Done your bit playing around with ChatGPT? Or maybe, tickled yourself by generating some images, comic strips, caricatures, and whatnot on Gemini? These Generative Artificial Intelligence, or GenAI, tools are powerful, educative, and entertaining, and they may well be the harbinger of the AI revolution hitting mankind, but now it is time for Agentic AI.
Agentic AI is the next step in the evolution of AI taking over our lives, and guess what—its presence is already being felt in fields as varied and diverse as financial services and healthcare. Put simply, Agentic AI refers to systems that can autonomously take decisions and act on them, with minimal human intervention. Or, as you will realise, with as much or as little human monitoring as you deem fit—unlike 'traditional' GenAI where you need to give prompts or commands for a task to be performed.
Unlike GenAI, which scrapes data off a vast database built on Large Language Models (LLM), Agentic AI uses multiple ‘agents’ (hence the terminology) that work together, learn from their actions as well as from the LLMs to decide, reason, and accomplish the task at hand.
Think NBFCs using Agentic AI to screen loan applicants as well as put together the paperwork. Or hospitals triaging doctor appointments and follow-ups in real time. These are not just examples—they are actual use cases of Agentic AI being deployed in India already.
No wonder, Qualcomm global President Cristiano R. Amon recently told an Indian news channel that India is not just a consumer but a launching pad for the next wave of technology—a veritable ‘new laboratory for Agentic AI’, no less.
“Agentic AI is a logical evolution towards the quest for sentient machine-intelligence akin to human capabilities,” said Aiyappan Pillai, senior member of IEEE, one of the world’s largest technical professionals organisations and founder & CEO of Congruent Services, an IT and software development company.
“Agentic AI tools are autonomous, mimic human reasoning, adapt to data, and process flows to translate decisions to action,” he added.
Like is the case with most new technologies, privacy and security concerns also abound, considering how Agentic AI sees machines taking the decisions and even implementing them directly, with minimal or no human intervention.
“Cautionary sentiment still remains regarding reliability and governance (models can err or drift), data privacy and consent, and the need for strict policies, human escalation, and deep system integrations to avoid ‘black-box’ behaviour, but these are concerns that remain top-of-mind for us and are things that we actively keep an eye out for to mitigate risks,” said Dikshant Dave, co-founder & CEO of Zigment AI, a Bengaluru-based AI-powered platform that automates customer interactions for sales and marketing.
Interestingly, Agentic AI is being tried in several sectors across Indian business, but without the kind of hype that accompanied the first wave of AI implementation: “Unlike the hype cycles seen elsewhere, many Indian organisations are showing strategic restraint. They are not rushing into large-scale deployments,” pointed out Ashvin Vellody, partner with Deloitte India.
There is a reason. “This cautious optimism is shaped by India’s unique context, including DPDP compliance, cost sensitivity, talent dynamics, and sovereign data priorities,” Vellody said.
There is also wariness on what shape Agentic AI will take—it is all fine to automate your workflow, but ensure there is the requisite human monitoring in place—that seems to be the running thought process. “Agentic AI works well today when scoped to clear tasks with guardrails. These systems augment teams rather than replace judgment,” said Dave.
As Agentic AI gets set to bring in the second round of revolution as far as artificial intelligence is concerned, and faster than you think, perhaps it is time for governments also to enter the picture. As Aiyappan said, “Privacy and cybersecurity considerations must be inherent in the design of Agentic AI systems and not an after-thought. Embracing the digital domain inherently leads to data security and privacy concerns. These are live issues that must be addressed through proactive regulation, both self and by authorities, data governance, effective policies, ethical practices, and continuous monitoring.”