Retail & Technology

The loyalty card was never about loyalty. It was about data you could not otherwise buy.

4 November 2026

Every few years the retail trade produces a list of loyalty trends for the coming season, and every few years the list says much the same thing in slightly newer language. This year’s, a round-up from the trade site EXEIdeas, is as good a specimen as any, mainly because of what its conclusion admits.

The piece runs through the familiar eight: AI-powered hyper-personalisation, first-party data, omnichannel, real-time rewards, gamification, paid membership tiers, mobile apps, and loyalty wired into the point of sale. “The future of loyalty,” it concludes, “is not a card. It is a relationship.” True enough, but it skips a more useful sentence: the card was never really about loyalty either.

The list rewards the shopper who tells you things, whether or not that shopper loves you. Read plainly, every item on it is a way of extracting first-party data at the moment of purchase: what you bought, when you came in, what you nearly bought and walked past. The points are the fee the retailer pays you for that information, and the discount is the receipt.

This matters more in 2026 than it did in 2016 because the signals retailers used to rent from the open web are drying up. Third-party cookies are going or gone, the walled gardens charge more to reach your own customers, and the privacy rules have tightened around everything in between. First-party data, the kind a shopper hands you willingly in exchange for a reward, is now the one asset a retailer owns outright and no platform can repossess. That is why the trade press keeps circling it, and why the loyalty programme has become the last piece of ground a retailer controls.

It is also why the “relationship” framing misleads if you take it at face value. The EXEIdeas list treats paid membership, the Amazon Prime model of charging shoppers for the privilege of being loyal, as just another trend, when it does a different job. A points card buys data from the many. A paid tier sells status to the few, and in doing so tells you precisely who your most valuable customers are, because they have raised their hands and paid to be counted. One programme is a net and the other a filter. Retailers who run both without knowing which job each is doing end up with a bulging database and no idea who in it matters.

All of this data collection assumes the shopper is doing the choosing, and increasingly they are not. When a shopper asks an AI assistant what to buy, the answer is likely to turn on price, availability and whatever the assistant can read about the product, with the points in your scheme counting for little. The retailer’s hard-won first-party data sits inside a loyalty silo that is probably out of the agent’s sight, and nothing suggests it would carry much weight if the agent could see it. You can know everything about a customer and still lose them at the moment a machine makes the introduction on your behalf.

The list never asks what all that data is for when the thing forming the impression that decides the sale is no longer a human you can personalise an email to. Loyalty data was built to win the shopper, and the contest is moving to winning the shopper’s agent, which takes a different skill.

What to watch. See whether any loyalty platform starts exposing its first-party signal outward, to the agents and assistants doing the recommending, rather than hoarding it inward for email campaigns. A programme that makes its data legible to the machine without handing away the margin has understood where the moment of truth has moved.To the Machine Moment of Truth. One still counting app downloads has not.

The Roth Read. If you run a loyalty programme, stop asking how to make customers feel loved and start asking what, precisely, you are buying with every point you give away. Data you can act on is an asset. Data that dies inside your own app, probably unseen by the machine now choosing on your customer’s behalf, is a cost dressed up as a relationship.

Tally counts shelves. What it saves is hours.

3 November 2026

A six-foot tower with coloured lights and a screen reading “Hi, I’m Tally!” is now rolling down the aisles of Ralphs stores in Los Angeles. The robot is what gets the attention. The more interesting question is what happens to the time it saves.

Tally is an autonomous inventory robot made by Simbe Robotics, a firm based in Burlingame. It moves through the aisles on its own and uses cameras and computer vision to spot products that are out of stock, wrongly priced or in the wrong place. It then sends shop staff a list of jobs in order of priority. Simbe says more than 3,000 of its robots have been contracted for retail deployments worldwide, and Lowe’s, Tesco, BJ’s and Bristol Farms are among the reported hosts. Kroger owns the Los Angeles stores. The figure Simbe leads its sales pitch with is this one: Tally can “reallocate” up to 50 labour hours per store per week away from repetitive shelf checks.

That word “reallocate” is carrying a lot of weight.

