Ex-Salesforce Chief on Selling AI to Enterprises

- Bret Taylor has a history of building successful products and companies, starting his career at Google where he co-created Google Maps. He later co-founded FriendFeed, which was acquired by Facebook, and then founded Quip, a collaborative productivity software acquired by Salesforce for a reported $750 million. - Sierra, co-founded with former Google executive Clay Bavor, is focused on building "agentic AI" for the enterprise, with a primary goal of transforming customer experience. The company's vision is that in the future, a company's primary digital interface will be an AI agent, not a website or mobile app. - Sierra has seen rapid growth, reaching $100 million in annual recurring revenue just seven quarters after its launch in February 2024. The company has raised significant capital, with a valuation of $10 billion after a $350 million funding round. - A key differentiator for Sierra is its "outcomes-based pricing" model, where customers pay for successful resolutions by the AI agent and are not charged if an issue needs to be escalated to a human. This directly addresses the enterprise challenge of proving ROI on AI investments. - While many enterprises are investing heavily in AI, a large percentage struggle to see a tangible return on investment. Studies indicate that a significant majority of companies are not yet seeing measurable revenue increases or cost reductions from their AI deployments. - The enterprise AI market is increasingly focused on vertical-specific solutions rather than general-purpose tools. Taylor believes that companies providing value for the core workflows of specific industries like telecommunications or financial services will have a competitive advantage. - Sierra's approach involves more than just retrieval-augmented generation (RAG); their agents integrate with underlying APIs to perform actions like modifying subscriptions or processing returns. To improve accuracy and reduce hallucinations, they incorporate specialized small models alongside larger ones. - The competitive landscape for enterprise AI includes players like Glean, which focuses on enterprise search and knowledge management, and others like Kore.ai and PolyAI, which specialize in conversational AI for contact centers. Competitors often focus on specific channels like voice or have different go-to-market strategies, with some targeting SMBs while Sierra aims for large enterprises.