Adept AI was founded in 2022 by David Luan (former VP of Engineering at OpenAI) alongside several former Google and DeepMind researchers. Based in San Francisco, the company focuses on building AI models that can interact with any software tool the way a human would — clicking buttons, filling forms, and navigating interfaces.
The company raised over $400 million in funding, with investors including General Catalyst and Spark Capital. Adept’s core model, ACT-1 (Action Transformer), was trained to understand and operate software UIs, going beyond text generation into real-world task execution.
In 2024, Amazon hired much of Adept’s leadership and technical team, licensing the company’s technology in a deal reportedly worth over $400 million. The remaining Adept team continues to operate, focusing on enterprise AI agent products.
Adept’s vision of AI agents that work alongside humans in existing software — rather than requiring users to learn new tools — helped define the “AI agent” category that many companies now pursue. Their ACT models demonstrated that transformer architectures could learn to navigate complex software workflows with minimal human supervision.
