped on its own so you can weigh it on its merits.
There is a defined exit: Engagements end when you hire the full-time seat, when the capability transfers, or when the decisions are made. You keep the documentation and the systems either way.

The question to ask every candidate, including us

There is a test for this purchase that takes one question, and you can run it on every fractional CTO you speak to this month.

Ask which decisions they will own and which they will not.

Somebody selling you a seat answers in a sentence, and the sentence has edges: I own the roadmap, the sequencing and the stopping decisions; I do not own your data quality or your vendor contracts, and here is who should. Somebody selling you availability cannot answer it at all, because the whole value of an undefined scope — to them — is that it never has to be declined. That is why the most expensive error available here is not picking the wrong person. It is buying a title instead of a boundary, and it is an error you make before any money changes hands.

Our answer, for the record: the seat this page describes is what we call the fractional Chief AI Officer. It owns the ranked portfolio of which processes become agentic and in what order, the policy for what your staff and your agents may do with company data, the adoption work that decides whether any of it gets used, and the stopping rules that let a pilot be cancelled without it becoming a referendum on somebody's reputation. It does not own your systems portfolio — that is the Chief Information Officer seat — and it does not own whether your numbers can be trusted, which is the Chief Data Officer's, and which usually has to come first.

Book a Fractional Fit Review — two weeks, fixed fee, one working session with your leadership team. You leave with a written assessment of the decisions going unmade, what each has cost so far, and a one-page recommendation naming the seat that fits the next four quarters. It comes back as none of them often enough that we say so on the page. You get the written answer whether or not you hire us, and the decisions it names are ones you can start making this quarter. Or talk it through with us first, no obligation.

Facet has run technology transformation for companies since 2014 and was named a CIOReview Top 10 Digital Transformation Services Company in 2021. The fee is fixed and quoted before we start; we do not bill by the hour.