'This represents an important evolution in AI governance': Infosys lays out how your business can thrive with the new generation of autonomous AI agents
Reported by TechRadar1 source ↓
Infosys tells us how your business can prepare for the age of autonomous AI agents.
Key points
- Infosys says each agent needs a defined purpose, an accountable owner and limits on its access and actions
- Organizations should inventory non-human identities and track their owners, purposes and permissions
- Agents handling sensitive data or high-impact actions need tighter controls than agents doing low-risk tasks
- Some agents can prepare actions for human approval; independent operation should remain within approved limits
Details
In an interview with TechRadar, Infosys executive Umashankar Lakshmipathy says agents that access systems and act with limited human involvement create risks beyond those posed by assistants that provide information. An agent can have legitimate access yet act outside its intended purpose, and errors may trigger chains of actions across systems that are difficult to detect or reverse.
Lakshmipathy says organizations should record each agent’s purpose, owner, access and actions, and separately assess the identity and authority of agents that work together. He warns that a compromised or poorly governed agent could pass inaccurate information to another agent or try to use a more privileged agent to take action.
The interview recommends setting oversight according to the impact and reversibility of an agent’s work and the sensitivity of the information it uses. Low-risk tasks such as summarizing internal information may need lighter oversight, while financial transactions, customer data and production infrastructure call for tighter limits; some agents can also be required to seek human approval before acting.
Written from coverage as of
Images
- TechRadar
Stocks in this story
Change since open: 20 Nov to 21 Nov 2026
Last refresh