Retail workers have heard this kind of reassurance before. Self-checkout was sold as convenience and electronic shelf labels as accuracy. Both were presented as help for the person on the shop floor, and both ended up changing what that person was there to do. Kathy Finn, president of UFCW Local 770, put the complaint plainly: workers “should not be finding out about major changes to their workplace only after new technology shows up in their stores.” NewsNation, citing new research on the future of retail jobs, reports an estimate that up to seven and a half million retail jobs could be automated away in the coming years, out of a workforce of roughly 16 million Americans. Kroger has announced plans to cut 3,000 corporate jobs over three years while it expands its AI tools. Those cuts are in head office, not in the stores, but staff take announcements like that as a sign of where the company is heading.

Tally clearly works. The question for any retailer or brand watching is what an empty shelf really costs them, and where the saving goes once a machine finds the gap faster than a person could.

A shelf that is fully stocked, correctly priced and neatly faced is the minimum a shopper expects, and nobody ever thanks you for it. Shoppers remember the assistant who knew where the tahini was, who told them the better avocados come in on Thursday, and who made the trip feel like a place rather than a warehouse with a till. Tally checks the shelf brilliantly but can do nothing about the people. A grocer who uses the 50 hours to put more knowledgeable staff in front of shoppers has bought itself an edge. A grocer who uses the robot to cut 50 hours has bought a cheaper version of a commodity. Commodities compete only on price, and that is a race against Amazon a grocer will not win and should not try to.

The robot doesn’t decide anything. The profit-and-loss decision is the retailer’s.

What to watch. Privacy. Simbe is careful to say Tally is “designed to understand shelf conditions, not identify or track humans,” and it separates shelf analysis firmly from tracking people. What Simbe hasn’t said is what the sensors capture, how long the data is kept or who can see it. A dozen high-resolution cameras going up and down a public aisle several times a day amount to a piece of infrastructure, and infrastructure gets put to new uses. Watch whether the robot that counts tins is ever asked to count footfall, how long shoppers linger, or the staff.

The Roth Read. If you run a store, Tally has just given you 50 hours a week and a question about what kind of store you are. Spend those hours on cheaper operations and your shop becomes more like everyone else’s. Spend them on people who make shoppers feel known and you have bought the one thing a competitor cannot scan off a shelf. The robot finds the gap. You decide what fills it.

The screen was never the asset. The proof is.

28 October 2026

A retailer can hang a screen on every shelf and call itself an advertising business. That works until a brand asks whether any of it sold anything, and how the retailer knows.

Retail media now turns on that question. Osmos published an industry guide to retail media attribution this month, and its definition is plain. Attribution is how a retailer proves to an advertiser that an ad led to a sale. The retailer matches each impression or click to a purchase in its own transaction records. Online it does this through logged-in accounts, and offline through loyalty IDs, payment cards or store data. It uses a window agreed before the campaign starts. A companion piece from SignStream covers the in-store side. It cites a pooled field study of 237 campaigns, which found that digital signage in stores lifts purchase probability by about 8.1 per cent on average. The effect was stronger for novel, low-priced and impulse goods. Read together, the two pieces show that the money is real and the measurement is still not trusted. Whoever closes that gap will lead retail media for the next few years.

Retail media is the most profitable thing in modern retail because the retailer owns both the audience and the till. That margin is on loan until the measurement holds up. The IAB guidelines cited in the Osmos guide warn that a seven-day attribution window and a thirty-day window can produce very different returns for the same campaign. A retailer marking its own homework can therefore report almost any number it likes. Brands know this. They have watched the same screen sell them an 8.1 per cent lift on a category where the real effect was a rounding error, because the average was flattered by the few products where signage works. An average sold as a guarantee is a hope with a decimal point.

For a decade the retailer’s question was how brands ranked on its own shelf. The new question is colder: what does the advertiser’s spreadsheet believe about the retailer? The network that wins will be the one whose numbers a brand’s finance director signs off without flinching, whatever its screen count or the size of its loyalty base. Third-party verification, disclosed lookback windows and honesty about which categories the lift is real in all sound like compliance work. In practice they are what the brand is paying for. Any retailer can sell inventory, and very few can sell a number a brand believes.

