By Shishank Gupta
Today, AI is way more than just an incredibly efficient tool of automation. It is becoming the nerve centre of business, determining how organisations create and deliver value at unprecedented scale: a leading consulting firm estimates that analytical AI could add global economic value worth an estimated $9.4 trillion to $15 trillion in 2040. Separately, India’s public policy think-tank, Niti Aayog, says the country could see a $2 trillion addition to the country’s economy by 2035.
Little wonder then that adoption is soaring. UNCTAD’s Technology and Innovation Report 2025 forecasts that the global AI market will increase sharply from $189 billion in 2023 to $4.8 trillion by 2033. Organisations across industries are adopting AI technologies to decide better, grow faster, and innovate smarter.
Agentic AI has further expanded these possibilities by amplifying employee potential and accelerating time-to-value, to make organisations more competitive. For example, an energy company centralised scattered documents containing health, security, safety, and environment data in a SharePoint library and trained an AI agent to answer user queries. By providing employees instant access to information, the solution enhanced both productivity and quality of experience. Separately, an insurance firm used an agentic AI solution to streamline pre-authorisation processes and thereby reduced manual effort by 45%, while achieving 90% accuracy.
Anticipate, act, orchestrate, innovate
The above examples only hint at the potential of agentic AI systems, which understand, predict, analyse, and respond on their own, to enhance value chains across industries. Leveraging advanced machine learning capabilities, they interpret context, anticipate future events, and even trigger appropriate action without human intervention. AI agents are goal-oriented and can take informed decisions to achieve specific outcomes, such as deciding the best sequence of actions for resolving a customer’s problem expeditiously.
As opposed to rule-based, somewhat static, conventional AI, agentic AI is highly dynamic, learning and adapting in real-time to its environment to achieve a defined target. Consider healthcare, which is severely short of medical staff and other resources (India has only about 21 health professionals per 10,000 people versus the recommended 44.5.) AI agents can help bridge the gap by dynamically reallocating staff and redirecting patient flow based on real-time data.
What’s more, agentic AI can orchestrate multiple systems, such as CRM, payments, inventory management, and logistics to autonomously resolve complex, multi-step problems. In marketing operations, agentic AI systems analyse customer and transaction data to hyper-personalise customer experience, making customers feel understood and valued. Over time, this results in higher conversion rates, loyalty, and top-line growth.
Move from pilot to production
Although organisations in every industry are exploring agentic AI to accelerate innovation, many are struggling to unlock its full potential. In 2025, a leading global research and advisory firm predicted that more than 40% of agentic AI projects could be scrapped by 2027 owing to rising costs, unclear business value, and risk management challenges.
Swayed by hype, companies are investing in pilot projects without a clear line of sight to returns. To extract real value from agentic AI systems, they need to look beyond one-off tools with limited impact – such as smart assistants – to comprehensive use-cases that can improve productivity and enhance impact across the enterprise.
Organisations can start with small projects and scale them gradually, while managing risk and compliance. Along with clearly scoping use-cases, they should ensure robust system architecture, cybersecurity, and governance.
Agents should be fully compliant with applicable data security and privacy regulations. Training datasets must be clean, comprehensive, accurate, and unbiased to produce high quality, non-discriminatory outcomes, while the underlying algorithmic models should be transparent and explainable. It is also crucial to balance speed of innovation, with control. A Responsible AI framework, with human oversight of critical agentic AI decisions will ensure that agents operate within ethical, social, and regulatory boundaries.
Although agentic AI shows immense promise, much of its potential is yet to be realised. With the right agentic AI strategy and adequate guardrails, enterprises can transform every aspect of the business – from identifying business opportunities to core operations to employee experience to customer engagement – and unlock unprecedented value. Ultimately, this can amplify our ability to reimagine business through experiences that were previously inconceivable.
The writer is SVP & head of the Digital Workplace Ecosystem and Microsoft Practice at Infosys
Disclaimer: The views expressed are the author’s own and do not reflect the official policy or position of Financial Express.