COGNIZANT IS PUTTING GOOGLE AI AGENTS INSIDE ITS CLIENTS' SYSTEMS Here is why that should worry anyone whose data sits on a Cognizant-run network. Start with the track record. In August 2023, hackers called the help desk Clorox had outsourced to Cognizant and asked for employee passwords. They got them. No identity verification, according to Clorox's $380M lawsuit against Cognizant. The attack knocked out Clorox's ordering systems and caused nationwide product shortages. Cognizant disputes the claims. In 2020, Cognizant's own network was hit by Maze ransomware, which the company said cost it $50 to $70 million. That is the company now selling AI agents deployed into client job roles in weeks instead of quarters. Cognizant's own press release says these agents automate 60 to 70% of manual effort. Every agent needs standing access to the systems it works in. More standing access means a bigger blast radius when access control fails. Access control is exactly what Clorox says failed. Now add jurisdiction. Cognizant has 357,600 employees and over 250,000 of them are in India, roughly 70% of the company. Federal data rules like CJIS and IRS Pub 1075 restrict offshore access to criminal-justice and taxpayer data. Cognizant now runs the State of Iowa's entire IT operation. No one has shown the public whether those offshore-access rules are enforced, or whether Gemini tools will touch state systems. Two dates tell the story. April 29: Cognizant sets aside $270M for severance under Project Leap, with analysts modeling 7,000 to 15,000 job cuts. July 7: Cognizant announces Gemini for up to 200,000 of its own employees. Your data is now on these networks whether you picked Cognizant or not.