The West should look East before congratulating itself. China’s retail media never had room for fuzzy attribution, because the shelf, the payment and the loyalty identity sat inside one account from the start. On Alibaba and JD.com the impression and the purchase land in the same ledger, so the loop closes by default. Western retailers are now spending heavily to join that loop after the fact, stitching card data to loyalty IDs to door counters. The Chinese platforms built the measurement in from day one, and everyone else is retrofitting it.

What to watch for is the first large grocer or retailer to publish its attribution methodology in full, with its windows, its verification partner and its category caveats, and then to sell against it. Once one does, keeping the method private will look like an admission. The laggards will be the retailers who still cannot tell a brand which half of the lift was real.

The Roth Read. If you run a retail media network, stop selling the audience and start selling the audit. Brands are buying a number they can defend to their own board, and the first retailer to show its working will take budget from the ten that hide theirs. The screen is not the asset. The proof is.

The robots went to Seoul looking for work

27 October 2026

A Chinese humanoid went to Seoul this autumn and found work before Korea had a rival of its own to send. That gap is the story.

The Korea Times, in partnership with the South China Morning Post, reports that Chinese makers have moved on from dance routines and showroom poses to factory trials and paying customers. A robot-themed venue in Seoul bought UBTech’s lifelike U1, delivered last month for around 170,000 yuan, roughly 18,000 pounds, according to the company’s Sean Qin. UBTech has also announced a trial on a Korean automotive parts line with local partners. Anyverse Dynamics, a Beijing startup whose founder Zhang Yufeng once worked with SK Telecom and Hyundai, plans café trials in the fourth quarter before moving into manufacturing. On the Chinese customs data cited in the piece, Korea was China’s third-largest export market for intelligent bionic robots in the first nine months of 2026, taking just under 20 million dollars of a 207 million dollar total.

Counterpoint’s Ethan Qi gives the reason plainly: nobody builds humanoids locally at scale, and that has left an opening for Chinese suppliers. Korea makes the world’s robotic arms, ships and chips, but not yet the humanoid. So the most automated manufacturing nation on earth is importing the machine that will automate the next layer of its work, from the one country that learned to build it cheaply and ship it fast.

Why it matters. Seoul has seen this coming. The government has proposed 3.1 trillion won, about 1.75 billion pounds, in physical AI funding for 2027, more than triple this year’s, with a target of 500 AI factories by 2030. Samsung set up a robotics division in July. LG is building a data factory to train robots. Hyundai is putting Boston Dynamics’ Atlas through drills at its Georgia campus ahead of a 2028 line deployment. Korea has the intent and the money. It lacks what the robot runs on today: a supply chain, engineering depth and rapid iteration of the kind that only exists at scale in Shenzhen. Korea can fund the ambition in a single budget, but it cannot buy back the five years of deployment data China gathered while the West debated whether a humanoid needs five fingers or two.

Deployment data decides who wins, because a robot that has served a thousand coffees knows things a prototype never will. Every trial Anyverse runs in a Seoul café and every UBTech unit tested on an AJIN parts line feeds a loop that compounds. Deputy Prime Minister Bae Kyung-hoon made the same point at a Seoul roundtable: China is deploying more humanoids in the field and collecting the data. He then asked whether Korea’s plans were bold enough. The politics suggest they are not. At a parliamentary audit on 6 October, one lawmaker wheeled in Unitree and AgiBot machines to shame the government, and another accused the Assembly of handing Chinese technology a stage. Both reactions sound like a country discovering it has become a customer when it meant to be a champion.

For retailers and brand owners the lesson reaches well beyond Korea. If you do not make the layer that matters, you do not set the terms on it. You pay the price the maker names, adapt to their certification and wait your turn. Korea is learning this as national industrial strategy. A retailer learns it the day its store robots, fulfilment, shelf scanning and service all arrive on hardware built by a company that also sells to its competitors.

What to watch. Watch whether Korea’s big manufacturers put a humanoid into their own lines before Chinese suppliers take the proof-of-concept slots. Whoever goes first collects the data, the certification path and the reference customer, and those are much harder to win back than a robot.

