Dinesh Nirmal, Senior Vice President, IBM Software
After launching its GenAI Innovation Centre in Kochi — aimed at enabling enterprises, start-ups, and partners to explore, experiment with, and develop generative AI technologies — IBM has now unveiled its Agentic AI Innovation Centre at its Bengaluru office. This new centre is designed to foster co-creation and provide clients with experience in working with autonomous AI agents.
In July, IBM also introduced its Ecosystem Incubation Centre (EIC) in Kochi, a dedicated space to drive innovation, entrepreneurship, and collaboration among start-ups, enterprises, and academic institutions. The EIC will serve as a hub for nurturing early-stage ventures, promoting skill development, and strengthening Kerala’s innovation ecosystem through a suite of structured programs that act as a launchpad for emerging innovators.
Dinesh Nirmal, Senior Vice President of IBM Software, shares insights on these latest initiatives in India and IBM’s growing focus on Agentic AI.
What role does the Kochi centre play in IBM’s broader innovation strategy in India?
Since the establishment of our Software Lab in Kochi, we have witnessed the city’s rise as a tech hub, characterised by a mix of talent and innovation. This newfound spirit inspired us to launch the IBM Ecosystem Incubation Centre (EIC) in the city. Our goal is to boost this effort by accelerating innovation, entrepreneurship, and collaboration among startups, enterprises, and academia. The EIC will provide structured support to Kerala’s innovation ecosystem through many initiatives, including hands-on mentorship and guidance for developing proof-of-concepts, technical workshops, hackathons, and integration with academic curricula to facilitate continuous learning and upskilling. The Government of Kerala’s exploration of AI for state administration also enables a strong collaborative environment.
What capabilities will the new Agentic AI Innovation Centre in Bengaluru unlock for IBM?
The new Agentic AI Innovation Centre is designed to unlock advanced AI capabilities for IBM, its clients, partners, and startups. This centre will offer access to IBM’s agentic AI tools, allowing users to develop and fine-tune Small Language Models (SLMs), orchestrate AI agents via watsonx Orchestrate, and rapidly audit performance. The centre intends to accelerate AI adoption, foster collaboration, and empower developers through immersive learning. It will help enterprises shift from AI experimentation to implementation at scale by focusing on high-impact use cases, boosting productivity, enhancing decision-making, and enabling scalable innovation. This centre will significantly advance IBM’s agentic AI capabilities by focusing on agent-to-agent communication, governance, security, and “agent ops” for managing and governing thousands of AI agents in large enterprises.
Can you share examples where agent-based systems are delivering measurable ROI?
We will see the full potential and results of AI agent return on investment (ROI) in 12-24 months after the agents have been fully integrated. However, we are already seeing great examples of agent usage across sales, procurement, and HR. Agents should be fit-for-purpose, meaning they should eliminate time-consuming workflows that bog down teams. Using IBM agents and/or a combination of other agents from our partners gives enterprises a lot of flexibility to customise and dictate how impactful an agent can be. Internally, IBM is also witnessing the highest adoption of AI, particularly in HR, IT help desks, and procurement. We saved $3.5 billion by integrating AI into over 70 enterprise workflows over the past two years.
Will AI agents eliminate the need for human intervention or create new opportunities?
We still see AI as augmenting human potential. For example, AI agents will help us complete mundane tasks and workflows that currently take teams hours, creating new opportunities for enterprises to use AI in a fit-for-purpose manner. The better the data, the more ways it can be utilized—and the less intervention and tuning will be required for these agents.
A recent Gartner report noted that 40% of agentic AI projects will be cancelled by 2027 and may not scale to the anticipated extent. How do you think Agentic AI will play out in the long run?
We see organisations that are enterprise-data ready as the winners of AI agents and what happens next in the tech landscape. However, companies may also see a lot of agentic agent projects fail because they are not set up to scale, are using bad data, or their AI agents are not governed properly. Having an open, governed data platform with intelligent automation will become a massive competitive advantage over the next five years.
Published on August 7, 2025