ElevenLabs CEO Says Companies Should Disclose AI Voice Agents to Callers, for Now

ElevenLabs, the voice AI company behind many of the synthetic agents answering customer service calls, thinks businesses on the other end of those calls have a disclosure obligation — for the time being. CEO Mati Staniszewski said this week that companies should be upfront with callers when a machine, not a person, picks up, framing it as a temporary norm rather than a permanent rule. His reasoning: once reaching an AI agent becomes as routine as punching through an automated phone tree, the announcement stops carrying information.

That position lands as voice cloning and real-time speech synthesis get good enough that listeners can no longer reliably separate a human agent from a generated one. ElevenLabs sells the underlying models that businesses use to build those agents, so its customers are the ones making the disclosure call in practice. Staniszewski's stance puts him closer to the transparency camp than to rivals who argue disclosure is a friction cost that slows adoption.

The company has grown quickly on the back of enterprise demand for voice agents that can handle support, sales and collections calls at a fraction of the cost of a human seat. ElevenLabs has raised at a multi-billion-dollar valuation and now counts large publishers, media firms and contact-center operators among its users. Margins in that business depend heavily on inference costs, which fall as models get more efficient — a dynamic Staniszewski has pointed to when asked how the company balances cheap per-minute pricing against compute bills.

IPO timing is the other thread Staniszewski has been willing to discuss, though he has kept the specifics vague. He has signaled the company is in no rush, and that going public is a milestone rather than a strategy. For a company at ElevenLabs' stage, that reads as a signal it can fund growth from private markets for now, and that it would rather let revenue compound than face quarterly scrutiny while the voice AI market is still shaking out.

Disclosure is not just a branding question. Regulators in the US and Europe are weighing rules that would require bots to identify themselves in some contexts, and states including California have already pushed automated-caller statutes. A CEO publicly endorsing disclosure gives ElevenLabs a cleaner story with lawmakers, and gives its enterprise buyers cover if they choose to announce the bot. It also pre-empts the scenario the industry fears most: a viral recording of someone being deceived by a synthetic voice, followed by a regulatory crackdown.

What to watch is whether ElevenLabs' customers actually follow the advice, and whether any of them break ranks to argue that disclosure hurts completion rates on calls. The harder question is what happens when Staniszewski's predicted normal arrives — if callers assume a machine is listening, the burden shifts from announcing the bot to proving a human is on the line. That inversion, not the current disclosure debate, is where the next regulatory fight sits.