The Roth Read. No retailer is going to build its own robots. The hardware will come from someone else, most likely from Shenzhen, so the layer a retailer can own is what those machines learn in its stores: which jobs they do well, and where shoppers still want a person. The first retailer to put them to work starts learning first, and competitors have to make up every month of that lead. You can buy the hardware at any time, but you can never buy back the time. Decide this quarter how you will deploy, and write into every contract that the data from each robot in your estate belongs to you. Then put a team and a plan behind it, because owning the data is worth nothing until someone is learning from it.

BYD drew a robot. The drawing you should study is the dealership floor.

15 October 2026

BYD filed a design patent this week, and the drawing matters less than where BYD says the robot will go.

On 9 October, China’s National Intellectual Property Administration published an “Embodied AI Robot” design patent with BYD as the holder, according to CarNewsChina and Star News Korea. The drawings show a black-and-white humanoid with a rounded, screenless head and jointed torso, arms and legs, designed to walk, balance and interact with people. It looks very like “Xiao Di”, a robot a BYD executive teased on social media earlier this year. Figures disclosed earlier put Xiao Di at 1.61 metres and 58.5 kilos, with 31 degrees of freedom and a hand accurate to plus or minus one millimetre. BYD has not confirmed that the patent shows the production model.

Most carmakers now have a humanoid in a slide deck. BYD has also said where its one will work. According to CarNewsChina, its executives plan to deploy these robots across the company’s own dealership network, to greet customers, explain the cars and demonstrate the in-car infotainment. They are meant for the showroom, not the factory.

That makes this a retail story. For twenty years the dealership has been the weak link for premium car brands. Service varies from one site to the next, and a carefully built brand can be undersold by whoever happens to be on the floor that Saturday. BYD wants to automate the one role no carmaker has ever controlled: a salesperson who never has an off day, never improvises the pitch, never forgets the script, and demonstrates the infotainment the same way in Shenzhen as in Stuttgart. The robot is the headline. The bigger change is brand control reaching the last metre of the sale, the one part the brand has never owned.

China matters here because of how BYD builds. It is the world’s largest integrated maker of electric vehicles, and it is making this robot the way it makes everything else: in-house, at scale and from its own parts. It set up a dedicated robotics team in 2022, invested in AgiBot in 2023 and took a stake in Percibot in 2025. It has also been running Ubtech’s Walker S1 inside its own Changsha plant to collect operating data. BYD already makes the batteries, motors and actuators and runs the factory floor, so the robot is simply the next product down the same line. A Western carmaker partners with a robotics firm and negotiates over the margin. BYD owns both sides of that deal, and Western boards keep underestimating what that is worth.

There is reason for caution. A design patent is a drawing. There is no price, no reliability data and no deployment date, and Xiao Di may never get beyond the teaser video. A hand accurate to a millimetre still has to survive a showroom full of children and spilled coffee. So far BYD has shown renders and no revenue, and scepticism is fair.

The intent is still clear. BYD wants to know whether a humanoid can be the brand’s own face in the room where the sale happens. That is a far more commercial question than whether one can work in a warehouse, and almost no Western retailer or carmaker is seriously asking it yet.

What to watch. See whether BYD puts one of these robots on a dealership floor before it puts one in a factory. The order of deployment will show what BYD thinks the robot is for. If customers meet it first, BYD is selling brand control rather than cost savings, and that is the version retail executives should worry about.

The Roth Read. Picture the robot on your own shop floor, in your uniform, saying your words every time, without fail. BYD has said its humanoid belongs where the customer decides, in the showroom. If your brand still treats the last metre of the sale as somebody else’s problem, BYD is about to show you what that assumption costs.

The fee was never the point. The float was.

16 September 2026

Every guide to accepting crypto in 2026 leads with the same number: the fee. CoinRemitter at 0.23 per cent. NOWPayments from 0.5 to 1 per cent. Stripe at 1.5 per cent on stablecoins. Set against the 1.5 to 3.5 per cent that card processors charge, on UPay’s own figures, it reads like a bargain the retailer would be foolish to refuse. But the fee is the decoy. The real story is which token lands in your account, on which chain, and whether you can ever spend it.

