The hotel booking test: what Martha Lane Fox gets right about AI, and what SMEs should actually do about it

A line from Martha Lane Fox in The Times stuck with me. AI might kill us all, and yet it still cannot reliably book her a hotel. I have spent 30 years watching technology promises collide with what actually ships, and that gap between the apocalyptic headline and the everyday failure is exactly where I have learned to look for the real story.

Her point is not that the risks are fake. It is that the extinction prediction and the booking failure are both true at once, and only one of them tells an SME owner anything useful about what to do on Monday morning. I agree with her instinct completely. Confident predictions about where technology ends up have a poor track record, and nobody building websites in the 1990s predicted the internet we actually got.

What struck me most was the adoption number she cites: AI use among UK businesses has risen from about 12 per cent to 35 per cent in under two years, yet only one in ten of those businesses says it uses AI extensively. That gap is not a technology problem, it is an organisational one, and it is the exact gap Bield exists to close. Getting hold of the software is the easy bit. Redesigning how the work actually gets done around it is where almost everyone gets stuck.

Her point about the government's Digital Champion years lands hardest for me, because I have lived a version of it from the supplier side. Putting a clever front end on a broken process does not fix the process, it just makes the frustration faster. I have watched businesses buy a platform, switch it on, and wonder six months later why nothing changed. The tool was never the constraint. The workflow built around thirty years of "that's just how we do it" was.

Her taskforce finding, that the people doing the work need to be involved in changing it, is the same lesson I took from building Bield with AI rather than just reading about it. The technical capability was never the hard part. Knowing which process to change, and having the people who actually do that process in the room when you change it, was.

Where I would push slightly further is on where the risk conversation should actually sit for most businesses. Lane Fox is right that safety belongs in how AI is governed in practice, not in speculative headline numbers. For most SMEs, that practical risk almost never lives at the infrastructure layer that dominates the sovereignty debate. It lives at the application layer: what data goes into a tool, what a contract actually says about retention, who has access to what an agent can now touch. That is solvable with discipline and the right agreements, not with a different data centre.

AI is not weather heading towards us with a fixed outcome, and neither is an SME's adoption of it. Between an agent that cannot book a hotel and whatever comes in ten years lies a great deal of unglamorous, doable work. Most businesses have not started it yet. That is the opportunity, not the extinction debate.

Source reference: Martha Lane Fox, "AI is not an inevitable disaster, we still have power to shape it," The Times, 14 September 2026, https://www.thetimes.com/article/fd04bc6f-9fc8-4662-8f1e-5f7201c8c33c