The tool, named Claude Cowork, is an agentic coding and workflow automation system that has existed for nearly a year. What Anthropic has now done is expand its enterprise AI assistant with an added automation layer designed to handle complete business workflows. The move underpins how quickly agentic tools are evolving from developer-facing products into systems capable of executing end-to-end tasks across organisations.
That puts the AI company — roughly valued at $350 billion — into direct competition with traditional Software-as-a-Service (SaaS) platforms and IT services firms. Analysts are beginning to question how essential many legacy software layers will remain if AI agents can independently solve complex operational problems.
Mayank Gupta, founder of TraceLM.ai, who has used Claude Cowork, said that from a pure capability standpoint, the tool does not feel radically different from Claude Code, which developers have had access to for about a year. Claude Code, he said, has already demonstrated how powerful agentic tools can be in the hands of developers.
“If a general-purpose agentic tool like Claude Cowork, combined with 11 open-sourced and customisable plugins, can handle research, analysis, legal document reviews and operational tasks end-to-end without relying on traditional SaaS interfaces, then a large part of today’s legacy software becomes structurally vulnerable,” Gupta said. “This is an early indication that AI is moving beyond being a productivity add-on and starting to replace entire layers of legacy software.”
The sense of disorientation around AI’s pace of change was echoed by Aditya Agarwal, former chief technology officer of Dropbox and a former board member at Flipkart. In a post on X, Agarwal wrote that writing code by hand may soon be obsolete. After spending a weekend building with Claude, he said it no longer made sense to write code manually, describing skills he once specialised in as “free and abundant”. He added that both the “form and function” of his early career were now being produced by AI. Anthropic’s update automates tasks across legal, sales, marketing, and data analysis, intensifying competition between incumbent technology firms and AI companies pitching gains in efficiency and productivity.
JPMorgan analyst Mark Murphy wrote in a note that Claude is evolving “from a chatbot that answers text-based questions to an agent that executes labour across your Mac’s local files and browser”.
The software development life cycle is already undergoing a tectonic shift, with AI agents writing large volumes of code that humans primarily verify. Beyond engineering, functions such as finance and accounting, legal operations and supply chains are also being reshaped as agents increasingly assist employees. For Indian IT companies, already grappling with slower client spending, such systems pose challenges to traditional software-driven workflows and raise questions around long-term hiring plans.
“What Anthropic has really done with Cowork is package these agentic capabilities for non-developers... people who don’t write code but still manage complex workflows using a patchwork of SaaS tools and manual processes,” Gupta said.
Investor unease spilled into markets on Tuesday, triggering a selloff in technology stocks in the US and Europe, with Indian IT shares also sliding on Wednesday.
Executives have cautioned that AI’s impact on legal work remains underestimated. Cognizant Chief Executive Officer Ravi Kumar said last year that classical software has barely touched legal functions, making them “the right candidate for agentification”.
Ashok Soota, chairman and chief mentor of Happiest Minds Technologies, said Anthropic’s plugins lower the entry barrier to building software but do not eliminate the need for IT firms. “In fact, it enhances the same,” he said.