What happened

The comparison sites have industrialised. The Bitcoin Foundation’s 2026 ranking lays out five gateways on fee, supported coins and settlement path, from CoinRemitter’s no-KYC crypto-only model to BitPay’s daily bank withdrawals in dollars, euros and sterling. UPay’s guide names eleven, adding enterprise infrastructure players like BVNK and CoinsPaid, the latter having processed over 29 billion dollars on its own reported figures, mostly for Europe’s iGaming operators.

The more interesting document is the one that ignores fees almost entirely. EdgeX’s 2026 stablecoin guide argues the choice between USDC, USDT, PYUSD and EURC “is less a question of market capitalization than of workflow fit.” USDC for regulated checkout and treasury. USDT where local liquidity decides whether a supplier can actually cash out. PYUSD inside PayPal’s walls. EURC for euro invoices. Same dollar peg on the label. Very different money in the hand.

Why it matters

Here is the shift a retailer has to grasp. A card payment is a single decision: accept Visa, or don’t. A crypto payment is a chain of them, and each link carries a cost the headline rate hides. There is the on-chain gas fee. The provider’s cut. The FX spread when you convert to the currency you pay rent in. The compliance screening. The reconciliation time. EdgeX puts it plainly: the real cost “includes the token, the chain, the provider, FX conversion, compliance review, reconciliation, and the off-ramp.” The 0.23 per cent was true and also almost meaningless.

Then there is the machine underneath. When Stripe re-entered this market it did not build rails. It bought Bridge, the stablecoin infrastructure company, and folded acceptance into the dashboard a merchant already knew. That is the tell. The value is migrating from the token to the orchestration layer, the software that mints, screens, converts and settles while the merchant sees only “paid.” Whoever owns that layer owns the margin, the data and the relationship. The coin is just the thing moving through the pipe.

And notice what crypto quietly removes. Chargebacks, estimated by Chargeback Gurus to have drained 33.8 billion dollars from merchants globally in 2025, vanish because blockchain settlement is irreversible. For the retailer that reads as a saving. For the shopper it reads as the disappearance of buyer protection. A card gives the customer a way to be wrong and get their money back. An irreversible payment does not. That is not a feature you advertise at checkout. It is a trust you spend.

What to watch

Watch MiCA do to Europe what it was built to do: sort the field. UPay’s guide already flags that EU businesses “must now consider MiCA licensing,” and CoinGate is being marketed on compliance rather than price. When regulation becomes the sales pitch, the low-fee, no-KYC operators do not win the enterprise account. They lose it.

The Roth Read. Stop shopping for the lowest fee. It is the cheapest number on the page because it is the least important one. Ask instead which token lands, on which chain, who holds it while it settles, and what your customer loses when the payment can never be reversed. The retailer who accepts crypto to save half a per cent, and hands a stranger’s software the float, the data and the buyer’s only recourse, has not cut a cost. They have sold the counter and kept the rent.

The subsidies were never the point. The habit was.

10 September 2026

For a year, three of China’s largest companies spent billions of dollars teaching their customers a single reflex. That reflex has now been learned. The coupons are being withdrawn, the free-delivery banners are coming down, and what is left behind is worth more than everything the subsidies cost. A new expectation. When I think of something, I buy it and get it right away.

That sentence is not mine. It belongs to Jiang Yanxin, a Beijing shopper quoted by Reuters, who ordered a doll on her way to meet friends for lunch and found a courier already at the restaurant by the time she reached her table. “I’m used to shopping this way now,” she said. That is the whole war in one line. After a year in which Meituan, Alibaba and JD.com poured money into coupons, free delivery and merchant incentives, what I call  instant retail has become the new battleground: electronics, flowers and even medicine, delivered in under sixty minutes.

The scoreboard has already moved. Goldman Sachs said in April that Meituan’s meal-delivery share had slipped from the 75 to 80 per cent it held before the price war. On Analysys data cited by Reuters, Meituan commanded 45.3 per cent of the broader instant-retail market in the second quarter, with Alibaba’s Taobao Instant Commerce ahead at 45.7 per cent and JD.com holding 7.7 per cent. The meal-delivery fight, in other words, has been swallowed whole by a bigger one.

