You and AI: Breaking Boundaries Everywhere | Asia Summit 2025

Panelist Profiles & Strategic Focus

• Hans Dekkers (IBM APAC GM) focuses on hybrid cloud, enterprise/government AI, and quantum computing.
• Grace (Prudential Chief Data AI Scientist) prioritizes embedding AI into corporate strategy to drive value for both the company and customers.
• Rodrigo Lima (Microsoft Asia President) predicts AI will transform every industry and profession, drawing on early experiences with Watson.
• Pavel (Stanford AI researcher) focuses on incubating research ideas to move them from the lab to real-life impact.

Adoption Challenges & ROI Reality

• Industry reports cite significant struggles with AI implementation, including low productivity gains and high costs without visible ROI.
• Rodrigo Lima counters negative narratives by noting 98% of customers buying Microsoft Copilot continue to purchase more annually.
• Grace highlights a disconnect: while AI raises individual productivity (e.g., via Copilot), measuring downstream P&L impact remains difficult.
• The panel argues that 1% of companies reporting "true value" often stems from measuring the wrong metrics, such as focusing solely on GenAI rather than traditional algorithms.

Strategic Implementation Principles

• Success relies on solving business problems through industry expertise and process redesign, not just deploying technology.
• 70% of value in AI comes from traditional AI and propensity models, not just generative AI.
• Effective adoption requires a "bottoms-up" approach (enabling all employees with tools like Excel) combined with "top-down" strategic use cases (e.g., customer support automation).
• Companies must avoid "death by a thousand pilots" by focusing on scalable, proven use cases rather than diluting resources across hundreds of experiments.

Technical Architecture & Data Sovereignty

• Enterprises face five key hurdles: multi-environment management (on-prem/cloud), exploding data volumes, lack of automation, skills gaps, and security risks.
• The future AI ecosystem will feature 10–12 large foundational models alongside hundreds of proprietary, domain-specific small models.
• Small, domain-specific models trained on proprietary client data are expected to become major enterprise assets (e.g., for steel manufacturing blast furnaces).
• A robust orchestration layer is required to route specific tasks to the appropriate small model (e.g., oncology models vs. financial crime models).

Talent, Skills, and Change Management

• The primary challenge is reskilling the workforce; the AI gap is compounding existing shortages in cybersecurity and data science.
• Traditional coding skills are becoming obsolete, replaced by the need for "AI business translators" who possess strong domain knowledge and business acumen.
• CEOs must invest heavily in change management to upskill employees rather than simply cutting jobs, ensuring the workforce can pivot to new roles.
• Power law dynamics in talent acquisition mean top AI researchers will cluster around major hubs, requiring innovative retention strategies for specialized talent.

Global Competitive Landscape

• Asia is emerging as a new center of gravity for AI innovation, driven by strong government roadmaps, conglomerate R&D, and high startup activity.
• Asia now accounts for 70% of global patents and more GPU consumption than the US and Europe combined.
• The region is transitioning from a manufacturing hub to a creator of global digital products and AI solutions.
• Sovereign strategies in countries like Singapore, China, and Japan are lowering barriers to AI deployment and testing.

Future Outlook & Forward-Looking Statements

• Humanization: AI will evolve from logical reasoning to possessing empathy, creating seamless human-machine interactions.
• AI as Literacy: AI will become "the new math," a fundamental requirement for all professionals regardless of their specific field.
• Spatial Intelligence: The next frontier involves models capable of acting in physical environments, fusing digital and physical realities.
• Lab-to-Reality Gap: Only ~10% of AI research currently translates to real-world market impact, necessitating better cross-pollination between academia, startups, and enterprises.
• Investment Horizon: IBM reports saving $4.5 billion in costs after applying AI internally over three to four years, validating the long-term transformative potential.