Anthropic announced that it has acquired Vercept, a startup focused on enabling artificial intelligence systems to interact with software applications in the same way humans do. The acquisition is aimed at accelerating the development of advanced computer-use capabilities within Anthropic’s Claude models, allowing AI systems to execute complex, multi-step workflows directly within live applications.
As enterprises increasingly rely on AI to automate sophisticated tasks, the ability for AI systems to operate software environments rather than simply generate text or code has become a key frontier in the industry. By integrating Vercept’s technology and expertise, Anthropic aims to enhance Claude’s ability to perceive on-screen interfaces, interpret user intent, and carry out actions across digital tools and operating systems. These capabilities allow AI to perform tasks such as navigating spreadsheets, managing workflows, and completing multi-step processes across applications, much like a human working at a keyboard.
Vercept was founded with a focus on solving the technical challenges required for AI to both understand and interact with graphical user interfaces. Its founding team Kiana Ehsani, Luca Weihs, and Ross Girshick has spent years researching how AI systems can observe, interpret, and act within the same software environments used by human workers. As part of the acquisition, the Vercept team will join Anthropic to further develop these capabilities and contribute to the company’s broader roadmap for intelligent AI agents.
“People are using Claude for increasingly complex work writing and running code across entire repositories, synthesizing research from dozens of sources, and managing workflows that span multiple tools and teams. Computer use enables Claude to do all of that inside live applications, the way a person at a keyboard would.”
“Vercept was built around a clear thesis: making AI genuinely useful for completing complex tasks requires solving hard perception and interaction problems. The Vercept team including co-founders Kiana Ehsani, Luca Weihs, and Ross Girshick have spent years thinking carefully about how AI systems can see and act within the same software humans use every day.”
Also Read: WordPress Unveils Built-In AI Assistant: What It Means for DevOps and Businesses
Following the acquisition, Vercept will wind down its external product in the coming weeks as its team transitions fully into Anthropic’s research and engineering efforts. The move reflects Anthropic’s strategy of bringing specialized teams into the company to strengthen its capabilities in AI agents and advanced automation.
The acquisition comes shortly after Anthropic introduced its latest model update, Claude Sonnet 4.6, which significantly improved the AI system’s ability to operate within software environments. According to the company, the model’s performance on the OSWorld benchmark a widely used evaluation for computer-use capabilities has improved from under 15% in late 2024 to 72.5% today, approaching human-level performance in tasks such as navigating complex spreadsheets and completing forms across browser tabs.
Anthropic noted that Vercept is the latest team to join the company as it continues to expand its AI research capabilities. The organization stated that it prioritizes teams whose technical ambitions align with its focus on building advanced AI systems grounded in safety, reliability, and rigorous engineering practices.
As the race to build autonomous AI agents intensifies across the technology sector, the integration of Vercept’s expertise is expected to accelerate Anthropic’s efforts to develop AI systems capable of executing real-world digital workflows, helping enterprises unlock new levels of productivity and automation across software ecosystems.
Source: Anthropic
