In 2010, Blockbuster had 9,000 stores, $6 billion (roughly £3.9 billion) in revenue and a brand that everyone recognised. By 2013, it was gone. Not because the team wasn't talented. Not because the finances were broken. Because the operating model couldn't adapt to a structural shift in how their industry worked.
AI is that kind of shift. Not because it's flashy or because the press says so, but because it's changing what it costs to do work, and therefore what businesses need to look like.
This isn't about technology. It's about operating models.
If your competitor can serve the same customers with 40% fewer manual processes, respond to regulatory change in days rather than weeks and personalise their service at a fraction of the cost, they don't need to be better than you. They just need to be cheaper, faster and more consistent. AI makes that possible for mid-market businesses in a way it wasn't five years ago.
The question isn't "should we use AI?" The question is "what does our business need to look like in three years, and what role does AI play in getting there?"
Where mid-market CEOs get stuck
Most CEOs we speak to aren't sceptical about AI. They're stuck between three uncomfortable truths:
First, they know they need to act. The board is asking, the PE sponsor is asking and the competitors are moving. BDO's 2026 mid-market survey found 42% of UK mid-market firms now cite AI and productivity as their primary growth route. Second, they don't know where to start. There's no shortage of vendors offering tools, platforms and "AI solutions," but nobody is helping them answer the prior question: where does AI actually matter for *this* business? Third, they're worried about getting it wrong. In regulated sectors (insurance, financial services, legal, professional services) getting AI wrong doesn't just waste money. It creates regulatory risk, reputational damage and operational fragility. The cost of a failed AI project isn't just the budget; it's the organisational confidence to try again.
What "doing AI" actually means for a mid-market business is not buying a platform, hiring a data scientist or giving everyone access to ChatGPT and hoping for the best.
It means understanding your business capabilities (the things your organisation does to create value) and working out which of those capabilities can be materially improved, automated or redesigned using AI technologies. Some of the answers will be obvious (document processing, data extraction, customer triage). Some will be surprising (pricing models, risk assessment, workforce planning).
The starting point is always the same: a structured look at what your business does, scored against where AI creates real leverage. Not where a vendor wants to sell you something. Where the business case is real.
The "Blockbuster test" for your business
Ask yourself three questions:
Could a competitor deliver our core service at meaningfully lower cost by using AI to automate what we currently do with people and spreadsheets? If the answer is yes, that's your cost exposure.
Are there parts of our customer experience that are slow, inconsistent or manual, and that a competitor could make instant, consistent and automated? If the answer is yes, that's your service exposure.
Is our operating model designed for the way business was done five years ago rather than the way it will be done in three years? If the answer is yes, that's your structural exposure.
If you answered yes to any of those, you don't have an AI problem. You have a business model question that AI is part of the answer to.
The businesses that get this right don't start with technology. They start with strategy.
They map their business capabilities and score each one against AI opportunity, not in theory but with data about their actual processes, costs and pain points. They produce a roadmap that the leadership team (not just the IT team) can stand behind. And they start with a focused pilot that proves value before scaling.
This isn't a twelve-month programme. The first phase (understanding where AI matters and building a prioritised plan) takes days, not months. The businesses moving fastest are the ones that invested a week in getting the strategy right before spending a penny on technology.
The cost of waiting
The World Economic Forum research is clear: AI can help mid-market companies scale "much faster" than traditional approaches. But the advantage goes to early movers, not because the technology will be different in two years but because the organisational learning compounds. The businesses that start now will have built the muscle memory, the governance frameworks and the operating model changes that make them structurally faster and more efficient.
The businesses that wait will find themselves trying to catch up against competitors who've already embedded AI into how they operate. That's the Blockbuster position: not wrong, just late.
A practical first step
If you're the CEO or MD of a mid-market business and you're not sure where to start, the answer isn't to hire a data scientist or buy a platform. It's to invest a few days in understanding where AI actually matters for your business.
A structured discovery sprint (mapping capabilities, scoring opportunities, building a roadmap) gives you the clarity to make real decisions. Not a slide deck. A plan your leadership team can execute.
That's what we do at Oxygen Bubbles. Our Breathe engagement takes 5–10 working days and gives you a capability-scored AI roadmap, a 90-day pilot blueprint and a board-ready summary. No tools to buy. No lock-in. Just clarity about what your business needs to do next.
Not sure if your business model is AI-ready? That is what the Breathe engagement answers. Get in touch to talk it through.