For years, the pitch for AI pricing tools sounded like a growth story. Smarter offers, better margins, less money left on the table. Last week a federal regulator read the same pitch and heard a different story.
What happened
On 28 October the Federal Trade Commission issued orders to nine companies: Delta, United, JetBlue, Lyft, Uber, DoorDash, Instacart, Target and Walmart. The Record reports that the agency wants to know whether the price of "a ticket, ride share, food delivery or other product or service varies from one person to the next," based on things like loyalty programs or "analysis of personal data by superintelligence algorithms." For the delivery and rideshare companies, it also wants the "inputs, tools and procedures" behind how drivers are paid.
The companies have 45 days to respond. It's worth being precise about what this is. As Recording Law's breakdown of the orders puts it, "This is a study, not an enforcement action." Nobody has been accused of anything.
That distinction matters legally. It matters much less to a headline.
The question just went mainstream
Chairman Andrew Ferguson's statement, quoted by PYMNTS, set the terms: "Consumers deserve to know whether companies use their web surfing habits or other personal data to estimate the highest price they can charge each consumer."
That one sentence is now the frame. CNBC's report, drawn from Reuters, noted the timing, a week before the US congressional elections, with voters ranking the cost of living as their top issue. It also reported that state lawmakers have been moving to ban personalized pricing, and that the National Retail Federation has defended data-driven discounts as a way to reward loyalty.
So there are already two stories competing for the same facts. One says algorithms find the most each person will pay. The other says data helps the right customer get the right discount. Which one a reporter tells about a given company depends a lot on what that company has already said in public.
Why this lands on tech companies, not just retailers
The FTC's statement noted that its 2024 study looked at the middlemen, the consultancies selling these tools. This time it went straight to the companies that set the prices.
That moves the pressure in two directions at once. The brands now have to explain systems they often bought rather than built. And the vendors who sold those systems are about to get calls from customers asking what they can safely say, and from reporters asking who else uses the same tool.
If you build pricing, personalization, loyalty or offer-optimization software, your marketing is now research material. Case studies, conference talks, investor decks, help docs and old blog posts about "willingness to pay" will all be read through the FTC's sentence, not yours.
Where AI companies get this wrong
Letting the growth pitch become the public pitch. The language that wins a sales meeting ("capture more margin per customer") can read very differently next to a regulator's quote. Your customers' buyers and a consumer reporter shouldn't be getting the same deck.
Assuming "we're just the vendor" is a story. It's a fact, not a story. Reporters will want to know what the tool can do, not only what a client chose to do with it. If you can't explain that plainly, someone else will explain it for you.
Telling three versions of the truth. Sales says the product personalizes. The policy team says it doesn't set individual prices. The founder says something else on a podcast. Inconsistency is what turns a routine inquiry into a feature story.
Hiding behind complexity. "It's a sophisticated model with many inputs" is the opposite of what the moment needs. The FTC used the word "opaque." Sounding opaque in an interview proves its point.
Forgetting the other side of the marketplace. For gig platforms, the inquiry is as much about pay as price. Any company whose software shapes what workers are offered should expect that question too.
Getting ahead of your lawyers. Anything touching an active regulatory inquiry needs legal review before it goes out. What you can say, and when, is their call. How clearly you say it is ours.
Why it matters now
The companies named have about six weeks. Everyone else in the category has a quieter window to decide how they want to be described before the next round of coverage, the study's eventual findings, or a state bill pulls them in.
The best position isn't silence, and it isn't a pre-emptive press release. It's knowing, before anyone asks, how to explain in one honest paragraph what your system does with a customer's data, and being able to say the same thing on every channel.
The algorithm got harder to explain. Your story doesn't have to.
Building or selling AI that touches price, offers or pay? Request a briefing.