The AI Fog Is Real. Stop Making 5-Year Plans. Start Making 90-Day Bets.

Harvard Business Review just put a name to the thing keeping you up at night.

They call it “AI fog”. The idea, coined by Toby Stuart at UC Berkeley-Haas, is simple. AI is rewriting the future faster than any of us can plan for it. So the long-range decisions you used to make on autopilot, the ones that assumed tomorrow would look broadly like today, suddenly feel impossible. Source

If you’re running a UK business with 5 to 100 staff, you’ve felt this for months. You just didn’t have a label for it.

You sit down to write the 2027 plan. You stare at the screen. Then you close the laptop and go and do something useful, like answering an email or fixing the coffee machine.

The thing is, that’s not procrastination. That’s a rational response to a foggy world.

What HBR actually said

Stuart’s argument is worth reading slowly. He’s not saying AI will eat everything. He’s saying we genuinely can’t see far enough ahead to plan in the way our parents did.

His advice, and HBR’s, is to “swap long-term planning and optimise for optionality.” Source

That’s a fancy way of saying: stop trying to pick the right answer for 2030. Start making smaller bets you can change your mind about.

Let’s be honest, most strategy decks are theatre anyway. Lovely fonts, five-year hockey-stick charts, a slide called “Our North Star”. And then year one happens and the whole thing goes in a drawer.

The fog just made the theatre obvious.

The world we built assumes a stable tomorrow

Look around. Almost every big decision a normal person makes has a long tail attached to it.

A four-year degree. A thirty-year mortgage. A five-year business plan to wave at the bank. A pension you won’t touch for two decades. A strategy offsite where everyone agrees on goals for 2030 over a lukewarm pastry.

All of these assume the world in 2030 looks roughly like the world today, only a bit shinier.

Now look at what’s happening in AI in a single week.

Anthropic, the company behind Claude, just reported revenue growth of 80x in a single quarter. Yes, eighty. They shipped three significant product features the same week. NYTand The Neuron

You cannot plan a five-year operating model around a tool that doubles in capability between board meetings.

So either the assumption is wrong, or your plan is. Probably both.

What the big firms are doing wrong

Big corporates are responding to AI fog by doing the corporate thing. Hiring a Head of AI Transformation. Commissioning a 60-page report. Forming a steering committee. Building a roadmap for 2028.

Spare me the hype. By the time the steering committee has agreed on a logo for the initiative, the underlying tech will have shifted twice.

You don’t have a steering committee. You’ve got you, maybe a partner, maybe a head of operations who’s also doing payroll. That’s not a weakness here. That’s your edge.

You can change your mind on a Tuesday afternoon.

A 200-person company can’t. A 2,000-person one definitely can’t. They’re paid to be consistent. You’re paid to be right.

The optionality move

Optionality is a posh word for keeping your hands free.

Instead of one big bet for five years, you make several small bets for ninety days. You set up each bet so that if it works, you double down, and if it doesn’t, you walk away without bleeding out.

Three rules sit underneath this.

Smaller bets. If a project would sink the business when it fails, it’s too big. Cut the scope until it wouldn’t.

Faster reviews. Ninety days is the longest you should commit to anything you can’t reverse cheaply. Friday afternoons exist for a reason.

Willingness to abandon. This is the one most owners can’t do. You fall in love with a project. You’ve told your team about it. You’ve told your accountant. So you keep funding it long after the evidence says stop.

Here’s what works. Treat every initiative like a tenancy, not a marriage. Three months. Renew or leave.

Here’s what doesn’t. Treating the business plan as a sacred document because you spent £4,000 on a consultant to write it.

A 90-day rhythm you can actually run

You don’t need software for this. You need one whiteboard, or one side of A4, and a recurring Friday slot.

The cadence looks like this.

Day 0. Pick three bets for the next ninety days. Write them down. Each bet has a clear, boring outcome you can measure. Not “explore AI”, that’s not a bet, that’s a vibe. Try “automate the quote-to-invoice process for our top ten clients” or “test one AI sales agent on cold replies for one channel”.

Every Friday. Fifteen minutes. Are we still on the bet, or has the world moved? Anything to kill? Anything to double down on?

Day 90. Sit down properly. Three questions. What worked. What didn’t. What do we now believe that we didn’t believe ninety days ago. Write the next three bets. Repeat.

That’s it. No offsite. No Miro board. No facilitator in a beige cardigan asking how the brand makes you feel.

The boring rhythm is the strategy. Strategy isn’t the document, it’s the willingness to keep choosing.

Why this matters more for you than for a corporate

Here’s the thing nobody says out loud. AI fog is actually better news for small UK businesses than it is for big ones.

The big firms have spent twenty years optimising for predictability. Long contracts. Long planning cycles. Long approval chains. Their entire operating model assumes the fog clears. When it doesn’t, they freeze, and then they form another committee.

You don’t have any of that machinery. You can decide on Monday and ship on Wednesday. The same nimbleness that used to feel like a disadvantage when the world was stable is now the actual moat.

A SaaS competitor with thirty staff and a clean stack will beat a 300-person legacy player who can’t update their systems without a project manager and a six-week sprint plan. Not because the small one is smarter. Because the small one can change its mind.

At the end of the day, the question isn’t “what’s the right five-year answer”. The question is “what’s the cheapest way to learn what’s true in the next ninety days, and how fast can I act on it”.

That’s a question you can answer this afternoon.

The Practical Bit. The 90-Day Bet sheet

One side of A4. Stick it on the wall above your desk. Write it by hand, it sticks better.

Three sections.

The three bets. Three concrete things you’ll commit to over the next ninety days. One sentence each. Each one has a measurable outcome. Each one is small enough that if it fails, the business is fine.

The three Nots. Three things you are deliberately not doing this quarter. This is the bit owners skip. Strategy is what you say no to. Maybe it’s “no new website”, “no new hire”, “no new market”. Write them down so you have something to point at when the shiny thing tries to drag you off course.

The Day 90 decision. One question you’ll properly revisit on day 90. Not before. Things like “do we hire an AI ops person” or “do we sunset the legacy service line” or “do we double the marketing budget”.

That’s your strategy doc until 2027.

Review it every Friday for fifteen minutes. Kill what’s dead. Keep what’s working. Don’t add a fourth bet, three is the cap, your brain has limits and so does the business.

When the fog rolls in next time, and it will, you’ve already got the answer. You’re not planning for 2030. You’re planning for August. And in this weather, that’s the only honest move.

PS. If your accountant asks for a five-year forecast for the bank, give them one. Then put it back in the drawer where it belongs and get back to your 90-day sheet.