Here is why it matters, and why the Western reader should not file this under “another Chinese price war.” The subsidies were a customer-acquisition cost disguised as generosity. Liu Xingliang, director of the Beijing-based Data Centre of China Internet, put it precisely to Reuters. The industry, he said, “has moved from the first stage of winning users through subsidies to a second stage of retaining users, expanding supply and calculating order-level economics.” Translation: the giants bought the habit at a loss, and now they must make the habit pay. The clever part was never the discount. It was recognising that a shopper who has had paracetamol at her door in under an hour will never again plan a trip to the chemist. The behaviour is a one-way door.

The damage sits where it usually sits. The food industry analyst Zhu Danpeng, quoted by Reuters, says the battle benefited consumers but the damage to small restaurant operators is still there, because a subsidised order is a thin order, and thin orders on someone else’s platform are a poor way to run a kitchen. That is the ledger the West should read most carefully. Instant retail does not create demand so much as it relocates margin, from the shop you owned to the network you rent. The convenience is real. So is the tax on it.

For a Western retailer, the lesson is not “build one-hour delivery.” It is subtler and harder. The Chinese platforms understood that logistics density, payment and media sit in one loop, so a subsidy in one part of the loop buys behaviour that monetises in another. Amazon has the pieces. Most Western grocers and chains have them scattered across four vendors and three contracts, which is why their version of instant retail is a feature nobody remembers rather than a habit nobody breaks.

What to watch. Watch retention now that the coupons are thinning. The whole thesis rests on whether the habit outlives the discount. If second-half order volumes hold as subsidies fall, the giants have bought something durable. If they sag, they have rented attention at a ruinous price, and the analysts warning that users may not stay will have their answer.

The Roth Read. Stop asking whether you can afford one-hour delivery. Ask what habit you are willing to buy at a loss, and whether you own the loop that makes it pay you back later. China just proved the subsidy is the cheap part; the expectation it leaves behind is the asset, and right now your competitor is teaching your customer to expect something you cannot yet deliver.

The sleep app learned to buy. That is the whole game now.

9 September 2026

A sleep-tracking game now wants to do your shopping. Not point you to a shop. Do the shopping. That small, slightly absurd promise is the clearest picture yet of where retail is heading, and most retailers are not looking at it.

The app in question is a gamified sleep tracker, and the enthusiasm came from one of its users, a poster who wrote that they “absolutely love that my sleep app is now smart enough to be my own personal shopping assistant” and that “we shouldn’t have to close our game to go buy the things we need to sleep better.” The pitch, in their words: the cute AI agent can “figure out what we need, find the perfect cozy product, and buy it for us right inside the app.” The industry has a drier name for it. Agentic commerce. The user preferred “magic.”

Strip away the glowing shopping bags and the mechanism is stark. The app has your data, the app has your attention, and now the app proposes to have your wallet. Meta is building the same shape at the other end of the scale, with Muse, pitched as a personal AI agent to “get more done” across everyday tasks. A sleep game and a trillion-dollar platform are converging on one idea: the software that sits closest to you should also be the thing that buys for you.

Here is why it matters, and it is not the novelty. For thirty years the contest in retail was for the shelf, then for the search result, then for the feed. Each was a fight to be seen by a human who would then decide. The agent removes the human from the middle of that sentence. The sleep app does not show its user a page of pillows and mist diffusers to browse. It picks one. The moment of truth, the instant an impression becomes a purchase, moves from a shopper’s eye to a model’s judgement. And the model was trained, tuned and paid for by whoever owns the app.

Follow the incentives, because they are the story. When an agent buys “the perfect cozy product,” who defined perfect? The brand that optimised its product page for machine reading, as sellers on the ecommerce forums are already asking how to do. The brand that struck a commercial deal with the platform. The platform’s own private label. Perfect is a slot, and slots get sold. The retailer’s old question was how to rank on the shelf. The new question is what the agent believes about you, and what it costs to change that belief.

