Agentic AI is exploding in India – but 71% of companies can't scale beyond experiments

Every day, more Indian businesses are exploring the potential of Agentic AI—autonomous systems that can function with minimal human oversight. According to Deloitte’s fourth wave of the State of GenAI report, more than 80 percent of organisations in India are actively looking into the development of AI-driven autonomous agents. “The growing interest in multi-agent workflows, where smaller AI sub-agents operate under a master agent to execute complex tasks,” validates this shift. Half of the surveyed businesses identified these workflows as a key focus area.

The broader push towards generative AI (GenAI) is also evident, with 70 percent of businesses eager to use the technology for automation. AI adoption is accelerating, with 61 percent of companies running over ten GenAI experiments, reflecting an industry-wide effort to integrate AI at scale. Sectors such as IT, customer service, marketing, operations, and product development have emerged as early adopters.

The report also highlights the tangible benefits of AI integration. Over 67 percent of businesses reported a positive impact on all phases of the software development lifecycle, while nearly 70 percent said their AI efforts met or exceeded return on investment (ROI) expectations.

However, despite the enthusiasm, businesses are struggling to scale AI solutions effectively. While many companies are experimenting with AI, only 29 percent have managed to scale even 30 percent of their AI proofs of concept. Additionally, AI adoption remains inconsistent across workflows, with 61 percent of organisations reporting that less than 40 percent of employees actively use GenAI tools.

The challenges aren’t just operational. Concerns around AI errors with real-world consequences (36 percent), bias and hallucinations (30 percent), and data quality (30 percent) continue to slow deployment. Organisations remain optimistic, though, with most expecting to overcome these roadblocks within the next 12 to 24 months.

“As Indian organisations explore Agentic and GenAI, the key to unlocking their potential lies in moving from experimentation to large-scale deployment. Businesses must build trust in AI systems by addressing concerns about errors, bias, and data quality through strong governance,” said Moumita Sarker, Partner, Deloitte India. She emphasised that firms need to ensure they have a suitable ecosystem to scale AI experiments and measure their impact effectively.

A major factor influencing AI adoption strategies is the preference for off-the-shelf AI tools. The report found that most Indian businesses opt to buy AI solutions rather than develop them in-house. While this approach enables faster deployment, it also presents challenges in customisation, differentiation, and dependence on external vendors. Companies that fail to tailor AI solutions to their specific operational needs may struggle to realise their full potential.

The financial stakes in AI adoption are high. According to the survey, 94 percent of firms would need more than six months to exit an AI project that fails to meet ROI goals, while 76 percent anticipate the process taking over a year. Given the rapid evolution of AI, 28 percent of businesses worry that their current AI solutions will become obsolete within two years, raising concerns about the long-term viability of third-party tools.

“Embracing an agile innovation approach is essential to stay ahead of AI advancements and optimise long-term returns. Balancing rapid adoption with sustainable strategies will be key to successful AI investment and growth,” opines Sarker.