The entire payments industry is organized around one moment: the transaction. Analysts have been predicting that moment will stop existing. So what will the physical world do when it can no longer rely on checkout to know who is standing inside it? It is already building new presence infrastructure. There are four questions you should you should ask when you are evaluating technologies that can verify presence.
For decades, the payments industry has been organized around a single moment. The tap, the swipe, the scan, the click. Every major player in the ecosystem has built its business on owning some piece of what happens when a consumer is ready to pay. Apple Pay owns the phone. Visa owns the rail. PayPal owns the redirect. Buy-now-pay-later providers own the financing screen. All of them compete for the same sliver of time: the last few seconds before money moves.
eMarketer’s report, “The Next Era of Payments,” argues that this sliver of time is about to collapse. Digital wallets have become a commodity. By 2029, more than 207 million Americans will use them. Tapping to pay, storing credentials, clicking through a one-step checkout have stopped being differentiators and become table stakes. The next phase of growth, the report argues, will come from wallets that move beyond the transaction entirely, orchestrating commerce so seamlessly that the checkout step simply disappears from the consumer’s experience.
Most of the industry will read that prediction as good news, a bigger pie and a longer runway. But there is another way to read it, and it is considerably less comfortable: if checkout disappears, then every company that built its competitive position around owning checkout and using the checkout as a deterministic measurement signal is about to discover that the thing it owns no longer exists.
What will the moment of transaction be replaced by?
Invisible, ambient commerce already exists, and it reveals a pattern.
Several versions of ambient commerce are already running in production, and each one reveals the same structural dependency.
Amazon pulled its Just Walk Out technology from its own grocery stores in 2024, but the same system is reportedly operating in hundreds of third-party locations across airports, stadiums, and convenience stores. A customer enters, picks up what they want, and leaves. There is no register, no checkout screen, no line. The transaction settles automatically because the system can identify exactly who is inside the store and what they took. If the system cannot do that, the entire model collapses. The commerce is ambient only because the presence verification is deterministic.
Stadiums have begun deploying facial authentication for both entry and concessions. Enrolled fans walk through gates and purchase food without pulling out a phone, a wallet, or a ticket. The environment recognizes the person and charges them. The technology has crossed millions of uses across deployments and continues to expand. Once again, the transaction is triggered not by an action the consumer takes, but by the fact that the system can confirm exactly who is standing in front of it.
Gateless parking follows the same logic. License plate cameras identify a vehicle on entry, and the fee debits automatically when the vehicle exits. No ticket, no barrier arm, no app. The system knows the car is physically at the exit point, and that knowledge alone is what triggers payment.
These are different technologies serving different environments, but they share a single structural truth: none of them work unless the environment can verify, with certainty, that a specific person or vehicle is physically present at a specific place at a specific moment. That verification is the foundation of ambient commerce. Without it, there is no transaction to trigger, no account to charge, no experience to make seamless. Presence verification is the load-bearing layer underneath all of it.
The infrastructure for Phase 3 — ambient commerce — does not exist at scale.
eMarketer’s report maps three phases of wallet evolution. In the first phase, wallets become commerce enablers by surfacing loyalty programs, embedding financing, and influencing where consumers shop.
In the second phase, wallets become AI-driven orchestrators that choose the optimal payment method, apply rewards, and resolve disputes automatically.
In the third phase — ambient commerce — wallets become commerce environments themselves, invisible, triggered by context rather than by any deliberate action the consumer takes.
Ambient commerce is the phase where checkout disappears. It is also the one with a dependency that the broader infrastructure conversation has not caught up to: every ambient commerce scenario assumes the system knows, with certainty, where the consumer is. The stadium charging for a beer has to verify that the fan is standing at the concession counter and not sitting three sections away. The parking system has to confirm the right car at the right exit. The walk-out store has to match a specific person to a specific set of items.
The physical environment, as it exists today, has no reliable and scalable way to do this. Everything the industry is building toward in Phase 3, the AI orchestration, the contextual triggers, the invisible settlement, is software waiting on an infrastructure foundation that has not been poured.
The 4 questions every presence technology has to answer.
Several technologies can verify presence today, and each has earned its place in specific environments. The way to evaluate them — the way to evaluate any presence infrastructure, including ours — is to ask 4 questions:
1 Signal certainty. Can it confirm that a specific person is at a specific place with enough confidence to move money on? Ambient commerce settles transactions on presence alone. A probabilistic signal is not a smaller version of the answer. It is a different product.
2 Regulatory exposure. Does the method depend on data that legislators are actively restricting — biometrics, precise location — or does it operate outside those statutes entirely? Infrastructure is a decade-long bet. A method that is legal today and banned in three states by next year is not infrastructure.
3 Cost structure. Does the economics pencil at your revenue per square foot, on hardware the environment already owns? Or does it require a capital program per site that only the highest-margin formats can absorb?
4 Reach. What fraction of the people who actually walk through the door can it verify — including the guests who never enrolled, never downloaded, and never will?
