New Delhi, Nov. 11 -- Even as banks and financial institutions step up investments in artificial intelligence (AI), many are struggling to turn early pilots into large-scale deployments. A new global study and TechCircle's own interviews with Indian industry leaders reveal that while most financial firms have adopted AI, data silos, legacy systems, and governance gaps are holding back full-scale transformation.
According to the research-which surveyed CIOs, CTOs, and technology heads from financial institutions worldwide, including India-97% of respondents said they are using at least one AI or machine learning use case. Yet, only 26% reported enterprise-wide adoption, and nearly 48% admitted they remain stuck between experimentation and integration. The findings echo TechCircle's observations across India's BFSI ecosystem: AI enthusiasm is high, but scaling it remains elusive.
Hybrid AI takes centre stage
A key shift emerging from the study is the rise of hybrid AI models, as firms balance innovation with compliance. Around 62% of respondents said they are deploying AI through hybrid environments-combining public cloud, on-premise data centres, and edge computing-while 91% rated hybrid AI as "highly valuable."
This model, experts say, is helping banks and NBFCs modernise without completely overhauling legacy systems. "Hybrid architectures allow financial institutions to process data wherever it resides - in the cloud, on-premise, or at the edge," the report notes. "They're essential for managing complexity, meeting regulatory obligations, and scaling AI efficiently."
In India, TechCircle's August 2025 survey found that 84% of BFSI organisations are already using AI, with 67% experimenting with generative AI (GenAI) for use cases like fraud detection, compliance, and customer analytics.
Data silos and security slow the shift
Despite near-universal AI adoption, data silos remain a major barrier. About 97% of global respondents cited fragmented data as the biggest obstacle to effective AI. TechCircle's special report, "Beyond the Silo: Why Fragmented Data is Holding Back Indian Enterprises," found that while data volumes have surged, access to and context for that data have lagged.
"Data may exist in abundance, but access is uneven and context is missing. You can't drive intelligence if every team operates on a different version of the truth," said a senior executive quoted in the report.
The issue is particularly acute in India's banking sector, where outdated core systems and inconsistent data policies persist. "Fewer than 20% of banks are in the phase of applying AI to drive measurable business performance," said Ranga Reddy, CEO of Maveric Systems. "Most are still wrestling with data readiness, governance, and platform consolidation."
Security, governance take board-level focus
As AI systems increasingly handle sensitive financial data, security and governance have moved to the top of boardroom agendas. "Agentic and generative AI expand the attack surface significantly. If a single agent is compromised, it could unintentionally expose sensitive data. Security can no longer be perimeter-focused," warned Pratik Shah, Managing Director, India & SAARC, F5.
Globally, 84% of surveyed financial firms said a unified approach to data governance and security across environments is "critical" or "very important." In India too, security is shaping AI investment priorities-25% of firms cited it as the top factor when selecting AI platforms.
"AI can only deliver its full potential when data sovereignty, privacy, and trust are guaranteed," said Gary Wright, Managing Director, Finextra Research, who co-authored the study. "Our results highlight that AI success depends not just on investment, but on infrastructure, partnerships, and governance."
Indian BFSI doubles down on responsible AI
Leading financial institutions are now aligning AI strategies with responsible data practices. "Machine learning tools sift through vast amounts of data to detect patterns, predict risks, and automate routine processes. At the same time, we are investing in strong AI governance frameworks to ensure responsible and ethical use of emerging technologies," said Madhusudhan Warrier, Chief Technology Officer at Mirae Asset Sharekhan, in a recent TechCircle interview.
Adrien Chenallier, Global Director, AI Solutions for Financial Services at Cloudera, added, "A data-anywhere, hybrid strategy is non-negotiable. Infrastructure alone isn't enough - financial institutions need unified data and AI platforms that ensure consistent governance and security across environments. That's the only way to build trust and accelerate AI adoption at scale."
The road ahead
Experts agree that scaling AI in financial services is as much a data challenge as it is a technology one. As firms integrate AI into lending, risk, and customer operations, governance will define the winners.
"Integrating AI with legacy systems and upskilling the workforce are key challenges," noted Anand Krishnan, Head of AI Practice at Infosys Finacle, in an earlier TechCircle interview. "Institutions that can unify their data and talent strategy will move faster from experimentation to enterprise AI."
In essence, India's BFSI sector is past the AI trial phase but not yet at scale. To get there, firms must turn fragmented data into trusted intelligence, security into confidence, and AI pilots into measurable business outcomes. Until then, the industry will continue to navigate the delicate balance between AI promise and operational reality.
Published by HT Digital Content Services with permission from TechCircle.