The Agentic AI invasion: It’s here—but does the software industry face an existential threat?

Subscribe to enjoy similar stories. It was swift and brutal, the global sell-off in tech stocks this week after Anthropic’s launch of Agentic AI tools with startling skills. From America to India, software stocks lost $285 billion in market value globally within the span of a single trading session.

Indian IT majors also slid sharply, with investors seeming convinced that AI agents had finally arrived to replace software products and services at scale. The trigger was Anthropic’s decision to open-source 11 ‘plugins’ on GitHub for its Claude Cowork AI platform. Compatible with Claude Code, its coding assistant, these plugins let users set goals for its AI model to achieve.

Each plugin guides Claude on which tools and what data to use for every task. Together, this toolkit spans productivity, sales, marketing and legal work. What rattled markets the most was its legal plugin for reviewing contracts, compliance and risk.

The panic was understandable, but contracts and revenues have not changed. So, did investors overreact, or do IT firms face a mega-threat? Until now, humans wrote the script. AI tools such as Replit, Bolt and Claude helped refine it as assistants.

Now, an agent can write the script, analyse it, test it, improve it and even deploy it, which explains the market’s extreme reaction. The first casualties were predictable. Large tech firms with product-heavy business models driven by software-as-a-service (SaaS)—especially those behind banking, financial service and insurance workflows—were hit the hardest.

Indian IT majors that sell licensed platforms alongside services suddenly looked exposed. One may argue that IT firms already use AI platforms like Kiro, Claude and Replit to build their own for clients. But