Salesforce agrees to buy Listen Labs, adding AI customer research to its CRM

Why it matters

Salesforce is buying a way to collect and simulate customer feedback for use across its CRM and AI products, while the undisclosed price and unverified simulation performance leave the deal's economics and limits unclear.

Salesforce signed a definitive agreement on September 29th to acquire Listen Labs, the AI customer-research startup whose software recruits people, interviews them and analyzes their responses. The companies did not disclose a purchase price. Salesforce expects the deal to close in the fourth quarter of its fiscal 2027, subject to customary conditions and regulatory clearance.

For Alfred Wahlforss and co-founder Florian Juengermann, the agreement brings a product born from a startup's own customer problem into Salesforce's sales, marketing and service software. In January 2023, the pair moved into a San Francisco hacker house after an AI avatar app they had built drew 20,000 users overnight. They did not know who those users were or why they had arrived. As Wahlforss recounted in Listen's announcement of the Salesforce deal, a hackathon project to interview them became the foundation for Listen.

The startup launched its commercial research product in April 2025. Its software helps design a study, recruit participants, conduct audio or video interviews with follow-up questions, and turn the responses into reports and presentation materials. Salesforce says Listen can draw from a network of more than 50 million participants and conduct interviews in more than 120 languages. Those are company-reported scale figures; they do not establish how many people participate in a typical study or how representative a particular sample is.

Listen has also added digital twins: simulations, the company says, grounded in real customer behavior. Businesses can use them to explore possible reactions to a product or marketing idea before further testing with customers. Salesforce says the acquisition will extend customer-understanding capabilities across Marketing Cloud, Service Cloud and its broader AI portfolio. The strategic fit is direct: Salesforce sells software that helps businesses act on customer information, while Listen supplies a way to collect qualitative feedback and feed it into decisions and AI-agent workflows.

That combination raises the value of the information layer, while leaving an important distinction between interviewing real people and simulating likely responses. The former produces feedback from recruited participants; the latter estimates how people may react based on existing behavior. Salesforce's announcement describes both as capabilities it will bring to customers, but does not explain how the simulations will be validated or how the product will be integrated into its services.

Listen had already attracted substantial venture backing. In its Series B announcement, the startup said its latest round was led by Ribbit Capital, with Evantic, Sequoia Capital, Conviction and Pear VC participating, and described total funding as $100 million. Listen also reported more than one million interviews and 15-fold annualized-revenue growth in the nine months after launch. Those are company-reported metrics, and the announcement did not provide the underlying revenue figure.

The acquisition follows a disrupted fundraising process. On September 9th, TechCrunch reported that Listen had signed, then walked away from, a $125 million Series C term sheet at a $1.5 billion valuation, and that Salesforce acquisition talks were underway. That report described the proposed transaction value as around $2 billion; Salesforce and Listen have not confirmed that figure, and it should not be treated as the deal price.

Wahlforss wrote in the acquisition announcement that he will remain CEO, Juengermann will remain CTO, and the existing team will join Salesforce AI Labs. He also said the partnership would help Listen reach more companies faster. For Salesforce, the deal adds a customer-research product to a platform already built around customer records and business workflows. Whether that research becomes useful context for agents at scale will depend on how Salesforce integrates it and how reliably Listen's simulations reflect real customer behavior.