Agentic AI is redefining the future of banking, financial services, and insurance (BFSI) by moving beyond traditional AI models to systems that act autonomously, make complex decisions, and execute tasks with minimal human intervention. From real-time claims settlement to fraud detection and personalized financial planning, it is unlocking unprecedented efficiency and transforming customer experience. Yet, this transformative potential also raises critical challenges around governance, transparency, and data privacy—making responsible deployment imperative. In an exclusive interview with Abhineet Kumar of Elets Technomedia, Rishi Aurora, Managing Partner, IBM Consulting India & South Asia, shares insights on IBM’s long-standing heritage of trust and its commitment to AI governance.
Edited excerpts:
Agentic AI is emerging as a transformative force across industries. How would you define Agentic AI in the context of BFSI, and how is it fundamentally different from traditional AI models?
Agentic AI describes AI systems that are designed to autonomously make decisions and act, with the ability to pursue complex goals with limited supervision. As India’s financial services industry moves towards AI-driven transformation, Agentic AI is emerging as a powerful model with significant potential in enabling transformation such as real-time claims settlement, dynamic loan underwriting, fraud detection, and personalized financial planning, with minimal human intervention.
Unlike traditional AI, these Agentic AI systems are unique in that they can independently set goals, make decisions, take actions across business workflows, and unlock new capabilities in areas like onboarding, compliance, and risk management. They enable banks and insurers to move from reactive to proactive decision-making, driving speed, precision, and compliance at scale.
Can you share specific real-world use cases where Agentic AI has successfully improved onboarding, compliance, or risk assessment functions in banking or insurance?
The ability of Agentic AI to understand complex financial language, reason across diverse data sources, and take corrective action—such as filling forms and analyzing documents—makes it particularly impactful for India’s BFSI industry. It can dynamically operate in fast-paced, data-heavy environments and significantly improve decision-making, optimize workflows, and enhance compliance. For instance, it is being used to perform continuous, autonomous risk audits to detect unusual patterns and respond to emerging threats. Using similar logic, it is well-positioned to support compliance monitoring and loan underwriting, both of which involve large volumes of data-intensive, repetitive tasks.
On the customer-facing side, Agentic AI and virtual assistants are delivering AI-driven financial advisory services by automating specific wealth management activities or crafting investment strategies based on market conditions and individual risk tolerance. There is significant opportunity to enhance customer experience and enrich customer interactions in financial services in areas such as hyper-personalization of financial products and fast-tracking the onboarding and KYC process.
With greater autonomy comes greater risk. What are the most pressing governance and ethical challenges you’ve observed while deploying Agentic AI in BFSI? Data privacy and control are becoming increasingly sensitive topics, especially with Agentic AI systems making decisions autonomously. What kind of safeguards need to be put in place?
Agentic AI is ushering in a new era for financial services. From managing portfolios and detecting fraud to automating compliance and transforming customer engagement, Agentic AI is redefining financial operations. As AI systems become more autonomous, robust governance frameworks are non-negotiable. These frameworks must balance autonomy with accountability, speed with safety, and innovation with reliability.
Bias is another critical challenge. AI systems learn from historical data, but if the data contains biases, AI may amplify them. AI agents may make undesirable decisions such as prioritizing efficiency over fairness or privacy.
Other areas of concern include cybersecurity risks such as adversarial attacks, data leaks, and unauthorized access that expose sensitive information. To mitigate these risks, Agentic AI algorithms should incorporate access controls and authentication mechanisms to prevent unauthorized interactions. Another challenge lies not only in AI itself but in aligning legacy systems, processes, and people. Many enterprises still operate on decades-old IT systems, making AI integration complex and slow.
IBM has a long heritage of trust and commitment to AI governance, and responsible AI is both a strategic and ethical imperative. We have developed a framework—“Pillars of Trust”—to make these principles of Responsible AI explicit:
- Explainability: AI systems should be transparent, particularly about the data used in their algorithmic recommendations.
Fairness: AI must ensure equitable treatment of individuals or groups. Properly calibrated, AI can help humans make fairer choices, countering human biases, and promoting inclusivity.
- Robustness: AI-powered systems must be defended from adversarial attacks, minimizing security risks and enabling confidence in outcomes.
- Transparency: To reinforce trust, users must be able to see how the service works, assess its functionality, and understand its strengths and limitations.
- Privacy: AI systems must prioritize consumer privacy and data rights, with explicit assurances about how data will be used and protected.
How is IBM Consulting helping financial institutions design AI systems that are not just performant, but also compliant by design? What role do explainability and transparency play here?
Agentic AI demands not just technological deployment but also a deep reconsideration of how work is structured, how decisions are made, and how humans and machines collaborate. It is critical that organizations integrate bias consideration and bias testing into their product development cycle.
IBM Consulting has deep expertise in applying an ethically responsible approach to AI and in working alongside AI to deliver business value, supported by user-centric design. We collaborate with clients to create responsible, transparent AI strategies and to build governance frameworks supported by automated AI governance platforms. This ensures that the data used to train models is free of bias, while the AI models themselves are transparent, explainable, and free of drift. Implementing strong governance structures that define roles, responsibilities, and oversight mechanisms is crucial for fostering accountability in AI initiatives.
IBM watsonx is IBM’s next-generation AI and data platform, designed to provide self-service access to high-quality, trustworthy data. It enables the effective adoption of responsible AI by training, validating, tuning, and deploying AI systems across enterprises with speed, trusted data, and governance.
Looking ahead, what’s your vision for the evolution of Agentic AI in regulated industries? Are we moving towards co-pilot models or full autonomy—and what should organizations start doing today to prepare for that shift?
The global economy is currently experiencing an AI super-cycle, driven by unprecedented progress and investment in AI technologies. This cycle is igniting transformation initiatives aimed at accelerating growth and uncovering new efficiencies. The true value of AI will be realized when humans can fully delegate both simple and complex tasks to AI systems, allowing them to focus on more strategic, higher-value activities.
This transition will mark a decisive moment in AI value realization. That said, human oversight will remain essential, even as more tasks are delegated to agents/systems. Businesses in regulated industries need confidence that the AI they deploy for mission-critical decisions is trustworthy and reliable—from securely using proprietary data to offering insights into how decisions are made, while ensuring compliance with regulations and preventing biased or harmful behavior.
Organizations should adopt a well-planned, gradual approach to incorporating Agentic AI. This includes identifying business value for potential use cases, defining personas and goals, setting risk appetites, updating risk assessment processes, and implementing controls to effectively manage AI-specific risks. Starting small and refining the approach ensures scalability and trustworthy adoption.
Also Read: Digital Inclusion by Design plutosONE’s Bharat Centric Playbook
With many enterprises using multiple AI models across diverse cloud and on-premises environments, coordinating governance and ethics initiatives becomes both critical and complex. This is where choosing a trusted partner like IBM Consulting to co-develop and co-execute these efforts is essential.
Elets The Banking and Finance Post Magazine has carved out a niche for itself in the crowded market with exclusive & unique content. Get in-depth insights on trend-setting innovations & transformation in the BFSI sector. Best offers for Print + Digital issues! Subscribe here➔ www.eletsonline.com/subscription/
