Run · $2,495/month
Who Actually Owns AI Decisions in My Business?
Right now, nobody. Which means everybody. Which means your most enthusiastic employee, and you've never asked what they're pasting in.
The Site Rescue starts the moment your order is placed. I’ll email you within one business day, usually much sooner, to get access and book the 30-minute intake call. If the site is under attack right now, call (754) 302-4631 after you pay.
In most small businesses, nobody owns AI decisions — which means everyone does, and the most enthusiastic employee's judgment becomes de facto policy. A fractional AI officer is someone senior who owns AI for your business a few days a month: what tools are approved, what data can go where, what gets automated, what stays human, and what's actually working. sosSTEVIE does this for $2,495 a month, for businesses too small for a full-time AI hire but too exposed to keep improvising.
Key takeaways
- You don't have no AI strategy. You have as many as you have employees.
- The risk isn't your staff being reckless. It's that nobody's ever told them what's allowed.
- Shadow AI is already load-bearing in most businesses — critical work running in personal accounts.
- A policy document isn't governance. Nobody reads it and it doesn't decide anything.
- The people using nothing are as much of a problem as the ones using everything.
Run · $2,495 a month
Who Actually AI Decisions in My Business?
Right now, nobody. Which means everybody. Which means your most enthusiastic employee, and you’ve never asked what they’re pasting in.
- Can I use this new tool?unassigned
- Is this data okay to paste in?unassigned
- Should we automate this, or is it too risky?unassigned
- The thing Dave built is now critical. Who maintains it?unassigned
- Is our competitor doing something we’re not?unassigned
- Is any of this working?unassigned
Decisions need an owner, not a document.
The direct answer
In most small businesses, nobody owns AI decisions, which means everyone does, and the most enthusiastic employee’s judgment becomes de facto policy. A fractional AI officer is someone senior who owns AI for your business a few days a month: what tools are approved, what data can go where, what gets automated, what stays human, and what’s actually working. sosSTEVIE does this for $2,495 a month, for businesses too small for a full-time AI hire but too exposed to keep improvising.
Key takeaways
- You don’t have no AI strategy. You have as many as you have employees.
- The risk isn’t your staff being reckless. It’s that nobody’s ever told them what’s allowed.
- Shadow AI is already load-bearing in most businesses: critical work running in personal accounts.
- A policy document isn’t governance. Nobody reads it and it doesn’t decide anything.
- The people using nothing are as much of a problem as the ones using everything.
The story
I asked a managing director how his company used AI. He said they didn’t, really.
So I asked his eleven staff. All eleven used it. Daily. Six different tools, paid for personally on expenses under vague descriptions, none approved, none reviewed.
One had built a genuinely brilliant thing that produced their weekly client reports. It had quietly become load-bearing: a workflow living in one person’s personal account, documented nowhere. Another was pasting client contracts into a free tool to summarize them. Depending on who you asked, a data breach in progress.
And two people were using nothing at all, doing everything the slow way, because nobody had ever told them it was allowed.
Not no AI strategy. Eleven, all in conflict, none written down.
The MD wasn’t negligent. That’s not an AI problem. It’s a governance vacuum, and in a vacuum the loudest, keenest person’s judgment becomes company policy by default.
The vacuum
Every business has the same three groups. Nobody has mapped them.
The enthusiasts
Fast, inventive, genuinely valuable, and building things nobody’s reviewed, in personal accounts, with company data. Your best asset with no guardrails. When they leave, whatever they built leaves with them.
The quiet majority
Using it a bit, unsure what’s allowed, so they don’t mention it. You can’t see it. This is where the data exposure lives, and it’s the rational behavior of people who were never told the rules.
The refusers
Doing it the slow way, sometimes on principle, more often because nobody said it was permitted. The group everyone ignores and usually the biggest cost: a productivity gap you pay for every day.
of people using AI at work bring their own tools.
Three groups, no coordination, and the enthusiast’s judgment has become your company policy without a decision ever being made.
Note what the figure says: 78% of AI users, not of all employees, and it’s enterprise-weighted. But your business isn’t more governed than a Fortune 500 with a compliance department. It’s less.
