Customer Experience

AI age verification opens self-service opportunities

Retailers, regulators and technology providers are deploying systems that can estimate a customer's age in seconds, determine whether additional identification is needed and connect the decision to a transaction.

Image: Adobe stock

July 21, 2026 by Richard Slawsky — Editor, Connect Media

Summarize withChatGPTGrokPerplexityClaude

Artificial intelligence has already transformed self-checkout, digital signage and customer service. Now it is poised to reshape another corner of self-service technology: age verification.

Once viewed primarily as a tool for alcohol sales, AI-powered age estimation is finding applications across kiosks and unattended retail. Retailers, regulators and technology providers are deploying systems that can estimate a customer's age in seconds, determine whether additional identification is needed and connect the decision to a transaction.

The trend is gaining momentum as governments tighten restrictions on age-sensitive products while businesses look for alternatives to labor-intensive manual checks. For kiosk deployers, the opportunity is not simply to add another peripheral. It is to build flexible verification systems that balance compliance, privacy and customer convenience across multiple vertical markets.

From alcohol to a broader range of products

Alcohol remains one of the most established uses of automated age verification. Systems deployed in Europe, Australia and North America analyze facial characteristics to determine whether a customer appears to be above a predefined threshold. If the customer is clearly above that threshold, the transaction can continue. If the result is uncertain, the system can request an identification document or employee assistance.

TendedBar, which provides automated beverage-dispensing technology, has found that alcohol compliance cannot be addressed with a single operating model.

"Alcohol compliance is one of the most important parts of the TendedBar platform," TendedBar co-Founder and CEO Justin Honeysuckle said in an email interview. "We have learned that there is no one-size-fits-all answer because requirements can vary by state, venue type, service model and local regulations."

TendedBar is designed to work with different levels of verification depending on the use case, Honeysuckle said. The biggest lesson from regulated environments is that the technology has to be flexible.

"Some venues want maximum automation," he said. "Others need a staff-assisted model. Our job is to provide the right operating structure for the location while keeping compliance, accountability and customer experience at the center of the deployment."

The technology is also moving beyond alcoholic beverages. Lottery terminals are adopting age-checking capabilities as jurisdictions try to prevent underage gambling without requiring constant employee supervision. Vape products, nicotine pouches and electronic cigarettes are another focus as retailers face increasing scrutiny.

Pennsylvania Lottery documentation published in January describes a wireless barcode reader that checks the age encoded on Pennsylvania driver's licenses and identification cards. The agency said the equipment can help retailers screen lottery customers and may also be used when selling tobacco and alcohol.

Proposed rules and legislation provide additional signs of demand. South Dakota Senate Bill 221 would require nicotine vending machines to include age-verification systems and require customers to complete verification before buying. A 2026 Georgia Department of Revenue proposal would clarify that tobacco vending customers must be 21. Point-of-sale solution provider Modisoft has announced support for a 2026 Altria Group Distribution Co. program that includes ID-verification technology for tobacco retailers.

Cannabis presents another long-term opportunity. As legal recreational markets expand, automated dispensaries and self-service kiosks must comply with strict purchasing-age rules. AI age estimation can reduce the number of customers asked to produce identification, although higher-risk transactions may still require document authentication and a facial match.

Other candidates include knives, fireworks, high-caffeine energy drinks and certain medications.

Recent regulatory action in the U.K. illustrates how quickly the potential market can expand — and how regulation can limit it.

The U.K. Department of Health and Social Care announced that England plans to ban the sale of high-caffeine energy drinks from all vending machines beginning in April 2027, subject to parliamentary approval, arguing that enforcing age restrictions on unattended equipment would be difficult. The Vending & Automated Retail Association criticized the decision, stating more than 80% of vending machines are in workplaces, factories, warehouses and gyms where children generally are not present.

The association also argued that the government overlooked alternatives such as AI-powered age checks and location restrictions. Whether the policy is reconsidered or not, the dispute illustrates a central issue for the industry: Technical capability does not automatically produce regulatory acceptance.

How automated verification works

AI age estimation generally does not attempt to establish a person's identity. A camera analyzes facial characteristics and determines whether the individual appears above or below a threshold. Retailers commonly set that threshold higher than the legal purchasing age. An alcohol retailer, for example, might automatically approve someone estimated to be older than 25 and ask a younger-looking customer for identification.

