Time to ‘go all in’ on AI for the BFSI Sector: Thoughtworks

India’s Banking, Financial Services and Insurance (BFSI) sector is undergoing a major transformation, driven by the rise of Artificial Intelligence (AI). Industry reports suggest that AI could reshape 35-50 per cent of jobs in Indian banking. Adopting this technology is necessary for banks to manage cost structures and boost productivity.

A series of discussions, titled ‘AI: The next leap in BFSI’, saw Bharani Subramaniam, CTO for India and the Middle East, Thoughtworks, and Deepak Bhatia, who leads the BFSI business for the region for Thoughtworks, share their insights on AI’s impact in the sector from foundational changes to its strategic role in future growth.

Detailing the foundational shift

The most fundamental shift AI enables is genuine hyper-personalisation. Unlike traditional software with its fixed user flows, AI gives banks much greater flexibility at runtime. This enables institutions to address each individual’s needs uniquely, moving beyond grouping them into small customer clusters. This innovation is driving smarter applications from proof-of-concepts (POCs) into full-scale production.

Operationally, AI is already enhancing critical business areas. In customer experience, AI enables sophisticated interactions across various channels, often retaining a “human in the loop” to ensure quality service and oversight. For loan approvals and underwriting, the process is faster and more accurate as AI efficiently analyses the extensive documentation required in lending. Even in areas like risk management and fraud detection, newer AI models are providing enhanced capabilities.

Bhatia highlighted the industry’s accelerating maturity, and said, “We have a number of pilots that people have already tried and now people have already scaled that into production when it comes to either loan approvals or customer experience.”

Overcoming legacy and preparing for scale

A significant hurdle in the Indian BFSI sector is the presence of widespread legacy systems, making the integration of cutting-edge AI challenging. Subramaniam felt that with legacy infrastructure, the main difficulty lies in managing data. “How you deal with data and how trustworthy is the data and what’s the lineage of data,” he said, noting these issues must be resolved for a successful transformation. Thoughtworks is addressing this by using GenAI to accelerate modernisation. By analysing legacy code with AI tools, the transformation process can be drastically reduced from years to just weeks or months.

For systems that cannot be immediately overhauled, a tactic involves embedding “agents” within the existing legacy flow, thereby creating a safe, contained environment for new AI applications to function. To successfully scale AI adoption, institutions must first focus on foundational steps – firstly, boosting AI literacy across the entire organisation to manage the overhyping of the technology and ensure teams understand how AI should solve problems. Secondly, ensuring adequate access to tooling and infrastructure is provided to support the new AI-driven processes.

A future with Agentic AI

Looking ahead, the strategic future of AI in BFSI is dominated by Agentic AI, where solutions possess a degree of autonomy in their workflow. While this excitement is real, Subramaniam advised caution, stating, “I think the banking sector is actually finally waking up to this fact that not all problems have to be agentic.” He felt that internal applications are ideal for testing agentic systems, whereas external, client-facing applications often benefit more from intelligent, but still defined, workflows.

Bhatia strongly urged BFSI leaders to “go all in” on AI, integrating it as a core component of the business across three dimensions, namely Business Model Redesign, which involves fundamentally rethinking operations and reaching new customer bases through local language interfaces; improving Software Delivery by using AI across the entire lifecycle of building and managing technology; and fostering Cultural Adoption by making AI tools available to all employees for their daily tasks, ultimately cultivating an “AI culture”.

Serving the underserved and meeting regulations

In a country like India with all its diversity, AI also holds potential for advancing financial inclusion by serving the underbanked population. Many segments face barriers such as language, social factors and limited branch access. AI, particularly through local language voice interfaces, can drastically lower these barriers, simultaneously reducing the cost of servicing the last mile for banks.

When considering ROI in the low-margin Indian market, Bhatia proposed a shift: “The way we look at ROI is what do I lose if I don’t do it.” He cautioned that not adopting AI now risks significant loss of market share later. He also advocated for ‘Indianising’ solutions to suit local economic and market requirements. Leaders must also try to balance rapid innovation with strict compliance and trust to operate in a heavily regulated environment.

Institutions must also ensure that AI decisions are transparent and bias-aware, which requires “setting a set of principles for using AI in the organisations to start with” and demanding clarity on internal builds or from external vendors, moving away from “black box” systems.

Realising AI’s full potential necessitates a collaborative ecosystem. Bhatia concluded by urging BFSI companies, technology partners, start-ups and regulators to collaborate on developing an “AI stack for India” that mirrors the success of the India Stack, ensuring localised solutions and shared investments for the collective benefit of the financial sector.

Note to the Reader: This article has been produced on behalf of the brand by HT Brand Studio and does not have journalistic/editorial involvement of Mint.