What is Agentic Artificial Intelligence (AI)? Explain its working. Describe its applications with suitable examples. Discuss the advantages, risks and challenges associated with Agentic AI systems. (250 words)

Agentic AI systems pursue a defined goal through iterative perception, reasoning, tool use and feedback, rather than merely generating a single response.

A goal and constraints are converted into a plan; the agent observes data, decomposes tasks, selects tools or APIs, executes actions, checks outputs and revises the plan. Memory, retrieval, simulators and human approval can improve continuity. Guardrails define permissions, audit trails and stop conditions.

Examples include an agricultural agent combining weather, soil and market data for advisories; a health workflow agent scheduling tests under clinician approval; cybersecurity agents triaging alerts; disaster systems allocating resources; and public-service assistants navigating forms. High-impact actions require human-in-the-loop controls.

Automation of multi-step work, personalisation, faster analysis, accessibility, productivity and continuous monitoring can improve governance and enterprise efficiency.

Hallucinated plans, unsafe tool calls, reward hacking, prompt injection, privacy breaches, bias, cyber misuse, opaque accountability and cascading errors are serious risks. Compute concentration, energy use, liability, interoperability and workforce displacement add policy challenges.

Risk-based regulation, sandboxing, least-privilege access, provenance, red-teaming, independent audits, incident reporting and human appeal are essential. India should combine innovation with DPDP compliance and accountable public procurement.

science and technology; emerging technology; strategic application; India