Executive Summary

The IDC article highlights that AI adoption in Asia-Pacific financial institutions has moved beyond experimentation and is now a board-level strategic priority. According to IDC, 74% of financial institutions have reached managed or optimized stages of GenAI adoption, with primary deployments focused on IT operations, customer service, and finance functions. The industry is also rapidly progressing toward Agentic AI, with institutions already operating numerous AI agents in production and planning significant expansion through 2026. Governance, data quality, and talent shortages remain the key barriers rather than technology limitations. AI is increasingly viewed as a resilience and competitive-growth capability rather than solely an efficiency tool.

Dr. Sekar’s Perspective:

Key Opportunities

1. AI can enhance operational efficiency, automate middle- and back-office processes, improve research productivity, and strengthen cyber resilience.
2. Agentic AI could support faster risk analysis, regulatory reporting, client servicing, and operational continuity.
3. Custom-built AI platforms may provide greater control over proprietary data, intellectual property, and differentiated client services.

Key Privacy & Risk Considerations

1. Governance must extend beyond model risk management to have a holistic approach towards Data Privacy, Agentic AI Usage and MCP standards.
2. Direct customer interactions without human intervention increase risks around confidential information exposure, inaccurate responses, and regulatory accountability. Exception handling and human in the loop should not be removed from all workflows.
3. Data quality concerns are particularly critical because investment banks manage highly sensitive client, trading, and market information.
4. As firms shift toward self-built AI stacks, privacy-by-design, data residency controls, encryption, and strict segregation of client data become strategic requirements.

Bottom Line

For investment banks, AI is becoming a competitive necessity. However, success will depend less on model capability and more on the institution’s ability to establish robust governance, privacy controls, and regulatory compliance frameworks that allow AI innovation without compromising client confidentiality and trust.