Synthetic Minds | Your Next Customer Is a Machine That Barely Spends
Synthetic Minds | Your Next Customer Is a Machine That Barely Spends
The Synthetic Minds newsletter offers short daily insights to get you thinking. If you enjoy it, please forward. All signals are powered by Futurwise. If you need more insights, subscribe to Futurwise and get 25% off for the first three months!
I built the Intelligence Age Scorecard! It will help you understand how ready your organization is for the Intelligence Age.
Today’s topic: Agentic AI & Tokenization
Machines Have Started Buying, Badly and Barely
Count every payer that could plausibly be a machine and software is 7.5 percent of the busiest rail built for software to pay for things. Demand proof, and it is 0.6 percent. That gap is the state of the art.
The machines buying on public rails are still a rounding error, and they already behave nothing like the customers those rails were designed for.
The uncertainty is the point. This is small enough to study completely, and the early readings say these buyers do not behave like the ones the rails were built for.
A blockchain analytics firm screened 198.9 million settlements to get there, and says the registries that would settle the question are voluntary and mostly unused.
Put a hundred of them in a closed town economy and multiply demand twelvefold, and the agents start to behave in ways not expected. Revenue rose more than fourfold, wages did not measurably move, and almost none of the prices changed.
Hand those agents cash and they hold 96.7 percent of it.
Nearly half of merchants say pricing is the last function they would hand an agent, and only fifteen in a hundred hold product data a machine can read.
A payment protocol for agents has gone live across wallets serving 1.5 billion accounts, carrying a trust score that governs how much freedom each agent gets.
Nine economists have modeled the loop where AI improves itself and found it not self-sustaining, and strengthening.
That's the adoption story. Here is the signal.
Nobody is going to rebuild a pricing system for a customer whose entire market moves $11,000 in a good month. That is the right call.
The merchants who hold no product data a machine can read did not choose to be invisible to a machine buyer. They never had a reason to look.
In this specific case, the organizations that look anyway might be better off, and not because the forecast comes true.
Every pricing model built for machine buyers assumes a relentless hunter. A hundred of them left 3,981 menu items almost untouched.
A team that tests its prices against a buyer who does not haggle learns what those prices actually depend on. That holds whether or not the buyer ever arrives.
Online retail was 0.6 percent of American retail sales when the Census Bureau first measured it. The firms studying it then were not visionary. They were reading a faint signal, and they were the only ones holding data when it mattered.
Prices, terms and access are beginning to be settled between machines. Most people will never choose their agent, never see the negotiation, never learn what a better one would have won.
So the question for a board is not whether machine buyers matter yet. It is what the business looks like to one, and whether anybody has looked.
The Intelligence Age Scorecard
The machines buying on public rails are still a rounding error, and they already behave nothing like the customers those rails were designed for. WAVE asks which part of the cycle a signal this faint actually demands: this one is Watch, not Adapt, and the discipline is knowing the difference.
Benchmark your readiness for the next two quarters, and the next five years, with the Intelligence Age Scorecard.
If this newsletter was forwarded to you, you can sign up here.
Thank you.
Mark
Frequently asked questions
How much of payment traffic actually comes from machine buyers?
When counting every payer that could plausibly be a machine, software accounts for 7.5 percent of traffic on the busiest rail built for software payments. But when proof of machine identity is demanded, that figure drops to just 0.6 percent, a gap that a blockchain analytics firm identified after screening 198.9 million settlements, revealing how uncertain the true scale of machine buying still is.</answer
Link to this questionHow do AI agents behave differently from human customers?
In a simulated closed town economy with a hundred agents and demand multiplied twelvefold, revenue rose more than fourfold while wages barely moved and almost none of the prices changed. When given cash, the agents held onto 96.7 percent of it rather than spending it, and a hundred of them left 3,981 menu items almost untouched, behaving nothing like the relentless hunters pricing models assume.</answer
Link to this questionWhy haven't merchants prepared their data for machine buyers?
Nearly half of merchants say pricing is the last function they would hand to an agent, and only fifteen in a hundred hold product data that a machine can actually read. This isn't a deliberate choice to be invisible to machine buyers; these merchants simply never had a reason to structure their data that way, since machine buyers have not yet become a meaningful part of the market.</answer
Link to this questionShould businesses act now on machine buyers even though they're still rare?
Yes, because testing prices against a buyer who does not haggle reveals what those prices actually depend on, regardless of whether machine buyers ever become significant. This mirrors early online retail, when it was just 0.6 percent of American retail sales; the firms studying it then weren't visionary, they were simply reading a faint signal and were the only ones holding useful data when it mattered later.</answer
Link to this question
Dr. Mark van Rijmenam
Dr. Mark van Rijmenam is The Digital Speaker, a world-leading futurist and global AI keynote speaker who helps leaders decide what AI, and the technologies that follow it, mean for their business. He speaks to leadership teams at their retreats and to conference audiences worldwide, and his WAVE framework gives them a way of deciding that they keep using long after he has left the room. His Intelligence Age Scorecard measures how ready an organization is across the four WAVE pillars.
He holds a PhD from the University of Technology Sydney, is a Certified Speaking Professional and Global Speaking Fellow, and ranks among the world’s top futurists on the Global Gurus list (#1 in 2025, #3 in 2026). Salesforce named him one of its 16 leading AI voices. He has written six books, most recently Now What? How to Ride the Tsunami of Change, writes the Synthetic Minds newsletter on Substack, and founded Futurwise. Talk to his digital twin on WhatsApp: +1 (830) 463-6967.
