After a decade of making our digital lives frictionless, artificial intelligence is finally moving beyond screens to reshape how we move through airports, stores, and cities, and it could render physical credentials entirely obsolete.
The digital world already knows who you are. Your phone recognizes your face, remembers your preferences, and anticipates what you want to buy before you have thought about it. Yet the moment you step into an airport, an office building, or a hospital, you are essentially transported back to the 1990s. You fumble for a boarding pass, swipe a badge, or show an ID card to a security guard who glances at it for half a second.
This jarring gap between digital intelligence and physical infrastructure is starting to close. As Fortune recently explored, a convergence of mature computer vision technology, scalable AI systems, and shifting consumer expectations is making it possible for physical spaces to recognize and serve people without requiring them to repeatedly prove their identity. The concept, sometimes called the "Recognition Economy," moves verification away from objects like keys and cards and toward the individuals themselves. For startups and incumbents alike, the implications are substantial.
Physical access and commerce have always relied on tokens. Keys, tickets, badges, and cards are all proxies for identity, stand-ins designed to authorize a transaction or grant entry. The system works reasonably well, but it carries fundamental flaws. Credentials get lost, stolen, copied, or forged. The global identity fraud market alone costs businesses and consumers tens of billions of dollars annually, with the Federal Trade Commission recording over $10 billion in fraud losses in the United States in 2023 alone. Much of this fraud persists precisely because identity is mediated through transferable objects rather than anchored to the person standing in front of you.
Meanwhile, there are an estimated one billion surveillance cameras deployed globally, a vast majority of them sitting on networks in commercial buildings, retail environments, and transit systems. For years, these cameras served limited purposes: recording footage for security review, perhaps feeding basic analytics. Computer vision has advanced to the point where existing camera infrastructure can be leveraged as a sensor network for real-time, presence-based identification and transaction processing. The hardware is largely already in place.
From Metropolis to the Broader Built Environment
Metropolis, the parking and mobility platform, offers an early blueprint. The company started by deploying computer vision in parking facilities so drivers could enter and exit without rolling down a window or pulling out a ticket. The system recognizes the vehicle and charges the associated account automatically. It is a focused application, but the underlying architecture is designed to be universal. Restaurants, hotels, stadiums, office towers, healthcare facilities, and transit hubs all operate on the same fragmented logic of repeated credential verification.
Consider the modern airport experience. A traveler verifies their identity at curbside check-in, again at terminal security, again at the boarding gate, and again at a lounge desk. Each checkpoint exists because identity data sits in siloed systems that do not communicate with one another. A unified recognition layer could allow identity to flow securely across the entire environment. Security would not be weakened; in fact, continuous verification across multiple touchpoints is arguably more robust than a single glance at a driver's license by a tired contractor at 6 a.m.
Why the Timing Matters Now
Three forces are making this shift realistic today rather than a distant ambition. First, AI systems have reached a level of reliability that allows them to operate in complex, unpredictable physical environments rather than only in controlled digital ones. Computer vision models that once struggled with variable lighting, angles, and occlusions are now commercially viable at scale. Second, the hardware layer is already deployed. The camera networks embedded in physical spaces around the world can be upgraded through software rather than requiring expensive retrofits. Third, consumers have been conditioned to expect seamless experiences. People who open their phones to instant facial recognition and walk out of an Amazon Go store without checking out are not going to remain patient with manual badge checks and paper tickets indefinitely.
History offers a clear pattern here. Transformational technologies do not simply make existing systems faster. They render them unnecessary. The printing press did not speed up scribes. GPS did not improve paper maps. Each leap made the previous baseline irrelevant. Presence-based identity and transaction systems have the same potential for physical commerce and access control.
For startups building in this space, the opportunity is significant but the challenges are real. Privacy concerns are the most obvious obstacle. Any system that identifies people in physical spaces by definition collects sensitive biometric data, and regulatory frameworks like Illinois' Biometric Information Privacy Act and the EU's GDPR impose strict requirements on how that data is handled. Companies that get the privacy architecture right will have a durable competitive advantage. Those that do not will face costly lawsuits and eroded trust.
The physical world is the last major environment where AI has not yet fundamentally reshaped how we interact with infrastructure. That is changing quickly. The companies that figure out how to make presence-based systems reliable, private, and invisible will define the next era of physical commerce. Watch the airports, stadiums, and office towers. The first deployments there will tell you how fast this future arrives.