Source: Microsoft and LinkedIn, Work Trend Index. Outside any IT approval.Why a policy document won’t fix it
Governance isn’t a document. It’s someone answering the question when it arrives.
The instinct is to write a policy. It feels like control. It goes in a folder. Nobody reads it.
A policy answers “what are the rules.” The questions arriving every week are decisions, and decisions need an owner. A policy is a snapshot of a judgment made once, in a domain where the ground moves every quarter. Written in January, it’s wrong by April, and still in the folder being wrong with great authority.
The risk and financial math
$2,495 a month deserves arithmetic, not adjectives.
About 15% of the alternative, for the decisions rather than the seat. For a business of ten to fifty people, the full-time hire is obviously absurd.
The invisible dependency
Harder to price, and often the one that actually happens. Somebody’s personal account runs your weekly client reporting. They leave. Nobody can reconstruct it, and the client notices before you do. Not a breach. Just a business that quietly stopped being able to do something it could do last month.
The gap at the bottom
Your refusers and quiet majority, working at a fraction of the speed available to them, every day, because nobody told them what was allowed. It never appears in a risk register and it’s usually the largest number on this page.
Set those four against $30,000 a year. The math isn’t close. It’s about the fact that nobody is currently deciding any of this.
What a fractional AI officer actually does
Not a strategy deck. A standing job, a few days a month.
Owns the decisions
Tools approved and paid for properly. Data rules specific enough to follow: not “don’t put confidential data in AI” but “client contracts go in this tool and nowhere else.” And a named person to ask, which is most of the value.
Finds the shadow AI
Maps what’s actually running. Turns the good stuff, Dave’s report workflow, into a documented, owned company asset. Shuts down the genuinely dangerous. Usually the biggest single win in month one, and it’s mostly just asking people.
Closes the gap at the bottom
Gets the refusers and the quiet majority to a competent baseline. Unglamorous. Nearly always the largest measurable return, because zero to competent is a much bigger distance than good to excellent.
Decides what stays human
There are things you shouldn’t automate, and someone needs the standing to say so and be listened to. Harder from inside than out.
Watches the outside
Models change quarterly. Most of it is noise. Reading the noise so you don’t have to, and telling you the two things a year that actually matter to you.
Reports honestly
Including “this isn’t working, stop paying for it.” An internal hire who championed a tool can’t easily say that. I can, and I will, and that’s most of what you’re buying.
Why fractional
Someone senior, in the room, with no incentive to sell you anything.
You can’t justify a full-time AI hire. The salary is absurd for the volume of decisions, and the role is too new to recruit for. You’d be interviewing for a job you can’t yet describe.
But the decisions arrive whether or not someone owns them, so they get made badly, by default, by whoever is most confident that week. Fractional fits the shape of the work: a few days a month, no empire-building, no budget to defend.
And I can tell you to stop.
I’m not protecting a department or justifying a headcount. When the answer is “you don’t need this, cancel it,” I’ve got no reason not to say so.
Fractional AI Officer
A standing role, not a project.
- Month one: mapping what’s actually running
- Owns the decisions, finds the shadow AI, closes the gap at the bottom
- Decides what stays human, watches the outside, reports honestly
- Roughly $30,000 a year, about 15% of a full-time senior hire
Six-month minimum, because anything shorter is a consulting project wearing a retainer’s clothes. The value compounds after month one.
Or call (754) 302-4631.
Somebody has to be in charge
Nobody was reckless. Nobody was in charge. Those turn out to be .
The MD didn’t have no AI strategy. He had eleven, in conflict, none written down, and one of them was pasting client contracts into a free tool.
Frequently asked questions
Before you decide.
What is a fractional AI officer?
Someone senior who owns AI decisions for your business a few days a month rather than full-time. They decide which tools are approved, what data can go where, what gets automated and what stays human, and they report honestly on what’s actually working. Mine is $2,495 a month with a six-month minimum — for businesses too small to justify a full-time AI hire but too exposed to keep improvising.
Isn’t it cheaper to just hire someone?
Not at this scale. A senior technical leader who could own this costs around $200,000 a year in total compensation before recruitment and benefits, for a role most small businesses can’t yet define well enough to hire against. At roughly $30,000 a year, the fractional version is about 15% of that, and you’re buying the decisions rather than the seat.
