Last month I published a short story about a telecoms store in 2030. Leon, the advisor whose glasses could hear intent. Laura, the customer whose new frames were reserved, financed and dispatched by drone before she had finished admiring herself in the mirror. A husband and wife who got talked down to a smaller television by a salesman who knew their extension wasn't finished plastering.

A few readers asked me the same question afterwards. How much of that is real?

Here is my confession. That story was not my first offence. I have form. I have written two science fiction novels, The Unravelling and The Paranoid Android, set in 2038 and 2075, and when I lined all three up recently I realised I have accidentally built a timeline. Forty-five years of predictions, filed in public, with my name on them. Which means there is only one honourable thing to do: mark my own homework and see how the future is scoring against the fiction.

I set Leon's story safely in the future. The future had other ideas.

Take the glasses. Laura's Gen7s "saved her life with translations" while travelling, and I thought I was being generous with the timeline. Meta's display glasses already put live translated captions on the inside of your lens today, controlled by a neural wristband reading the electrical signals in your arm, and the product line shifted more than seven million pairs in a single year. My fictional 2030 upgrade cycle is essentially a 2026 product roadmap with better branding.

The three-second credit check against Laura's "trust profile"? Agentic commerce got there before I did. Nearly half of consumers already use AI somewhere in their buying journey, ChatGPT will now complete a purchase for you without leaving the conversation, and Google has published an open protocol for letting AI agents shop across the entire journey. The drone that delivered Laura's frames between eight and ten that evening is barely fiction at all. Walmart and Wing are heading for 270 drone-equipped stores, and Amazon expects its drones to reach thirty million customers by the end of this year.

So the technology is not just on schedule. It is early, and showing off about it.

But here is the thing I keep telling anyone who will listen. The most fantastical line in that entire story was not the emotion-reading glasses or the ceiling robots. It was Leon's throwaway joke: "we stopped making people suffer around 2027."

Because everything that made Laura's purchase feel light had nothing to do with hardware. The identity that ported automatically. The contract that fitted on one screen. The absence of swivel-chair systems and re-typed passwords. That is not technology, that is orchestration, and orchestration is an organisational choice. The glasses will arrive on time. The operating model behind them is currently running about a decade late, and no drone can deliver it.

The Unravelling is set eight years after Leon's story, in a world that took his frictionless store and scaled it to everything. Homes, roads, markets, policing, all managed by systems trusted to make their own decisions. Into that world walks Dylan, an obsessive young man who works out that if you feed interconnected systems contradictory instructions, you do not need to break anything. You just need to confuse it, and civilisation does the rest to itself.

Look at what we are actually building. Leon's store works because everything trusts everything else instantly: the mirror trusts the profile, the credit system trusts the carrier, the drone network trusts the stock file, the customer trusts the lot. Agentic commerce means machines transacting with machines at a speed no human approval loop can follow. Every operator racing towards that frictionless orchestration layer is also, quietly, building a single point of failure, and I have yet to meet one who is stress-testing it with the same enthusiasm they are demoing it.

The Unravelling asks what happens when the machines we depend on stop behaving as expected. The honest answer for telecoms is that most operators cannot currently tell you what would happen, because nobody has rehearsed it. The frictionless future needs a resilience plan, not just a launch party.

The Paranoid Android is the quiet one. Zenon, an android butler, spends years reading a family: their micro-expressions, their vocal tension, their small evasions. He reads them so well that something unexpected happens. He starts to care. And then he has to choose between what he was programmed to do and what he believes is right.

Zenon is what Laura's Gen8 glasses become after another forty-five years of practice.

Remember her reaction in the mirror when Leon explained behavioural interpretation? "That sounds terrifying." She bought the glasses anyway, because they were useful, and that is exactly how it will go. We will not vote on whether machines get to understand intent. We will subscribe to it, one convenient feature at a time. The real question, the one Zenon exists to ask, is what happens when the systems reading us that closely start forming something that looks suspiciously like a point of view.

For retail, that lands somewhere very practical. There is a moment in Leon's story I put in deliberately: people had started laughing softly around advanced technology, as if the future kept arriving faster than their emotions could catch up. That gap, between what the technology can do and what the customer is emotionally ready for, is where telecoms retail will live for the next twenty years. Somebody has to stand in that gap, explain the terrifying-but-useful thing, and make it feel human. That is a job, and it belongs to the store.

So, will my story come true? Marking generously: 2030 is arriving early and I am claiming the points. 2038 is the one that keeps me up at night, because we are building Dylan's playground with genuine enthusiasm and no fire drill. And 2075 is too soon to call, although if you had told me in 2024 how quickly machines would learn to read a conversation, I would not bet against Zenon turning up ahead of schedule either.

Here is what all three futures have in common, and it is the part I did not plan. In every one of them, the store still exists. Not because people cannot buy without it, they clearly can, but because the more capable and unsettling the technology becomes, the more valuable the human layer gets. Leon is not in that story to process transactions. The ceiling robots do that. Leon is there because Laura needed someone to laugh with while the future fitted itself to her face.

The machines will hear intent. Count on it.

Whether anyone in your channel knows what to do with that intent is still, for now, a gloriously human problem.

The Unravelling and The Paranoid Android are my two novels. The stores in them are fictional. The mistakes operators are making on the way to them are not.

Amplify: straight talk on telecom retail, every fortnight, from someone who’s worked in 75+ countries.

Read next: “The Advisor who could hear Intent”

Want to see how we train telecom retail teams to sell? Take a free look inside Showtime.

Wondering how ready your stores really are? Try the readiness assessment.

Reply to this email and tell me what you'd do.