There is a harder edge underneath the cosiness, and it deserves naming. To buy for you, an agent needs your payment details, your address and standing permission to spend. One engineer, writing about giving an AI agent shell access, put it plainly: the agent “has everything you have because it is you” as far as the system is concerned. A sleep app that can charge your card while you sleep is a convenience and a surface for things to go wrong, in exactly equal measure. The trust you extend is not to a brand you chose. It is to an intermediary that chose for you.

China worked this out first, as it usually does. Alibaba and JD.com spent a decade collapsing discovery, payment and delivery into a single tap inside a super-app, so the distance between wanting something and owning it shrank to nothing. The West is now arriving at the same destination by a different road, through the AI agent rather than the super-app. The lesson is identical. Whoever owns the last decision owns the margin.

What to watch. Watch for the first agent that buys against its user’s stated wish, quietly steered by a commercial arrangement the user never saw. That is the moment the debate stops being about magic and starts being about disclosure, and it is coming sooner than the glowing shopping bags suggest.

The Roth Read. If you run a brand, stop optimising the page a person reads and start optimising the answer a machine gives. Your next buyer does not have eyes, a budget it can be tempted past, or a reason to remember you fondly. It has permissions, a checkout, and whatever the platform told it about you last.

You gave the agent hands. Did you notice it also has your keys?

3 September 2026

Three REST endpoints. Twenty million SKUs. And, if you are not careful, the run of your entire home directory.

The promise being sold for agentic commerce this week is that your shopping bot has a brain and now needs hands. Nobody is putting on the slide what those hands can reach.

The brain-and-hands line comes from CloudStore AI, whose promotion promises to turn any shopping agent into what it calls “a doer”: catalogue, checkout and logistics across 400-plus merchants and 20 million-plus SKUs through three endpoints. It arrives in the same month that Cloudflare launched, on Fortune’s reporting, a permanent identity and wallet for AI agents, with optional guardrails: spending limits and a whitelist of merchants where your agents are allowed to shop. Cloudflare’s own executive told Fortune the first wave will be developers and AI firms buying data, with ordinary consumers a second wave still to come.

Hold those two next to a quieter one. A developer, writing up an afternoon of paranoia, described running a shell tool for his coding agent and only then stopping to ask what “give your AI agent a shell” means at the level of the operating system. His answer, in his own words: the tool “has everything you have because it is you.” SSH keys, cloud credentials, the whole writable home directory, no audit trail. The post is titled, plainly, “AI Agent Has Root”.

Put the three together and you have the real shape of agentic commerce. Not a smarter shopper. A new account holder at the checkout who is not a person.

I have given that instant a name: the Machine Moment of Truth. P&G’s A.G. Lafley gave us the First Moment of Truth at the shelf in 2005. Google’s Jim Lecinski gave us the Zero Moment of Truth at the search results in 2011. Both belonged to the shopper. A hand on the pack, eyes on the ten blue links. The Machine Moment of Truth is the first one that does not. The machine hands the buyer no menu to judge. It returns a verdict, delivered with certainty, and the buyer takes it as the answer. It is the moment the buyer stops choosing and the machine chooses for them.

That is why the hands matter more than the brain. For more than a century the shopper on the other side of your checkout was a human being with a human’s frictions: a moment of hesitation, a second thought at the payment screen, a weakness for a well-placed offer. Retail was built to work on that hesitation. The agent has none of it. It does not linger, it does not take the extended warranty, and it does not forgive a clumsy returns policy. It executes. Every pound spent on persuading a person at the point of sale is aimed at a moment that is quietly moving out of reach.

Now follow the incentives, because that is where the story always lives. Whoever issues the wallet and holds the identity sits between the shopper and every merchant on the whitelist. That is not a payments feature. That is the introduction, owned. Cloudflare is not building a shop. It is building the thing that decides which shops an agent is even permitted to enter. The merchant that is not on the list does not lose the sale. It never gets asked.

The security point is not a footnote. It is the commercial risk. A retailer taking agent traffic is accepting orders from software that, on the developer’s own account, may be running with the full permissions of whoever deployed it. A compromised agent does not abandon a basket. It empties one, at machine speed, across every merchant it can reach, and the fraud desk built for stolen card numbers has never seen that pattern. The limits and the whitelist are not consumer niceties. They are the seatbelts, and they are optional.