Is my team already using AI without telling me?
Almost certainly. Microsoft and LinkedIn’s Work Trend Index found 78% of people using AI at work bring their own tools without IT approval, and when I survey staff directly the usage is consistently far higher than the owner believes. It’s rarely recklessness — people weren’t told what’s allowed and don’t want to be the one who asks. The more useful question is what they’re pasting in.
What’s the biggest AI risk to a small business?
Not the dramatic one. It’s that critical work is quietly running in someone’s personal account, undocumented, and leaves when they do. Data exposure matters too — IBM’s 2025 research found shadow AI added around $670,000 to average breach costs — but shadow dependency is the one that actually bites. Second biggest is the people using nothing, costing you every day, invisibly.
How is this different from hiring a consultant?
A consultant delivers a project and leaves. This is a standing role — the decisions keep arriving, so someone has to keep answering them. Closer to a fractional CTO than a strategy engagement. The other difference is incentive: I’m not protecting a headcount or a department, so “cancel that, you don’t need it” is an answer I can give freely, and it’s the answer I give most.
Stop guessing
Find out what’s actually running.
One fixed price. I find what’s wrong, fix what’s causing the damage, and tell you plainly what’s left. If anything beyond $997 is needed, you’ll see the scope and the cost first.
About Stevie
I’ve been in the room for the version of this that goes wrong. My own business was taken apart by something running quietly, confidently, with nobody watching it, and by the time I understood what I was looking at, the decision had been made for me.
I do this work because “nobody’s in charge” and “everything’s fine” look identical right up until they don’t.
My own website was taken over once. It looked completely normal in my browser while tens of thousands of spam pages ran underneath it. I found out when the calls stopped, and I had to take the whole thing down. I diagnose before I quote, because I know exactly what it costs to be certain about the wrong thing. The whole story is here.
Frequently asked questions
What is a fractional AI officer?
Someone senior who owns AI decisions for your business a few days a month rather than full-time. They decide which tools are approved, what data can go where, what gets automated and what stays human, and they report honestly on what's actually working. Mine is $2,495 a month with a six-month minimum — for businesses too small to justify a full-time AI hire but too exposed to keep improvising.
Isn't it cheaper to just hire someone?
Not at this scale. A senior technical leader who could own this costs around $200,000 a year in total compensation before recruitment and benefits, for a role most small businesses can't yet define well enough to hire against. At roughly $30,000 a year, the fractional version is about 15% of that, and you're buying the decisions rather than the seat.
Is my team already using AI without telling me?
Almost certainly. Microsoft and LinkedIn's Work Trend Index found 78% of people using AI at work bring their own tools without IT approval, and when I survey staff directly the usage is consistently far higher than the owner believes. It's rarely recklessness — people weren't told what's allowed and don't want to be the one who asks. The more useful question is what they're pasting in.
What's the biggest AI risk to a small business?
Not the dramatic one. It's that critical work is quietly running in someone's personal account, undocumented, and leaves when they do. Data exposure matters too — IBM's 2025 research found shadow AI added around $670,000 to average breach costs — but shadow dependency is the one that actually bites. Second biggest is the people using nothing, costing you every day, invisibly.
How is this different from hiring a consultant?
A consultant delivers a project and leaves. This is a standing role — the decisions keep arriving, so someone has to keep answering them. Closer to a fractional CTO than a strategy engagement. The other difference is incentive: I'm not protecting a headcount or a department, so "cancel that, you don't need it" is an answer I can give freely, and it's the answer I give most.
Also in Run
Stop guessing.
One diagnosis. Find it, fix it, get back online. Small and medium sites are usually fully recovered within the $997.
The Site Rescue starts the moment your order is placed. I’ll email you within one business day, usually much sooner, to get access and book the 30-minute intake call. If the site is under attack right now, call (754) 302-4631 after you pay.
The Backstory
Two years ago my own website was taken over. It looked completely normal in my browser while tens of thousands of spam pages ran underneath it. I found out when the calls stopped, and I had to take the whole thing down.
I diagnose before I quote, because I know exactly what it costs to be certain about the wrong thing. The whole story is here.