Of all the posts in this series, this one has the least to do with technology and the most to do with running a business.
Microsoft and LinkedIn’s 2024 Work Trend Index found that 78% of people using AI at work were bringing their own tools: personal accounts, personal subscriptions and personal devices. The report called it “Bring Your Own AI.” In small businesses, the US Chamber of Commerce Foundation found that employees exploring tools on their own drive adoption nearly twice as often as direction from the business.
So for most owners, your team is almost certainly using AI already. The useful question is who is accountable for how they use it.
This usually has little to do with discipline. Useful tools tend to arrive faster than a business can respond, as email, smartphones and cloud storage did before AI. There are two separate risks here, and most businesses only deal with the first.
Risk One: What Goes Into Those Accounts
Nearly half of small business workers (47%) say privacy and security concerns are a barrier to using AI at work, according to the Chamber Foundation. Those concerns are well founded.
A large firm has an IT team and a vendor review process. A small business usually has neither, and its client list, pricing and margins are the business. One document pasted into the wrong tool can tell a competitor more than a year of market research.
The most likely failure is ordinary: a well-meaning employee pasting a client’s financial information into a free tool to get a faster summary, because nobody has told them where the line is.
What you need is a one-page policy. A long governance document is something nobody in a twenty-person business will read. One page covering four points, stated plainly, will do more:
| Policy point | What to write down |
|---|---|
| What never goes into an AI tool | Name the categories: client identifiers, financial records, employee personal data, anything covered by a contract |
| Which tools are approved | List them, and name the person to ask about adding another |
| Where AI-assisted work isn’t allowed | Name the client deliverables, regulated documents or contract work that must stay fully human |
| What happens when someone gets it wrong | Ask people to report it immediately so you can deal with it together |
The last point matters most. If people expect to get in trouble, they hide mistakes, and you never hear about the serious ones. You want to be the first to know.
Print the policy, walk your team through it in fifteen minutes, and review it once a year.
Risk Two: Who Owns What the AI Learns
This is the risk fewer owners have thought about, and for a small business it is often the bigger one.
Your business’s know-how, meaning how you actually do the work and what you have learned about your market, used to live in your people’s heads. You kept it by keeping good people. Now some of it also lives in the AI tools your team uses. Every time someone teaches a tool how you handle a situation, or builds up a knowledge base, a piece of that know-how moves into a system someone else may own.
Vinay Gidwaney, one of the book’s authors, puts the question to clients this way:
“What you’re doing is you’re taking the know-how of how to run your business, and you’re putting it into these AI systems. … You’re spending a lot of money retaining people because they have that know-how. But are you spending the same amount of effort to secure and retain the knowledge that you’re giving all these AI systems?”
Vinay Gidwaney, Workforce Intelligence, Chapter 5
There is a simple test for every AI tool you pay for:
If we stopped paying for this tomorrow, what would we lose?
If the answer is the software, you’re fine. If the answer is everything you have taught it, with no way to get it back, you have been building an asset that belongs to someone else.
Businesses learned this lesson with customer records years ago. Nobody sensible keeps their client list somewhere they can’t export it. The same rule applies to the know-how you put into AI tools. Prefer tools that let you export what you have built, ask the question before you buy, and keep your most important processes documented somewhere you control.
The Seven Note-Takers Problem
The book describes a finding from inside OneDigital that most owners will recognize. By early 2026, the firm had about seven different AI note-taking tools in use. Different teams had chosen them independently, for sensible reasons. Each stored transcripts in a different format and a different place, so the firm couldn’t combine what those meetings captured.
Your version will be smaller. One person uses one assistant for writing, another uses a different tool for scheduling, a third has a subscription nobody else knows about, and a fourth avoids AI entirely. Nothing has gone wrong, but nothing connects either, and what your team learns stays scattered.
Mandating a single tool usually fails, because people chose what worked for them. A middle path works better:
- Consolidate where tools overlap. Agree on one note-taking or transcription tool for the team, through a conversation.
- Agree on one shared place to record what works, such as prompts that produced good results and approaches to recurring problems.
- Give one person responsibility for that shared place, with twenty minutes a month in their calendar. Without an owner, it will be out of date within a quarter.
When the Person Who Knows Leaves
One more version of this risk has nothing to do with software. When the person who set up your AI tools leaves, their knowledge of how those tools fit into the work leaves with them: which tasks were handed over, which were deliberately kept human, and what was tried and dropped.
That person now has a rare and valuable skill, which makes them more likely to be recruited away. The fix is a short note for each tool you use, covering what it does, what it must never do, who owns it and what you decided to keep human. It takes ten minutes to write, and it means the capability stays with the business.
Your Next Steps
Ask your team, without accusation: “Who is using AI for work, and which tools?” You will probably learn more in ten minutes than in a month of reading.
Write your one-page policy using the four points above.
Ask the ownership question about every AI tool you already pay for and write down the answers. That list is your AI risk register.
If you’d like a second pair of eyes on your AI policy, the Focus HR team can help you put one in place.
Talk to the Focus HR teamSee all 8 posts in The 2027 AI Action Plan for Small Business
About the author: Clint Parry, MBA, SHRM-SCP is a Senior Business Consultant at Focus HR, where he helps Arizona small businesses with HR strategy, compliance and workforce planning. Focus HR is powered by OneDigital. This post is part of The 2027 AI Action Plan for Small Business series.


