Strategic for Banks to Move from Isolated AI Project to Organisation Value

Derrick Lim
Financial Services Industry Leader
Audit & Assurance Partner
Deloitte Thailand

The question facing banks globally today is no longer whether “they use AI”—it’s whether “they have moved beyond pilots to organisation-wide implementation and achieved tangible value”.

Artificial intelligence is no longer on the horizon for banking—it has arrived with force. As Thailand positions itself as a digital hub under the government's Thailand 4.0 initiative, the banking sector is leading this transformation. Deloitte's Thailand Digital Transformation Survey 2025 confirms the sector's leadership in digital transformation and AI adoption across industries in Thailand. In 2024, Thai banks have announced heavy investment over 25 billion baht in AI to automate back-office operations, personalise customer experiences, accelerate innovation, and enhance real-time risk and fraud management.  This reflects years of strategic focus and puts the sector among the most advanced in the Thai economy.

Yet despite this progress, a critical challenge has emerged. Deloitte’s 2026 Banking and Capital Markets Outlook reveals a divergence between “AI ambition” and “readiness” within the banking sector.   Many AI initiatives are trapped at fragmented proof of concept, especially Generative AI. While banks are now entering a phase where AI is embedded across the value chain—from personalised marketing and credit decisioning to fraud detection and customer engagement—yet these systems often operate in silos. The defining challenge ahead is scaling these initiatives into aligned, organisation-wide AI implementation. While this research centre on the US market, Thailand's banking sector will likely encounter similar challenges in the near future. 

Ultimately, AI success depends on delivering tangible, measurable business value. Without it, even the most sophisticated initiatives risk becoming expensive experiments rather than transformative investments. 

Strategic Imperatives for Banks

Deloitte has identified key strategic priorities for Thai banks seeking to move beyond isolated AI projects:

1. Establish a clear and unified AI vision and strategy: Banks must begin by answering a fundamental question: What role should AI play in our organisation's future—operational efficiency, personalisation, or innovation? The answer will vary based on each institution's market position, customer segment, and strategic priorities. Regardless of bank type, they need a unified, organisation-wide AI vision and strategy. An effective AI vision must articulate concrete outcomes; recognise risks, costs, and human implications; align with the bank’s broader mission; be communicated consistently across all stakeholder groups. Otherwise, AI initiatives will scatter across departments with each pursuing different objectives without a shared overarching vision, leading to duplicated efforts and uneven impact.
2. Implement concrete ROI measurement: Thai banks are accelerating AI adoption with aspiration to enhance customer experiences, drive innovation, and improve operational efficiency. However, the real challenge lies in demonstrating tangible business value or return on investment. The study by Evident in 2025 found that only 4 out of 50 banks globally reported realised ROI from AI use cases.  The challenge can become increasingly urgent as boards and shareholders demand clear justification for substantial technology investments amid growing economic uncertainty. This challenge arises as executives may struggle with subjective metrics—such as hours saved. Without standard baselines or KPIs, claimed benefits often rely on user perception rather than verifiable financial outcomes. This creates credibility gaps that make it difficult to connect soft benefits to tangible value like cost savings or revenue growth.

Common hurdles faced by banks in measuring return on investment

• Fuzzy value statements with subjective assessment: AI benefits are often described in vague, qualitative terms—like "work faster" or "serve customer faster"—rather than concrete numbers.
  Suggestion: Tie claim to financial or risk metrics (e.g., time saved → cases processed → revenue impact).
• No baseline: Can't prove AI drives the gains over other factors.
  Suggestion: Use control tests or historical benchmarks for counterfactuals.
• Double counting: Multiple teams claim same gains, leading to overall impact inflation. Suggestion: Centralise ROI validation and consolidation with attribution rules
• “Productivity” ≠ realised savings: Faster tasks don't always convert to cost reductions
  Suggestion: Link productivity to tangible outputs (e.g., more loan processed) 
• No standardised metrics: Business units measure impact differently, creating inconsistency at the organisation level.
  Suggestion: Establish organisation-wide ROI categories and dashboards. 
• Vendor comparison: Third-party AI platforms vary in cost, accuracy, and speed—lacking benchmarks hinders provider comparison.
  Suggestion: Develop a vendor evaluation scorecard.

To move past this readiness challenge and fully capture the promise of AI, banks must make bold choices to move beyond isolated projects.