Run the existing approaches through those 4 questions, and the pattern shows itself quickly.
Computer vision, the approach behind Amazon’s Just Walk Out, clears the certainty bar. It fails on cost structure. It requires dedicated camera infrastructure covering every angle of every aisle, and Amazon proved the point twice: first that the technology works, then that the per-store economics do not hold at the margins grocery operates on. The technology is thriving in small-format, high-revenue-per-square-foot environments like airports and stadium concessions, where the hardware investment is justified. It is not a viable path for the hundreds of thousands of retail environments where the math does not pencil out.
Facial recognition also clears the certainty bar, and fails on reach. It requires consumers to voluntarily enroll their biometric data, which limits it to environments where the relationship between the business and the consumer is strong enough to make that ask feel reasonable. A stadium with season-ticket holders can do it. A corner store, a transit system, a hotel lobby cannot.
Both approaches then fail a second time, on regulatory exposure — and this is the axis the industry is most consistently underweighting.
On the location data side, a growing number of states are moving to ban the sale of precise location data outright. Connecticut has already passed such a law, and Vermont is close behind. By the end of 2026, California, Massachusetts, and Washington could join Maryland, Oregon, Virginia, and others in imposing similar restrictions, bringing the total affected population to nearly a quarter of the country. Industry groups within the MMA and the NAI attempted to preempt legislation through voluntary codes of conduct for precise location, but those codes never advanced beyond the voluntary stage. The self-regulatory window has effectively closed.
On the biometric data side, the regulatory pressure is even more acute. Illinois’s Biometric Information Privacy Act has generated over a billion dollars in settlements, including Meta’s $650 million payout for collecting facial recognition data without consent. More than 20 states now classify biometric information, including faceprints and voiceprints, as sensitive personal data requiring affirmative opt-in consent before collection. Portland, Oregon has banned commercial facial recognition entirely in places of public accommodation such as retail stores, restaurants, and hotels, with penalties of $1,000 per day per violation. New York City held hearings in early 2026 on legislation that would extend similar bans to stores, restaurants, and apartment buildings.
Existing approaches clear only one or two items on the ambient commerce diligence checklist. But for an investment to operate at scale—across every store, transit gate, and venue—your tech stack must address all four underlying patterns at once.
The pattern is clear: the approaches that produce the strongest presence signal, cameras and biometrics, are the same approaches that regulators are actively moving to restrict. Any ambient commerce infrastructure built on watching consumers has a regulatory shelf life that is getting shorter every quarter.
Geofencing inverts the profile: broad reach, low cost, and a signal that fails the certainty question outright. GPS drifts five to ten meters and has no vertical resolution, which means it cannot reliably distinguish between adjacent stores in a shared retail center, let alone between aisle 3 and the parking lot. BLE beacons can work at closer range but require dedicated hardware at every touchpoint, bleed signal through walls, and the industry already spent a decade deploying and then ripping them out.
Every approach clears one or two of the four questions. Ambient commerce at the scale eMarketer describes — every store, every transit gate, every venue — needs an approach that clears all four at once. Keep those four questions. They are the diligence checklist for any presence-layer investment you will evaluate in the next five years, including the one this article is about to make the case for.
The presence layer has been sitting in your already installed speakers the whole time.
LISNR identified this gap a decade before the analyst community gave it a name and built Radius to close it.
The insight was simple: every store already has speakers, and every phone already has a microphone. Radius turns those existing speakers into secure handshake points by transmitting inaudible frequencies at 18 to 19 kHz. When a consumer with an opted-in device enters range, the phone and the speaker complete an encrypted exchange that verifies the person is physically present. The consumer does not have to tap a terminal or scan a code. The environment knows they are there because they opted into that capability through a relationship they already have, whether that is a merchant app, a loyalty program, or a wallet integration. Without that opt-in, no handshake occurs.
Hold Radius to the same four questions as everything else.
Signal certainty: the handshake is an encrypted, device-specific exchange between one phone and one speaker. It is deterministic — the same class of signal that makes walk-out stores and gateless parking possible, not a probabilistic inference about where someone might be.
Regulatory exposure: no cameras, no biometric enrollment, and no third party precise-location data of the kind the new statutes govern. The consumer’s opt-in is the trigger for every handshake, which means the compliance posture is built into the mechanism rather than bolted on after.
Cost structure: it runs on speakers the merchant already owns and has already powered on, deployed as software. The camera-based approaches are excellent in environments where the revenue per square foot justifies the investment and where regulators permit them. Radius exists for the vastly larger set of environments where neither condition holds.
That leaves the fourth question, reach — and here honesty matters, because reach is where every wallet-centric ambient commerce model has a problem, ours included.
1 Signal certainty with LISNR: the handshake is an encrypted, device-specific exchange between one phone and one speaker. It is deterministic — the same class of signal that makes walk-out stores and gateless parking possible, not a probabilistic inference about where someone might be.
