AI and Agentic Confusion Was Never About AI

ZoomInfo told its investors that customer growth had "regressed" because of "AI and agentic confusion."

That was weird. A company that sells go-to-market intelligence data to sales and marketing teams explains its own slowdown by saying its customers are confused about AI. The basic point is that this phrase only makes sense if you understand it as a cover story for a deeper problem: ZoomInfo's own AI products are cannibalizing the seat-based subscription model that has been the company's revenue engine.

Let's start with the timeline. Between November 2025 and February 2026, ZoomInfo executives repeatedly told investors that AI-integrated products were driving strong demand. CEO Henry Schuck cited "20%-plus operations growth" and improving net revenue retention. In February 2026, the company gave full-year revenue guidance of $1.247 billion to $1.267 billion, framing 2026 as a year of bringing its "all-in-one AI platform for go-to-market teams" to customers at scale.

Then, on May 11, 2026, ZoomInfo reported Q1 results and slashed that guidance down to $1.185 billion to $1.205 billion. Revenue for the quarter itself was fine — $310 million, a 35% adjusted operating income margin, both beating expectations. But the outlook told a different story. Customer growth had "regressed," Schuck said, due to this "trend of AI and agentic confusion" in late March and April. Customers were pausing their purchasing decisions. The stock fell $1.98 per share, a 33% drop, closing at $4.06 the next day.

To compensate, ZoomInfo announced a restructuring eliminating roughly 600 jobs — about 20% of its workforce — targeting $60 million in annual savings at a cost of $45 million to $60 million in charges.

Here's the part the "AI confusion" explanation doesn't fully capture. The lawsuit alleges something more specific: ZoomInfo was concealing the deterioration of its legacy seat-based subscription business while customers migrated toward consumption-based pricing models. The company was simultaneously building AI features that made its own per-seat pricing structure look increasingly awkward.

Think about the incentives. If you sell software by the seat — one license per user — and then you introduce AI agents that can do the work of multiple seats, you have a pricing problem. Your new products make your old revenue model look expensive. Your customers notice this. They pause. They ask whether they need as many seats when an AI agent can do the legwork.

That is what "AI and agentic confusion" means in the plumbing. It doesn't mean customers were philosophically unsure about artificial intelligence. It means customers were realizing they didn't need to buy the same way anymore, and ZoomInfo hadn't figured out how to charge them for that. The company is now transitioning from seat-based pricing to a consumption model based on "data, insights, applications, and agents." That transition is painful because it means the old revenue stream is decaying before the new one is proven.

There's another layer. The complaint alleges that customers weren't just pausing — they were building their own internal AI-driven go-to-market tools instead of buying ZoomInfo's products. That's the competitive threat disguised as a demand-cycle story. When management says "confusion," the economic reality may be closer to "we're being displaced by cheaper alternatives, some of them internally developed."

The lawsuit, captioned Tejada v. ZoomInfo Technologies, names CEO Henry Schuck and CFO M. Graham O'Brien as individual defendants under Section 20(a) of the Securities Exchange Act, alleging they're "controlling persons" who knowingly promoted optimistic metrics while concealing adverse facts. They're also accused of violating Sarbanes-Oxley Sections 302 and 906 by certifying SEC filings they knew or should have known contained inaccuracies. The lead plaintiff deadline — the part of the story that sent law firms scrambling to email shareholders — is August 24, 2026.

That procedural detail is worth noting only insofar as it tells you how many firms are now competing to represent shareholders in a claim that's fundamentally about whether management's public language was a strategic euphemism or a legitimate reading of market conditions. Securities fraud requires scienter — knowledge or reckless disregard for the truth. The question isn't whether ZoomInfo's customers were confused. It's whether management knew the seat-based model was fracturing and continued to certify revenue projections built on it.

The market has already rendered its own judgment on that question. GTM is trading at $3.88 today, down 62% year-to-date. The stock briefly bounced back toward $10 in recent weeks before rolling over sharply, and it's now near the lower end of its 52-week range of $2.54 to $12.51.

This is a familiar machine in a new wrapper. It's the same story as every enterprise software company that tried to ride a technology wave while its core pricing model aged out: sell the AI narrative, stretch the guidance, hope the transition is invisible to quarterly numbers, and hope the restructuring costs absorb the narrative hit. The difference here is that ZoomInfo was selling the category it was being disrupted in. You can't be the leading AI-powered go-to-market platform while also being the company whose customers are confused about whether to buy your platform or build their own.

The structural question isn't whether ZoomInfo will recover. It's whether the consumption-based model generates enough revenue to justify the capital the company raised on the old seat-based story. That's a harder sell than guidance implies, especially when part of the customer base has already walked toward internal solutions.

The lawsuit deadline is five days away. The more interesting question is what happens when the transition economics become clear enough that "AI and agentic confusion" stops being a useful phrase to say on an earnings call.