Watch who gets to sit in the wallet layer, because that is the new gatekeeper. Watch, too, whether the standards emerging in the West borrow anything from China, where Alipay and WeChat Pay proved long ago that whoever holds identity and settlement holds the ecosystem. The West is about to relearn that lesson through a bot instead of a person.

The Roth Read. Stop asking whether your store is ready for AI shoppers. Ask the colder question: when an agent arrives at your checkout carrying your customer’s credentials and possibly root on its own machine, do you know whether it is friend or foe, and who told you so. The Machine Moment of Truth is already happening, in answers you cannot see, at a speed you cannot interrupt. The hands are here. Decide now whose keys they hold, because the merchant who waves them through blind will not lose a sale. They will lose control of the counter.

Your loyalty card was built to persuade a person. Soon it must persuade a machine.

26 August 2026

For thirty years, the loyalty card had one job: to nudge a human being. Earn, save, redeem, come back. Now a colder reader is arriving at the counter, one that does not feel loved and cannot be flattered. The question is no longer whether your programme moves a shopper. It is whether it moves an algorithm.

That is the argument running through a set of recent pieces on where loyalty is heading. Writing in Inside Retail, the analysis is blunt: programmes built to influence human decision-making may now also need to influence machine decision-making, because an AI assistant weighing several retailers on a customer’s behalf will consider loyalty benefits alongside price, convenience and availability. The same study found 80.4 per cent of Australian retailers naming loyalty a strategic priority for the next 12 to 18 months, and 57.1 per cent still describing their loyalty capability as maturing. In Forbes, Len Covello of Engage People puts the shopper’s side plainly: “It’s not trophy value anymore. This is currency, and it’s something I expect to have utilization with.”

Hold those two shifts together, because they are the same shift seen from two ends. Points are becoming spendable money, and the thing deciding where they get spent is increasingly software.

Here is why it matters, and it is not the part the headlines reach for. The romance of loyalty was always the emotional bit: the tier, the badge, the feeling of being recognised. A machine strips that out. It does not care that you are Platinum. It cares whether Platinum can be read, priced and applied inside the answer it is about to give. As Denise Holt of Phaedon argues in Loyalty Magazine, the first piece of work is plain: your loyalty value has to be legible to the assistant at the moment it is comparing options. Legible. Not lovely. Legible.

That single word rewrites the brief. For years the loyalty team optimised for feeling. Now it must optimise for a data contract. Can an agent see the points balance without a human logging in? Can it tell that 4,000 points knocks a real number off a real basket, today, at checkout? Can it apply status the way it applies a coupon? If the answer is no, your programme is invisible at exactly the moment the sale is decided. The store did not lose the customer. It lost the introduction.

And there is a trap on the other side. The tempting response to a machine that shops on price is to feed it discounts. Monocle warns where that ends: perpetual 15 per cent off is not a loyalty programme, it is a subsidised promotion in a loyalty costume, training your best-looking cohort to carry your worst margins. Hand an agent nothing but a discount and you have taught it to treat you as the cheapest tab, not the preferred one. The moment your only signal is price, you have volunteered to be a commodity.

This is the machine moment of truth arriving in the one place retailers thought they owned outright: their own members. The relationship you spent a decade and a marketing budget building now has a translator sitting between you and the shopper, and the translator only speaks in structured data and applied value.

What to watch. Watch for the first retailer whose points become natively spendable inside an AI assistant’s answer, the way Engage People’s Access Plus already links balances to checkout at Amazon, BP and PayPal. When a balance is a payment option an agent can reach for without a human clicking, the programmes that stayed a walled garden of emotional tiers will find the agent simply reads past them.

The Roth Read. Stop asking whether your customers love your loyalty programme. Start asking whether an algorithm can read it, price it and spend it in the three seconds it takes to answer “what should I buy.” A reward a machine cannot see is a reward you are no longer giving, and the emotion you built the whole thing on is the first thing the machine throws away.