It is twenty to eight in the morning and the line is four deep. A man in his late twenties puts a pack of cigarettes on the counter. The register throws up an age prompt. The clerk taps it, the drawer opens, the customer is gone, and the next person steps up. Nine seconds, start to finish.
In every record that store keeps, that was a compliant sale. The prompt fired. The clerk cleared it. The transaction closed clean, and the exception report at the end of the month will not say a word about it. The one fact nobody wrote down is whether an ID ever came out of a wallet.
The risk is not the judgment call. It is the ask.
It is easy to assume underage sales happen because a clerk studied a real ID and got the arithmetic wrong, or was handed a good fake and got fooled. That is the version in the training video. The field data says something simpler, and far more fixable.
Indiana runs unannounced inspections of tobacco retailers every year under the federal Synar program, sending young buyers in to attempt purchases. Prevention Insights at the Indiana University School of Public Health published the 2025 results on April 2, 2026. The statewide retailer violation rate came in at 10.5%, the lowest since 2013, down from 14.5% the year before. Good numbers, trending the right way.
Then there is the line in the same release that should stop any operator cold. When the clerk asked for photo identification, the violation rate was 1.59%. When the clerk did not ask, it was 88.89%.
Read that twice. Nearly the entire age compliance exposure of a store sits inside one binary act. Not judgment. Not document expertise. Not spotting a fake. Whether the question got asked at all. Indiana requires a retailer to request photo identification from anyone who appears to be under 30, and the clerks who did that essentially never failed. The clerks who skipped it failed nine times in ten.
Which is exactly why the register cannot tell you where you stand. An age prompt at the point of sale is a button. It records that a button was pressed. It does not know whether a card was produced, glanced at, or scanned, and it has no idea whether the person holding it was the person buying. Every store in the country is sitting on this data and almost none of it means anything.
What it costs when the inspector is the one who finds out
Federal law sets the minimum age at 21 for every tobacco product, and the FDA enforces it with undercover buys where nobody identifies themselves until it is over. The agency's published enforcement schedule, current as of March 31, 2026, escalates on a fixed ladder. A first violation draws a warning letter and no fine. A second inside 12 months costs $365. A third inside 24 months is $727, a fourth is $2,920, a fifth inside 36 months is $7,300, and a sixth inside 48 months is $14,602. The ceiling for a single violation of the tobacco rules is $21,903.
The fines are not really the danger. Five or more repeated violations inside 36 months lets the FDA pursue an order barring that location from selling tobacco products at all, for a period the agency sets. For a convenience store, losing the tobacco set for a stretch is not a fine. It is the category that brings people through the door, along with the fuel, the coffee, and the basket that arrives with them.
A second meter runs above the store. The Synar rule requires each state to hold its retailer violation rate under 20% or risk losing up to 10% of its federal block grant for substance use prevention and treatment. States drifting toward that line lean harder on enforcement, and it lands on retailers.
Alcohol runs on a separate track with the same shape, except the penalty often falls on a person rather than a permit. A national survey of enforcement agencies led by Darin Erickson, published in Alcohol Clinical and Experimental Research in 2014, found that 43% of local agencies penalize only the server or clerk for an illegal sale, 10% penalize only the license holder, and 47% penalize both. The same paper put successful underage purchase attempts at 26% to 39% across the cities studied, down from roughly 75% to 100% in the early 1990s. Better, and nowhere near solved.
The behavior is visible. It just is not recorded.
Here is the part worth sitting with. The act that decides everything, whether the clerk asked, happens at the counter, in frame, at a timestamp the register already knows to the second.
The regulator understands this. We Card, the retailer training program, notes in its walkthrough of the process that when a compliance check happens, the FDA records whether an ID was asked for as well as whether the product was sold. Its March 2021 summary counted more than 1.1 million FDA store inspections since 2010, with close to 88% recorded as no violation. So the inspector tracks the exact behavior the store's own systems do not. The store learns the outcome months later, in a letter.
Joining two sources the store already owns closes that gap. Take the age prompt event from the point of sale. Take the few seconds of camera around the same timestamp. Ask one question: did the ID come out. A prompt cleared in under two seconds with no ID motion at the counter proves nothing on its own. It is a coachable moment with a clip attached, available the next morning instead of after an inspector has been and gone.
The pattern matters more than any single instance. One skipped check is a busy Tuesday. The same clerk clearing age prompts in under a second, on the same shift, on the same category, four days running, is a habit setting. That is what you want to catch while it is still a habit and not yet a $7,300 letter with two more coming.
This is a training problem in a compliance costume
Worth saying plainly, because it decides whether a program like this survives contact with staff: almost none of this is dishonesty. Underage sales mostly happen because someone was moving fast on a rush, or was never told the store rule is to card anyone who looks under 30 rather than anyone who looks under 21, or has rung up the same regular for a year and stopped seeing a stranger. Treating every instance as an accusation is how you lose the clerks you already struggle to keep.
The useful version is narrow. Show the clerk the clip. Ask what happened. Retrain if it is a gap, coach if it is speed, escalate only if it continues after both. Keep the record either way, because if an inspector does turn up with a violation, a store that can show it was already monitoring and correcting this is in a very different conversation than a store holding a pile of prompt acknowledgements that prove nothing at all.
None of this needs new hardware. The cameras are up, the register already fires the prompt, and the schedule already says who was standing there. This is the kind of join we spend most of our time on at ARGUS, and it is less exciting than it sounds. Most of the work is getting four systems that never spoke to agree on what happened at 7:40 in the morning.
If you run stores and you are carding thousands of times a week with no way to know which of those actually happened, we would be glad to hear how you handle it now. ARGUS is in private beta. You can talk to us or write to business@useargus.co.