The owner of a small accounting practice has decided that AI isn’t something her firm needs to think about yet. Then, at a team lunch, the conversation turns to work shortcuts. Her bookkeeper mentions using a free chatbot to draft client emails. A junior accountant has been pasting spreadsheets into an AI tool to summarise them. The receptionist uses an AI app to tidy up meeting notes.
None of them had asked permission, because nobody had said anything either way. And some of what they’d pasted in included client names and financial details.
It’s a scene playing out in businesses everywhere. Employees are adopting AI tools on their own, often to great effect, but without guidance on what’s safe, what’s accurate and what’s allowed. Global survey data suggests this is the norm rather than the exception, which means the real question for most small businesses isn’t whether to use AI, but how to use it well.
One Like insight: If you haven’t set rules for AI at work, your team has probably set their own.
Bring your own AI
In 2024, Microsoft and LinkedIn published their annual Work Trend Index, based on a survey of 31,000 people in 31 countries along with LinkedIn labour market data. It found that around three-quarters of knowledge workers were already using generative AI at work.
The more striking finding was how they were doing it. Of those using AI, 78% said they were bringing their own AI tools to work, a trend the report called BYOAI, short for “bring your own AI”. The pattern held for businesses of all sizes, and was even more common at small and medium-sized companies. Many employees said they used AI to save time, cope with workloads and focus on more important work. But a large share were reluctant to admit using it for their most important tasks, and many organisations had no clear plan or guidance in place.
The report also noted that leaders generally agreed AI was important for staying competitive, yet many felt their organisation lacked a plan or vision for putting it into practice.
The survey was conducted by Microsoft, which sells AI products, and it focused on knowledge workers rather than all employees. Adoption will also have grown since 2024. But the core finding is widely echoed: staff are often ahead of their employers in using AI, which creates both an opportunity and a risk.
A simple AI tools policy
A small business doesn’t need a long legal document. A one-page set of guidelines, agreed with the team, can capture most of the benefit and avoid most of the risk.
Approved tools. List which AI tools staff may use for work, ideally business versions with clear data protections, and which they shouldn’t.
What never goes in. Be explicit about information that must not be pasted into public AI tools: client names and personal details, financial records, passwords, confidential contracts and anything covered by professional confidentiality. In South Africa, POPIA applies to how personal information is processed and shared, as we explored in POPIA Has Teeth Now.
Always check the output. AI tools can produce confident, plausible errors. Anything going to a client, published or used for a decision must be reviewed by a person who understands the subject.
Be open about use. Encourage staff to share how they use AI, so good ideas spread and risky practices are spotted. Decide whether and how to tell clients when AI has helped produce work.
Learn together. Set aside time for the team to share useful prompts and lessons. As we explored in Nearly Two in Five Work Skills Will Change by 2030, working with AI is fast becoming a core skill.
[Infographic placement] A one-page AI policy. Five boxes: Approved tools (which ones, which versions) · Never paste in (client personal data, financials, passwords, confidential documents) · Always check (a person reviews before it’s used) · Be open (share how you use it) · Learn together (monthly tips session). Caption: Microsoft and LinkedIn Work Trend Index 2024: 78% of AI users brought their own tools to work.
Starting the conversation
The easiest first step is simply to ask. At your next team meeting, invite everyone to share which AI tools they use, what for, and what has worked well or badly. Treat it as a learning conversation, not an audit. You’ll probably discover clever uses worth spreading and a few habits worth changing.
Then agree your one-page guidelines together, choose one or two approved tools with appropriate business plans, and revisit the guidance every six months as tools and rules change.
A few questions can help:
- Which AI tools is our team already using, and for what?
- Has any client or personal information been entered into public AI tools?
- Who checks AI-generated work before it reaches a client?
- Do we have a clear, written view of what’s allowed?
Which brings us back to the accounting practice and its surprising team lunch. The owner didn’t need to ban AI; her team had already found real time savings. What they needed was clarity: a business-grade tool with proper data protections, a firm rule that client details never go into public chatbots, and a habit of checking everything before it left the office. With those in place, her team’s initiative became a genuine advantage instead of a hidden risk.
Sources
- Microsoft and LinkedIn, 2024 Work Trend Index Annual Report: AI at Work Is Here. Now Comes the Hard Part, May 2024.
