As the first half of 2025 draws to a close, the year is already defined as one in which agentic AI takes centre stage. While generative AI has shown the value of machine learning in automating tasks and producing content, agentic AI marks the next step, enabling systems to perceive, reason, and act independently, much like human decision-makers.
AIM spoke to Siddharth Murlidharan, vice president of financial services at Tredence to understand the current state of agentic AI, especially in the Indian context, and to find out what lies ahead beyond AI agents.
While the adoption of agentic AI in India is undoubtedly underway, its progress remains deliberate and measured.
According to Murlidharan, the real question is not whether the Indian or the US market is ready—it is whether organisations have the right mindset to adopt this technology.
India’s Position
Companies that are open to rethinking how they work are the ones that are moving forward, whether they’re based in the US, India, or the Middle East. The appetite to explore and adopt agentic AI is present in India, but full-scale integration is happening gradually.
“We are now seeing a lot of our customers use agentic AI for very specific use cases. Once you get past that first hurdle, the next stage in the evolution is to infuse agentic AI across the entire landscape,” he added.
Murlidharan talked about three main challenges that are responsible for slowing things down. First is adoption, as it takes effort and time to change established processes. Regulation is another concern, considering financial services operate under strict compliance rules, and agentic systems must meet high standards of transparency, fairness, and data protection. Education also plays a role, with many stakeholders lacking a clear understanding of what agentic AI can and cannot do, resulting in hesitation and unrealistic expectations.
Real Use Cases in Financial Services
Tredence has already worked on agentic AI use cases in financial services, including wealth management and lending.
For instance, in wealth management, agents assist with portfolio rebalancing, allowing financial advisors to shift their focus to strategy and client communication. In lending, agentic systems can process up-to-date information like employment verification and market volatility to improve credit decision-making. In fraud detection, agents act on patterns faster than traditional systems, provided they remain explainable and auditable.
The core reason for using agentic AI lies in speed and action. Traditional AI can detect patterns, but agentic AI can act on them in real time. This is especially important in areas like fraud prevention, anti-money laundering, and dynamic credit assessment, where rapid response is crucial. However, this only works when those decisions can be understood and justified to both business teams and regulators.
The Human Element Won’t Go Away
Despite the progress, Siddharth is clear that fully autonomous systems are not viable in finance, at least not in the foreseeable future.
Financial decisions involve trust, risk, and regulation, all of which require human oversight. A more realistic approach is hybrid decision-making, where agents manage the routine and humans handle exceptions, complexities, and customer engagement.
On India’s regulatory framework, he believes that the focus so far has been on preventing data misuse, ensuring unbiased decisions, and protecting end users.
Tredence’s approach is to work closely with clients’ compliance teams to ensure that any AI deployment fits securely within existing frameworks.
Agentic AI won’t eliminate jobs, but it will change them.
“”I don’t see agents right now replacing humans… What people need to be aware of is… if I don’t use agentic AI, then I will be on the back foot as compared to somebody who does,” he said.
Roles will shift from manual execution to higher-level work, like strategic decision-making and client engagement. For employees, the key is to adopt these tools and use them to do their work better and faster.
Next Step: Super Agents
He mentioned that companies are currently focusing on specific use cases for agents to build confidence and show value, with the next step being the development of advanced ‘super agents’ to handle full customer journeys or complex workflows.
“You start by going deep on a single use case to deliver clear value. The next stage is scaling agentic AI across the landscape with super agents managing entire customer relationships or complex workflows,” he added.
What sets India apart is the promising opportunity to deliver agentic services in multiple languages, potentially broadening access to financial services for a much broader population. With an infrastructure like UPI already in place, India is in a strong position to lead in real-time, AI-powered financial systems.
