DELTA STICKS THE LANDING: TCPA Suit Dismissed at The Gate.

There’s a growing category of TCPA cases that seem less about unwanted telemarketing and more about frustration with how a system responded. Courts are starting to draw that line more clearly—and a new decision out of the Eastern District of New York is a clean example of where that line sits.

In Sait Kurmangaliyev v. Delta Airlines, Inc., a pro se plaintiff brought a TCPA claim against Delta Airlines based on a text exchange with its virtual assistant. The underlying interaction wasn’t marketing. It wasn’t outreach. It wasn’t even initiated by Delta. The plaintiff texted Delta first, asking to cancel a flight and obtain a refund. The system responded exactly as you would expect a customer service bot to respond—by pulling up the reservation, explaining the fare rules, and ultimately telling him no refund was available.

At that point, the conversation should have ended. Instead, the bot followed up with a standard satisfaction survey—asking the plaintiff to rate his experience from one to five.

That’s where things went sideways.

The plaintiff responded with “STOP.” The system didn’t recognize it in that context and repeated the prompt. He then responded “5.” The system followed up again. He sent “STOP” once more. The system again failed to interpret it and continued the scripted interaction until the conversation eventually expired. Weeks later, the plaintiff texted “STOP” again, and Delta responded that the interaction was not a subscription service and no further messages would be sent.

From those few messages—five, in total—the plaintiff built a TCPA claim.

The court didn’t need much time to dismantle it.

The plaintiff brought his claim under § 227(b), which requires, at a minimum, that the defendant used either an automatic telephone dialing system or an artificial or prerecorded voice. The problem was that the complaint did not plausibly allege either.

On the “voice” side, the issue was straightforward. The plaintiff’s theory rested entirely on text messages. But as the Second Circuit has already clarified in Soliman v. Subway Franchisee Advertising Fund Trust Ltd., the TCPA’s reference to “voice” means exactly what it says – something audible. Text messages do not qualify. That alone foreclosed one half of the claim.

The ATDS theory didn’t fare any better. The complaint was devoid of factual allegations suggesting that Delta used equipment capable of generating or dialing numbers randomly or sequentially. There was no description of the system, no explanation of how messages were sent, and no factual content connecting the chatbot to the statutory definition of an autodialer. Instead, the plaintiff relied on the kind of bare, conclusory assertions that courts have repeatedly rejected, particularly after Facebook, Inc. v. Duguid.

And that deficiency was fatal.

Even though TCPA claims are not subject to heightened pleading standards, courts still require some factual basis to support the elements of the claim. As the court emphasized, it is not enough to simply invoke the statute or restate its language. There must be “some indicia” that an ATDS was used. Here, there was none.

But what makes this decision particularly useful for defendants is the court’s focus on context.

This was not a case involving unsolicited outreach. It was not a case involving marketing messages sent to a cold lead. It was a consumer-initiated interaction. The plaintiff voluntarily texted Delta, engaged with the system, and continued responding, even after the system failed to interpret his inputs as intended. One of the messages he challenged was even sent after he reinitiated contact nearly forty days later.

That matters, because the TCPA is fundamentally concerned with unwanted, intrusive communications—not routine back-and-forth exchanges triggered by the consumer himself. As the court echoed, quoting Duguid, expanding the statute to cover any system that merely stores and sends messages to known numbers would effectively turn the TCPA into a blunt instrument a “chainsaw”, when Congress intended something far more precise.

There was also a procedural wrinkle that didn’t help the plaintiff. In opposing the motion to dismiss, he failed to meaningfully address Delta’s arguments regarding the absence of an ATDS or artificial voice. The court treated that silence as a concession, noting that even pro se litigants cannot survive dismissal without engaging with the substance of the defendant’s position.

Taken together, the outcome was inevitable. The complaint lacked factual allegations supporting the core elements of a § 227(b) claim, relied on a legally foreclosed theory regarding “voice,” and arose from a set of facts that fall well outside the TCPA’s intended scope.

For defendants, this decision is another helpful reminder that courts remain willing to dispose of weak TCPA claims at the pleading stage, particularly where plaintiffs attempt to stretch the statute to cover ordinary customer service interactions. For plaintiffs, it underscores a point that continues to come up post-Duguid: alleging that a system is “automated” is not enough. The statute has a definition, and courts expect plaintiffs to engage with it.

And stepping back, the case highlights something broader about where TCPA litigation is heading. As more companies rely on chatbots and automated messaging systems for customer service, courts are increasingly being asked to distinguish between nuisance telemarketing—the harm the statute was designed to address, and routine, user-driven interactions that simply don’t fit that mold.

This case lands firmly on the latter side of that line.

Because at the end of the day, the TCPA is not a vehicle for litigating customer service frustrations, even when the bot doesn’t understand “STOP.”