2 Regulatory exposure with LISNR: no cameras, no biometric enrollment, and no third party precise-location data of the kind the new statutes govern. The consumer’s opt-in is the trigger for every handshake, which means the compliance posture is built into the mechanism rather than bolted on after.
3 Cost structure with Radius: it runs on speakers the merchant already owns and has already powered on, deployed as software. The camera-based approaches are excellent in environments where the revenue per square foot justifies the investment and where regulators permit them. Radius exists for the vastly larger set of environments where neither condition holds.
That leaves the fourth question, reach — and here honesty matters, because reach is where every wallet-centric ambient commerce model has a problem, ours included.
4 Reach with LISNR? That leaves the fourth question, reach — and here honesty matters, because reach is where every wallet-centric ambient commerce model has a problem, ours included. Wallets are not only competing with wallets but with other apps.
Ambient commerce has a second infrastructure gap that compounds the first.
The eMarketer model runs entirely through the digital wallet, and wallets are apps. That introduces a reach problem the ambient commerce conversation has largely ignored.
The average phone user lives inside roughly five apps. Global app downloads fell 2.3 percent in 2024 and 3.4 percent in the United States. The average app loses 77 percent of its daily active users within three days of installation and 90 percent within a month. The app economy has hit a ceiling that shows no sign of lifting.
Consider who actually walks through a merchant’s door on any given day. A small percentage are members: regulars, account-holders, loyalty enrollees. The rest are guests. A tourist riding a metro twice on vacation. A shopper who wanders into a store once. A fan attending a single game. A traveler buying one coffee. Guests represent the majority of physical-world interactions, and they are never going to download a native app for a single transaction. Not for two subway rides. Not for one purchase. The era when a business could reasonably expect consumers to download, install, register, and retain an app for occasional interactions is over.
If ambient commerce requires a native app on the consumer’s phone, then ambient commerce only works for the fraction of customers who already chose the merchant strongly enough to install its software. Everyone else, the guests, the one-time visitors, the consumers that app fatigue made permanently unreachable, simply falls through. When we ran Radius through the fourth question, that was the honest answer: deterministic presence, but only for consumers who already carried an app with the SDK inside it.
Because the Radius SDK can live inside the digital wallets consumers already use every day, it verifies guests without asking them to download a single new merchant app.
What to do with this on Monday.
If you run payments strategy for a wallet, a merchant, a transit system, or a venue, the eMarketer report is not a reading assignment. It is a capital allocation question. Three moves follow from everything above.
First, stop spending incremental dollars differentiating the checkout moment. Faster taps, cleaner one-step flows, another tenth of a second off authorization — that work is now maintenance, not strategy. The report’s own logic says the moment you are polishing is the moment that dissolves. Fund checkout like the commodity it has become and redirect the difference toward the layer that replaces it.
Second, audit your environments against the four questions before you commit to any Phase 3 roadmap. For each physical environment you operate or serve: Can presence be verified there with transaction-grade certainty? Will the method survive where the biometric and location statutes are heading? Does the cost pencil at that environment’s revenue per square foot, on hardware that is already installed? And does it cover the guests, not just the members? A Phase 3 vision that cannot answer all four is a demo, not a plan.
Third, sequence infrastructure ahead of orchestration. Phase 2 — the AI that picks the payment method, applies the rewards, resolves the disputes — is software. It deploys fast and it gets copied fast. The presence layer underneath Phase 3 is infrastructure: slower to put in place, and compounding for whoever places it first. The rational sequence is the uncomfortable one: pour the foundation now, while competitors are still polishing checkout, and start where the reach problem is most expensive — the environment where guests outnumber members the most. That is the pilot that tells you whether ambient commerce works for the customers you have, not just the members you have already converted.
The companies still fighting over checkout are optimizing for a moment that is about to stop existing.
Wallets will become the primary interface for how consumers shop, borrow, and manage money. Commerce will become ambient. The checkout moment, the sliver of time the entire industry has spent decades fighting over, will dissolve into the background of everyday life.
And ambient commerce is not the last stop. The next turn is already visible: agentic commerce, where the consumer no longer even initiates the purchase — an AI agent shops, chooses, and settles on their behalf. Everything this article traced gets sharper in that world. When the buyer is software, the one thing the physical world still has to confirm is the human: that the person the agent is acting for is actually standing at the gate, the counter, the curb where the transaction lands. Agents will run the digital half of commerce on their own. The physical half will still turn on presence. That shift deserves its own article, and it will get one.
The winners of what comes next will not be the companies that built the most elegant checkout experience. They will be the ones that built the infrastructure for what comes after: a layer that verifies presence with enough certainty to act, survives where the regulation is heading, runs on hardware the physical world already owns, and reaches the phones of consumers who never installed an app.
LISNR has been building that layer for a decade. The market the analysts just described is the market Radius was built for.
Sources: eMarketer, “The Next Era of Payments: How Digital Wallets Are Shifting From Payment Utilities to Orchestrators of Ambient Commerce,” April 2026.