Southeast Asia’s AI Moment: Turning Innovation into Boardroom Accountability
Frank Bignone, FPT Software Senior Vice President and Head of Corporate Strategy & Growth, FPT Corporation
Across Southeast Asia (SEA), AI adoption in the enterprise is accelerating, with businesses embracing AI at a faster pace than the global average. With that urgency comes a growing tension in the boardroom.
Many CIOs are discovering a difficult truth: the metrics that matter to AI teams rarely translate well to executives. Model accuracy, prompt volume, and automation percentages show progress, but they don’t tell CFOs, CEOs, and directors how AI will increase revenue, drive agility, reduce risk, or improve resilience.
With AI-related investment projected to surpass US$110 billion by 2028, the stakes have never been higher. Boards are becoming more hands-on in oversight, and regulators are turning their attention toward model risk, data governance, and safe deployment. To meet this moment, CIOs need to lead with a new value language, one that anchors AI programs in growth, profitability, risk reduction, and workforce capability.
For technical teams to justify AI investments, they need to speak the board’s language to earn credibility, get funding, and align teams for enterprise AI. CIOs must act now to shift ROI from activity-based to outcome-driven metrics, prioritising business performance over AI model development.
Distilling the Real Value of AI for the Board
For years, AI ROI has focused on efficiency gains, such as support tickets closed, shorter sales cycles, and server consolidation. It’s increasingly automating workflows and reshaping customer experience, revenue streams, and the way compliance and talent are managed.
However, as we transition from basic tools to Agentic AI and a fully AI-augmented workforce capable of completing complex, high-performance tasks, ROI needs to be measured by its value to the business, just as we judge humans. What these systems deliver and what they cost are now central concerns for the board.
Yet reporting on enterprise AI programs still relies on metrics that do not demonstrate spreadsheet value. As boards fund outcomes, not activities, they increasingly want to see:
• Growth: Higher conversion, improved retention, new revenue streams.
• Profit: Lower cost-to-serve, shorter cycle times, improved working capital.
• Risk: Reduced loss expectancy, fewer incidents, stronger controls.
• Capability: Faster upskilling and a more productive, AI-augmented workforce.
Amid the rapid AI implementations in SEA, too many AI dashboards remain filled with model statistics rather than business impact. This makes programmes look like experiments rather than strategic investments. Fragmentation adds to the problem, with non-specific data, unclear ownership, and inconsistent governance leaving organisations stuck in experimentation mode.
The Balanced AI Value Scorecard
A practical step to measure AI ROI clearly is to use the Balanced AI Value Scorecard model, which is easily understood by the board and leverages existing data in a format relevant to decision-making.
The Balanced AI Value Scorecard connects directly to the P&L by measuring:
Growth KPIs: Conversion lift, retention gains, cross-sell rate, lead-to-order cycle inform the board that AI drives revenue, not just automation.
Efficiency: What does it cost to achieve a specific KPI? Measuring costs per order, per contact, and per claim, and which other business metrics help the board understand the cost of performance relative to the investment. Understanding this allows them to see how efficiencies are impacting the bottom line and where improvement is needed
Risk: Understanding data exposure incidents, model misuse, audit exceptions, and loss expectancy is essential for CISOs and boards to reduce the downside and associated costs.
Capability: As AI continues to expand the workforce and understanding its capabilities today, and what will be required tomorrow, is essential to forecasting.
Different use cases may dictate variations on the above, but CIOs should focus on three to five metrics to avoid confusion.
The Outcome-First ROI Method
To shift from proof of concept to proof of performance, CIOs can employ a disciplined five-step ROI method designed for enterprise environments:
Start with a Value Thesis - Define the anticipated business impact in one sentence: “If we deploy [AI capability] in [workflow], we will improve [metric] by X%, unlocking $Y of value in Z months.” Conversely, consider what would have happened without AI, to avoid inflated results.
Establish Baselines - Build 3–6 months of operational baseline data, segmented by channel, issue type, and customer cohort. Use both direct (e.g., AHT, CSAT) and indirect indicators (agent satisfaction, turnover).
Prove It with Tests - Run A/B or holdout experiments: AI-assisted vs. non-assisted. Measure not just performance, but the reasons for AI-to-human handoffs and early failure modes.
Convert KPIs into Board-Ready Dollars - Use simple formulas CFOs trust. Only count capturable savings, such as avoided hires or reduced rework, not theoretical efficiencies.
Govern and Scale With Evidence - Scale only when uplift persists outside of pilots, guardrails are in place, and a capture plan ensures savings flow to the P&L. This repeatable method builds credibility and keeps AI investments grounded in financial reality.
As the region rapidly emerges as one of the world’s most dynamic AI growth regions, organisations need to embed rigorous governance, clear accountability and a shared understanding of value at the board level.
With a scorecard that speaks their language and a disciplined ROI framework, CIOs can elevate AI from isolated initiatives to enterprise-wide transformation.
(JUN/QOB)