CEOs say AI agents are no threat. Their lawyers are writing a different story in the fine print.
The Signal
Twenty-seven enterprise software companies have now flagged AI agents as competitive risks in their SEC filings this quarter, up from seven last year. That's a near 4x jump in legal disclosure while executives tell investors everything's fine on earnings calls. Figma, HubSpot, and Workday are all on the list, companies whose entire business model depends on customers needing their specific interfaces and workflows.
The disclosure language is telling. These aren't vague concerns about "AI disruption." The risk is specific: customers could use AI agents to replicate core functionality or pull data directly without ever touching the actual product. That's not abstract future-gazing. That's the legal department saying "our moat might be code-deep, not product-deep."
This gap between what CEOs say publicly and what gets disclosed legally isn't new, but the speed of the shift matters. A year ago, seven companies thought agent risk was material enough to disclose. Now it's 27. The lawyers see what's coming faster than the talking points can keep up. When HubSpot's CEO says agents aren't a concern while his 10-Q says they could fundamentally alter customer behavior, believe the filing. SEC disclosures have consequences. Earnings calls are theater.
The real tell: these are companies built on being the system of record, the central hub. Agents don't need hubs. They need APIs and data access. The product becomes the pipe, and pipes get commoditized.
The Implication
If you're building in this space, watch what companies disclose, not what they say. The disclosure velocity tells you how fast the threat is moving in real time. For investors, this is your canary. When legal teams at 27 companies suddenly agree something's risky enough to put in writing, the migration has already started.
Source: The Information