Salesforce agrees to buy Listen Labs, giving marketers an AI shortcut to test messaging before real customers see it

Salesforce has signed a definitive agreement to buy Listen Labs, an AI startup that runs customer interviews and simulates how buyers might react to a product, price, or message before it ships. The companies announced the deal on September 29. Salesforce hasn't disclosed the price, though MarTech reports that several outlets have put it near $2 billion. Salesforce expects the deal to close during its fiscal fourth quarter of 2027, assuming it clears regulatory review.

What changed

Alfred Wahlforss and Florian Juengermann started Listen Labs in 2023, building software that handles a research study end to end: it designs the questions, recruits the right people, runs the interviews without a human moderator on the call, and writes up what it heard. Salesforce's own announcement puts the participant pool at upward of 50 million people, reachable in over 120 languages, which is how a study that once took months gets done in days.

The feature drawing the most attention is what Listen Labs calls digital twins: AI models grounded in real customer behavior that produce a simulated response to a new idea, product, or marketing message before an actual person ever sees it. Salesforce plans to fold that into Marketing Cloud, Service Cloud, and its wider AI portfolio, giving Agentforce a source of customer opinion that previously sat outside the CRM entirely.

"Listen Labs has reimagined customer experience, market research, and customer simulation with AI agents that can conduct meaningful research and uncover context," said Aman Naimat, President of AI Labs at Salesforce.

Per MarTech's reporting, Listen Labs will sit inside Salesforce AI Labs once the deal closes, and the two founders keep their current titles, Wahlforss running the company and Juengermann leading engineering. Salesforce hasn't said how the product will be priced or packaged once it's inside the broader suite, or whether it will keep operating as a standalone purchase for companies that don't otherwise run on Salesforce.

Who it affects

Marketers already running campaigns through Salesforce are the most immediate audience. Research that used to live in a separate tool, or get commissioned out to an agency, is headed toward becoming a workflow step inside the same platform that holds the customer data and sends the campaign. A line item that used to read as outside research spend starts to look like a CRM feature instead, with the usual consequences: faster access, but also a harder case to make for keeping a separate research budget next renewal cycle. A product marketer drafting positioning for a launch could, in theory, pull a simulated reaction to three headline options the same afternoon the brief is written, instead of waiting a week for a panel to come back from an agency.

It matters for marketers outside the Salesforce ecosystem too, because of a distinction worth sitting with. An interview captures what a recruited person actually said. A digital twin estimates what a customer might say or do, based on a model of past behavior. Once both feed the same AI agent, they can look identical from the outside: an insight with a confidence score attached, no label showing where it came from. MarTech's analysis of the deal flags this directly, arguing that distinction has to stay visible as agent-driven decisions become more common.

What to do differently

Salesforce hasn't said yet how deep the Listen Labs integration goes, or which agent workflows get access to the simulated data first. That leaves time to set some ground rules before the feature shows up in your instance.

• Ask any vendor pitching AI customer research whether a given insight comes from a real transcript or a simulated model, and keep the two labeled separately in your own reporting
• Treat a digital twin's prediction as a way to narrow options before a test, not a replacement for checking a message or price against real customers
• If your team already pays for a standalone research tool, ask now how it plans to compete once this kind of simulation ships inside the CRM for free

The bigger signal is consolidation. Customer research has mostly lived in its own category of standalone platforms and independent agencies, separate from the CRM and marketing automation stack. Salesforce folding a research-and-simulation platform directly into Agentforce suggests that boundary is starting to move, the same way customer data platforms and marketing automation merged a decade ago. Marketers who treat research as something that happens upstream of the martech stack, rather than inside it, may find that assumption gets tested sooner than expected. The question worth raising internally now, before the integration ships, is who signs off when a campaign decision traces back to a simulation instead of a conversation with an actual customer.