Swiss technology provider Scandit's Age Verified Self-Checkout, rolled out in early July, uses a customer's smartphone as the verification device rather than adding a scanner and camera to every kiosk. When a restricted product is scanned, the self-checkout displays a QR code. Scanning it opens a browser-based verification page, so the customer does not need to download an app.

The customer first takes a selfie, which is analyzed on the phone to estimate age. A result above the retailer's threshold sends a pass signal to the checkout. If the estimate falls below the threshold, the customer is prompted to scan an ID. The software reads the document and compares its photograph with the selfie, helping detect a minor using someone else's identification.

Scandit said the verification normally takes less than 10 seconds and can resolve about 80% of age checks without an employee. The company said all biometric and identification processing occurs on the user's device and that neither Scandit nor the retailer receives or stores that data. The kiosk receives only the pass-or-fail result.

Scandit also says the system complies with data-protection and security standards including the General Data Protection Regulation, the California Consumer Privacy Act, ISO 27001 and the United Kingdom's Challenge 25 certification.

For kiosk manufacturers and software developers, one of the greatest attractions is that automated age verification is largely product-agnostic. A camera, processor, document reader or phone-based workflow can support compliance for numerous categories.

That flexibility could allow operators to use a common verification layer across a diverse kiosk or automated-retail fleet instead of developing a different compliance system for every product category. The strength of the check could then be adjusted according to the product, location and applicable law.

"Today, retailers are forced to choose between absorbing the ongoing labor cost of manual checks or investing in proprietary hardware, neither of which scales efficiently," said Christian Floerkemeier, CTO and co-founder of Scandit, in a press release. "Age verified self-checkout removes this tradeoff by automating compliance in a fast, secure and hardware-free way, which reduces intervention and optimizes labor."

Operations determine whether the technology succeeds

Traditional age checks can also interrupt the customer experience. Waiting for an employee to unlock a cabinet, inspect identification or approve a transaction undermines one of self-service's principal advantages: speed. Automated estimation allows many transactions to proceed without assistance while escalating uncertain cases.

Amy Pethick, head of retail at TPP Retail, previously led self-checkout projects at BP, including the rollout of systems capable of processing fuel purchases along with conventional retail transactions.

"Age verification technology has the potential to remove one of the biggest friction points in self-service, but its success depends on more than the technology itself," Pethick said in an email interview. "Retailers need to balance customer convenience with compliance, accuracy and colleague confidence."

"From my experience, the most effective solutions are those that integrate seamlessly into the wider store operation," Pethick said. "If colleagues are constantly required to intervene because confidence thresholds are set too low or exception handling is poorly designed, the productivity benefits quickly diminish."

That makes exception management as important as the primary verification step. Deployers must decide what happens when a customer has no smartphone, the camera cannot produce a reliable estimate, the ID is damaged or unsupported, the network is unavailable or an employee is not immediately available. They must also determine how failures are logged, how disputes are handled and who remains accountable for an improper sale.

"As retailers expand self-service into convenience, kiosks and unattended environments, age verification should be viewed as part of the overall customer journey rather than a standalone compliance tool," Pethick said. "The goal is to reduce friction while maintaining trust and ensuring regulatory requirements are consistently met."

Privacy and accuracy remain concerns

AI age estimation is unlikely to eliminate traditional identification checks. Retailers must test accuracy across ages, skin tones, facial characteristics and lighting conditions. They must also guard against presentation attacks, borrowed IDs and false rejections that inconvenience lawful customers.

Privacy advocates continue to scrutinize facial analysis, making clear disclosure and data-minimizing design essential. Systems that estimate age without identifying the customer may present fewer privacy concerns than systems that retain an identification image or biometric template, but deployers still need to explain what is processed, where processing occurs and whether anything is stored.

Legal acceptance varies by jurisdiction and product. Some laws require physical identification or prohibit vending in locations accessible to minors regardless of the machine's technology. Others permit automated checks but impose requirements involving signage, audit records, staff oversight or approved forms of identification. Deployers therefore need a configurable system and a jurisdiction-by-jurisdiction compliance review.

As governments extend age restrictions to additional products, demand for automated verification is likely to grow. What began largely as an alcohol-sales application is expanding into lottery, cannabis, nicotine products, regulated retail merchandise and potentially energy drinks.

The winners will not necessarily be the systems that impose the strongest possible check on every customer. They will be the ones that match the level of verification to the product and location, integrate cleanly with the transaction and exception workflow, minimize the collection of personal data and give regulators and operators a defensible record of compliance.

About Richard Slawsky

---