Most people interact with AI agents every day without thinking about it. Behind the technology handling their customer service calls, drafting their legal filings, and routing their freight are a small number of San Francisco founders who bet early that AI could do more than generate text. Here is who they are.
Sierra: Bret Taylor and Clay Bavor
The idea for Sierra began over a long lunch at a Mediterranean restaurant in Palo Alto in early 2023. Bret Taylor had just stepped down as co-CEO of Salesforce, where he had earlier led the acquisition of Slack and co-created Google Maps at the start of his career. Clay Bavor had spent 18 years at Google, most recently leading Google Labs and building the company’s AR and VR efforts. By the time they left the restaurant, they had decided to found a company together.
Sierra launched in February 2024 as an AI agent platform for enterprise customer service. Taylor has banned the word “chatbot” inside the company’s offices, describing the product’s goal as building agents that are empathetic and conversational rather than robotic. The company builds the agents for customers directly rather than selling tools for customers to do it themselves.
Sierra crossed $100 million in annual recurring revenue in November 2025, less than two years after launch. By October 2025, voice interactions had overtaken text as the primary channel on the platform. Customers include SoFi, Ramp, ADT, Brex, Casper, Rivian, Cigna, SiriusXM, and Tubi. The company raised $635 million across three rounds, reaching a $10 billion valuation in September 2025. Forbes named Taylor a billionaire in November 2025 based on his roughly 25 percent stake.
Decagon: Jesse Zhang and Ashwin Sreenivas
Jesse Zhang and Ashwin Sreenivas met at an Andreessen Horowitz retreat in Utah and began surveying businesses about where AI could deliver the most measurable value. Both had backgrounds at Google, Citadel Securities, and Palantir. They founded Decagon in San Francisco in August 2023.
The platform builds conversational AI agents for customer service across chat, email, voice, and SMS. Its agents handle refunds, cancellations, identity verification, and payment disputes without human intervention, achieving average deflection rates exceeding 80 percent. One client, Bilt, reduced its customer support team from hundreds of contractors to 65 employees. Clients include Hertz, Avis, Duolingo, Notion, Oura Health, Rippling, and Eventbrite.
Decagon raised $250 million in Series D in January 2026 led by Coatue Management and Index Ventures, reaching a $4.5 billion valuation and total funding of approximately $481 million.
Harvey: Winston Weinberg and Gabriel Pereyra
Winston Weinberg was a first-year litigator at O’Melveny and Myers when his roommate Gabriel Pereyra, a research scientist at Google DeepMind and Meta, showed him OpenAI’s GPT-3. Weinberg saw it could handle the repetitive work that consumed junior lawyers’ time. The two sent a cold email to Sam Altman that led to the OpenAI Startup Fund’s first check into the company. They founded Harvey in San Francisco in 2022.
More than 25,000 custom agents now operate on the platform, handling contract drafting, due diligence, document review, and litigation preparation for more than 100,000 lawyers across 1,300 organizations in 60 countries. Clients include a majority of the AmLaw 100 and enterprises such as NBCUniversal, HSBC, and Allen and Overy. Harvey reached $190 million in annual recurring revenue in January 2026. Its most recent raise brought in $200 million co-led by GIC and Sequoia at an $11 billion valuation.
LangChain: Harrison Chase and Ankush Gola
Harrison Chase was a machine learning engineer at Robust Intelligence in late 2022, just weeks after OpenAI released ChatGPT, when he launched LangChain as an 800-line side project on his personal GitHub. The framework pioneered the idea of chaining large language models to external tools and data sources so they could take action rather than just generate text. It spread faster than Chase anticipated. He had no plan to leave his job. He co-founded the company with Ankush Gola and incorporated in San Francisco.
LangChain has since become the standard framework for developers building AI agent applications, with 118,000 GitHub stars as of late 2025. Its suite includes LangGraph for orchestrating complex agent workflows and LangSmith for testing and observability. Enterprise customers include Cisco, Workday, ServiceNow, and Databricks. Chase has described the competitive position as having 500 competitors and zero competitors simultaneously, on the premise that most enterprises will use multiple agent platforms and many will be powered by LangChain underneath.
In October 2025, LangChain raised $125 million in Series B at a $1.25 billion valuation led by IVP, with CapitalG, Sapphire Ventures, Cisco Investments, Datadog, and Databricks among the participants. Total funding stands at $260 million.
