There is a strange tension developing around AI. Businesses are being told they need to move quickly, competitors are adopting it, new tools appear almost daily, and products that seemed impressive six months ago can suddenly look dated.
So I understand why a business owner might say, “We need to be using AI properly within the next three months.”
My first question would probably be: why three months?
Before deciding how quickly you can implement AI, you need to understand what you actually need it to do. That doesn't mean I think businesses should sit back and wait. Quite the opposite. I believe there is an opportunity for businesses that adopt AI successfully to gain an advantage while many of their competitors are still working out what to do with it.
But there is a difference between moving with urgency and rushing. Fear of missing out isn't much of an AI strategy.
AI adoption is growing, but what does “adoption” actually mean?
The latest UK figures make an interesting read. In July 2026, the Office for National Statistics reported that self-reported AI use among UK businesses with 10 or more employees had increased from around 12% in late 2023 to around 35%.
That sounds like rapid progress, but look a little deeper and the ONS describes adoption to date as “relatively shallow”. The average number of AI technologies being used by businesses that had adopted AI had increased only modestly, from around 1.4 to 1.6.
The UK Business Data Survey 2026 provides another useful clue. Among businesses already using AI, only 31% of small businesses and 31% of medium-sized businesses said their AI tools were integrated into their existing business systems.
I think this distinction matters because there is a considerable difference between someone in your company using ChatGPT to research a subject and redesigning a business process around AI or automation. Both might be described as AI adoption, but they aren't the same level of transformation.
Start with the business, not the tool
If I were sitting down with the owner who wanted their business “using AI properly within three months”, I'd want to understand two things initially. First, I'd ask which businesses they consider their competitors. Then I'd ask what evidence they have for how those competitors are actually using AI.
It's very easy to assume everyone else is miles ahead. A competitor announces an AI initiative, someone posts an impressive demonstration on LinkedIn, another business launches an AI assistant, and suddenly you're wondering why you haven't got an agent doing something clever.
That's when you can end up with what I call a pay-and-spray approach to AI: paying for tools and implementations in different parts of the business in the hope that enough of them will produce some value.
I'd rather start by understanding the business itself. Where are people wasting time? Which repetitive jobs are being done manually? Where are the bottlenecks? Where does information have to be copied between systems? What frustrates customers? Where are mistakes occurring or work being duplicated?
Once you understand those problems, you can start mapping potential AI and automation opportunities against their likely impact and the effort required to implement them. That gives you the beginnings of a roadmap, and once you have that, you can start talking sensibly about timelines.
You need the bird's-eye view and the ground-level view
I'd normally start this process with leadership because they can provide the bird's-eye view of the organisation: its priorities, problems, objectives and where they believe opportunities exist.
But I wouldn't stop there.
You also need to speak to the people who actually carry out the processes. Depending on the problem, that might be an Operations Manager, Ecommerce Manager, Customer Service Lead or another Product Owner who understands that part of the business.
Often you need follow-up conversations as well. You might map a process and then play it back to the Product Owner to check whether you've actually understood how it works. Questions and exceptions emerge, assumptions get corrected and you start seeing the process as it operates rather than simply how it was initially described.
Leadership can tell you what is supposed to happen. The people doing the work can tell you what actually happens. If you're trying to identify where AI or automation could make a meaningful difference, you need both perspectives.
There is another reason I think this matters. We've spoken to employees who understandably have concerns about what AI might mean for their jobs. Involving the people carrying out the work means they aren't simply on the receiving end of a technology decision. They can help identify the frustrating or inefficient parts of their work and contribute to shaping something that helps them do that work better.
That doesn't answer every legitimate concern about the effect AI will have on jobs, and I don't think businesses should pretend that it does. But it's a very different starting point from deciding what to automate in a boardroom and announcing it afterwards.
“We need an AI chatbot”
What happens when the owner already knows what they want?
Imagine the instruction is: “We need an AI chatbot for customer service.”
My response would still be to ask why. What problem are we trying to solve? Are customers waiting too long for answers? Are staff repeatedly answering the same questions? Are enquiries being lost? Is the existing help content poor? And, perhaps most importantly, is customer service actually the highest-value opportunity in the business?
