Feeling behind isn't being behind.
Everyone feels behind on AI. Most of the people you think are ahead, aren't.
You hear it in board meetings, supplier calls, networking events, internal management discussions. Are we behind? Should we be using AI more? What are our competitors doing? Should we have Copilot? Do we need a strategy? Are we missing something obvious?
The feeling has a name. AI FOMO. Fear of missing out.
And honestly — it's understandable. AI looks like it's moving at ridiculous speed. Every week there's another model, another agent, another product, another LinkedIn post from someone who's apparently rebuilt their business in an afternoon using nothing but prompts and positive thinking.
If you're a leader looking at all of that, it can feel as if everyone else has worked something out you haven't.
They probably haven't.
Most businesses that look ahead aren't. Some have bought licences they aren't using. Some ran impressive demos that never became operational. Some encouraged experimentation without sorting the governance underneath. Others have banned AI outright, which feels safe and isn't.
So the first thing worth saying out loud:
Feeling behind isn't being behind. It usually means you're paying attention.
The challenge is what you do next.
FOMO has two failure modes
A bit of FOMO isn't necessarily bad. It can be the prompt to look properly at slow admin, disconnected systems, the parts of the business that have always been a bit of a chore. That's useful.
It goes wrong in one of two ways.
Paralysis. The whole topic feels too big, too risky, too confusing. Nobody owns it. Nobody decides anything. Everyone waits for things to become clearer. Meanwhile, staff use AI quietly on personal accounts, on their phones, with nobody watching.
Panic buying. The business decides it must "do AI". A licence is bought. A flashy project is started. Nobody quite knows what problem it's solving, but the box on the slide deck has a tick in it.
Both are common. Neither is a strategy. The work is to slow the panic down just enough to make it useful.
You don't need to know everything about AI to make progress. You don't need a perfect strategy. You don't need to understand every model, framework or buzzword. You need to start asking better questions.
Start with the work, not the tool
The biggest mistake I see is starting with the product.
Should we use ChatGPT? Should we buy Copilot? Should we build an agent? Should we automate something? Not bad questions. Just not the best place to start.
A better starting question is:
Where is work harder, slower or more manual than it needs to be?
That moves the conversation from hype to usefulness. Look for the boring friction.
Where are people copying information from one system to another? Where are managers chasing approvals? Where are emails being triaged by hand? Where are reports being assembled manually? Where are the same questions answered week after week? Where does the business depend on one experienced person remembering how something works?
Sometimes the answer there is AI. Sometimes it's normal automation. Sometimes it's a missing integration between two systems that should be talking to each other. Sometimes the process needs cleaning up before AI goes anywhere near it.
Start with the tool, you end up looking for places to force it in. Start with the work, you can choose the right answer.
Make the safe route the easy one
A lot of businesses sit in a dangerous middle ground right now. They haven't formally approved AI tools. They haven't formally banned them either. So staff guess.
Some use AI carefully. Some paste in things they shouldn't. Some are on free tools without realising what happens to the data. Some are experimenting with customer, commercial or HR information because nobody told them where the line is.
That isn't the staff's fault. A simple usage policy doesn't have to be heavy-handed:
- Which tools are approved
- What data is fine to use
- What data is not
- Who to ask if you're unsure
- What to do if you've already done something you're now worried about
The point isn't to stop people experimenting. The point is to stop them guessing.
Done well, this also kills shadow AI — the version of shadow IT we now have where staff use unapproved tools because the safe route is too hard. The fix isn't a ban. The fix is making the safe route easier than the risky one.
Three lanes: assist, automate, integrate
A useful way to make AI less confusing is to split opportunities into three simple lanes.
Assist. AI helps a person do work faster or better — drafting, summarising, researching, rewriting, checking. The human is still doing the work, with help.
Automate. A task happens with less manual input — classifying tickets, routing requests, generating summaries, triggering workflows. A human reviews; the machine does the heavy lifting.
Integrate. Systems are joined up so information moves properly — CRM to finance, service desk to docs, forms to workflows, identity to onboarding. This is where a lot of mid-sized businesses find serious value, because much of their inefficiency isn't a missing brain. It's a missing pipe.
AI is powerful. AI connected to the right business systems, with the right controls, is much more powerful.
Pick one thing. Measure it in boring ways.
If AI feels overwhelming, shrink the problem. Don't start with how do we transform the business? Start with one process. Visible enough to matter, contained enough to manage.
Customer enquiry triage. Quote prep. New starter onboarding. Monthly reporting. Service ticket classification. Document drafting. Knowledge-base search. Whatever's broken in your operation that everyone already complains about.
Then measure it in normal business terms: hours saved, fewer manual steps, fewer errors, faster response, less chasing, less rekeying, more work done without adding headcount.
If a project can't be connected to a measurable business outcome, it might still be interesting — but it probably shouldn't be the priority.
The benefit of one finished pilot isn't only the win on that one process. It's the way it teaches the business how to think about the next opportunity. That learning often outlasts the project.
Close
There's no need to panic.
There's also no advantage in pretending this doesn't matter. AI will change how businesses operate. In some areas it already has. The companies that benefit most won't necessarily be the ones that moved fastest. They'll be the ones that moved deliberately.
They'll understand their risks. They'll train their people. They'll improve real processes. They'll connect systems that should already be talking. They'll use AI where it helps and be honest about where it doesn't.
That's the way through AI FOMO. Don't ignore the feeling — it's a useful signal. Don't let it drive the car either. A short readiness conversation is a better first move than a panic spend.
The real risk isn't being late. It's being passive.