The chatbot might be exactly the right project, but there could be another implementation elsewhere that creates considerably more value.
This is something I've seen throughout my career in digital delivery. Organisations can start looking for technology before they've spent enough time understanding what they need from it, both now and in the future. The desired solution becomes the starting point rather than the business requirement.
That doesn't mean turning the exercise into a battle with the CEO. If we've challenged the requirement, looked at the alternatives and they still want the customer-service agent, we can build it. But I'd still want them to understand the other opportunities we've identified.
Sometimes part of being useful is challenging the brief.
The same technology doesn't mean the same result
I learnt an interesting version of this lesson during my time at Essentra.
We were working with an IBM WebSphere implementation that, from my perspective at the time, was really poor. We were behind some competitors in terms of website functionality, there was functionality we wanted but didn't have, and I began to think the platform itself was part of the problem.
Then I attended a WebSphere community group organised by IBM.
I met people from other businesses using the same version of the platform we were using, and some of them had much of the “dream functionality” we wanted and needed. They had access to essentially the same technology, but they were getting much more from it.
That changed my view of the problem.
It wasn't enough to say, “We're using WebSphere.” The implementation mattered. The decisions that had been made mattered. How the business understood its requirements mattered. Simply having access to the same technology clearly didn't mean we would achieve the same outcome.
I think there's a lesson in that experience for AI.
Over time, your competitors may have access to exactly the same models, tools and underlying capabilities as you. If everybody can buy access to similar technology, simply having AI isn't much of a competitive advantage.
The difference may come from how well you've understood your business requirements, which opportunities you've prioritised, how you've implemented the technology, how well it fits into existing processes and, importantly, what your organisation has learned from actually using it.
So why is there any urgency?
This is where I think the AI conversation can become too binary. One side effectively says adopt AI now or get left behind, while the other responds that businesses shouldn't follow the hype and should wait until everything settles down.
I don't think either is particularly helpful.
AI isn't going to “settle down” conveniently while we all decide what to do. The UK Government has set an ambition for UK SMEs to become the most digitally capable and AI-confident in the G7 by 2035, and its current AI-adoption work is explicitly concerned with moving businesses beyond surface-level use towards transforming workflows, products and business models.
For me, the urgency isn't about implementing as many AI tools as possible before your competitors do. It's about starting the learning process.
Imagine two competing businesses with access to broadly the same AI capabilities. One spends the next couple of years watching the technology develop. The other starts examining its processes now, establishes where the friction is, prioritises opportunities, experiments, measures what happens, learns what doesn't work and gradually develops the capability to implement AI successfully.
Two years later they may still be able to buy exactly the same technology, but they haven't accumulated the same experience.
There may also be opportunities where speed has a more direct competitive effect. A business that identifies a genuine customer need and gets a useful AI-centred product or service to market ahead of its competitors could potentially gain market share while others catch up.
But that's very different from rushing out an AI product because someone has told the CEO that their competitor is “doing AI”.
Don't rush. Don't wait.
My concern isn't that businesses are moving too quickly towards AI. It's that some may move quickly in the wrong direction.
Buying several AI subscriptions, launching a chatbot and telling everyone to use Copilot might make an organisation feel as though it is moving forward. Perhaps it is, but I'd want to see the evidence.
What problem did we solve? What was happening before? What happens now? What changed for the customer or employee? What did it cost? What did we learn? And what's the next highest-impact opportunity?
Those questions lead towards a much more useful AI strategy than asking which tool everyone else is buying.
So I don't think “don't rush” and “start now” are contradictory. Starting now gives you time to understand the opportunities, build the right things, learn from them and improve. Rushing because you're worried about what competitors might be doing gives you a good chance of spending money before you've worked out what you're trying to achieve.
There may well be a competitive window around AI. I just don't think the way to take advantage of it is to let FOMO write your strategy.
If you're trying to work out where AI or automation could genuinely make a difference in your organisation, have a casual conversation with us at Flow Strategic. You don't need to arrive with an AI project already defined. In fact, one of the most useful things we might do is challenge whether the project you're currently considering is where you should start.