Opening Address by SMS Tan Kiat How at Snowflake World Tour Singapore

Distinguished guests

Ladies and gentlemen.

1. Good morning. It is a pleasure to join you at the Snowflake World Tour Singapore.

2. The theme for today is “Making AI Real.” I think that is exactly the right question, because AI is no longer something we are waiting for.

   1. It is here.

   2. It is everywhere, as pointed out by Jenny Koh (Country Manager - Singapore, Snowflake).

   3. It can create.

   4. It can reason.

   5. It can act through copilots and agents.

   6. And increasingly, it can pursue goals with less human guidance.

3. So the question is changing. It is no longer just what AI can do. It is: “What can we do with AI?”

4. I think that points to a bigger shift. AI is becoming abundant. Capability is becoming scarce.

5. What do I mean? You see, models are getting better every day. Access is getting cheaper. New applications are appearing all the time.

6. So access to AI alone will not determine who succeeds. The winners will be those who can put AI to work, learn from it, and adapt as the world changes.

7. That is what I mean by making AI real. Because AI transformation is not a one-off technology project. It is a capacity for continuous adaptation, and that makes the foundation incredibly important.

8. We often talk about data, infrastructure, security, governance, and skills as prerequisites for AI. But they are more than prerequisites. They are what allow AI to scale, and what allow it to keep working as conditions change.

9. Let me start with what we are doing in Singapore at the national level. We do not pursue AI for its own sake. We start with some simple questions:

   1. What are the important challenges Singapore will face?

   2. Where can AI help us create meaningful outcomes?

   3. And what do we need to do to create the conditions for Singapore and for Singaporeans to thrive with AI?

10. Our updated National AI Strategy sets out this direction: deepen sectoral and public-sector transformation, mainstream adoption, strengthen workforce readiness, build Singapore into a trusted global AI hub, and strengthen the wider ecosystem – from infrastructure and compute, to talent, research, governance and enterprise adoption.

11. But our ambition is not simply to deploy more AI. It is to build a country that can keep making good use of new technology. We cannot predict which models, applications, or business models will dominate five or ten years from today. But we can build the foundations that allow us to benefit from whatever comes next – trusted data, secure infrastructure, strong digital foundations, good governance, research and talents, and organisations that know how to experiment, learn and adapt.

12. These are no-regret investments. That is why our National AI Missions bring together government, industry and researchers from important sectors such as healthcare, finance, connectivity and advanced manufacturing.

13. We want these sectors not simply to adopt AI. We want them to become better at adapting with technology – more productive, more resilient, more competitive, more trusted.

14. And we want Singapore to be a place where companies can come to solve real problems, test what works, and scale successful solutions beyond Singapore. Because a national AI strategy only matters when it changes what organisations actually do.

15. Which brings me to businesses and enterprises.

16. For businesses, the technology is moving very quickly. Models are becoming more capable and more affordable. There are more proprietary and open options.

17. So the natural question is: “Where can we put AI?” But I think the better question should be: “What should we redesign because AI now exists?”

18. Those are very different questions. One adds AI to the organisation we already have. The other asks whether the organisation itself should work differently. That means starting with the problem.

    1. What is worth solving?

    2. What data and knowledge do we need?

    3. Can we bring them together securely?

    4. Does the workflow need to change?

    5. And how will we know whether we have actually created value?

19. This is where the foundation matters. AI does not create value in isolation. It needs the right data, the right context, the right systems, and the right people.

20. And the foundation has to do more than support a good demo. A successful pilot is not a successful deployment. And a successful deployment is not necessarily a lasting success. Because the world changes, customers change, markets change, rules change, data changes.

21. So the real test is: Can the system learn, adapt and keep creating value? This matters even more as AI moves from generating answers to taking actions.

22. An agent needs current and reliable data. It needs the right context. It needs access to the right systems. It needs clear permissions. It needs feedback. And we need to be able to monitor it and improve it over time.

23. So I would put it simply: The foundation for AI cannot just support deployment. It has to support adaptation.

24. That is why the enterprise journey is moving from data, to intelligence, to workflows, and increasingly to action.

25. But it must be governed action.

26. Healthcare gives me a good example.

27. Healthcare has enormous potential for AI. We have large amounts of data, specialised knowledge, and many processes where better information can lead to better decisions and better outcomes.

28. But healthcare also teaches us something important: A good AI model is only the beginning. It has to work with real data, inside real workflows, with real clinicians, for real patients, under real-world constraints.

29. It has to be safe, secure, interoperable and trusted. And ultimately, we have to show that it improves care.

30. So the hard part is often not building the model. It is making the model work in the system. And once it works, we cannot simply walk away. We have to learn from what happens in practice – improve the data, improve the workflow, improve the model, and keep the cycle going.

31. That is how advantages compound. An organisation that deploys AI successfully today may gain an advantage. But an organisation that can repeatedly learn, adapt and turn the next improvement in technology into another advantage can build a lead that compounds over time.

32. That, I think, is where the real competitive advantage lies.

33. Let me turn to my third point – it is about people.

34. For me, there are two responsibilities.

    1. First, AI should make a positive difference to people’s lives and work - better public services, better customer experiences, work that is safer, more effective and more meaningful. That’s the first responsibility: make AI work to improve people’s lives in home, in society, in the workplaces.

    2. But there is a second responsibility that is just as important. For a lot of you in the audience, in organisations, this is a responsibility that all of us must take seriously. AI should make people more capable, not simply make them faster. Technology can remove work that is wasteful.

35. But some work is formative. It is how people build judgement, experience and confidence. So as we automate, we should ask: If AI removes this task, where will people learn the judgement they used to gain by doing it?

36. That is why I believe we should automate the task, not the learning. Employers, all of us, need to redesign jobs and learning together. Training has to connect to the technologies people will actually use. And workers need opportunities to practise new skills in real settings.

37. I see it every day in the healthcare sector. You want to automate decision-making, decision support, optimise diagnosis, but you must ensure that the nurses, the clinicians and healthcare professionals are able to gain the skills, knowledge, experience and importantly, judgement, because that’s where the value lies.

38. Another good example is the partnership between Japfa and Nanyang Polytechnic. Japfa has built a regional data platform on Snowflake to monitor farming operations across Southeast Asia. Students have worked on applications including image-based farm surveillance and a smart assistant for workers.

39. What is powerful about this example is that it creates value on both sides. The organisation becomes better today. Students learn by solving real problems, with real users and real constraints, and the company develops people who understand both the technology and the business.

40. The best transformation does two things at once: it improves the organisation today, and builds capability for tomorrow.

41. So in conclusion, when I think about “Making AI Real”, I do not think it means simply putting more AI into more places. I think it means building the capability to continually turn technology into value.

    1. At the national level, we build the foundations that allow us to adapt.

    2. At the business level, we have to have organisations that can learn, redesign and improve continuously.

    3. And at the people level, we build the skills, judgement and agency to adapt alongside technology.

42. We cannot predict exactly what AI will look like five or ten years. But we can prepare for it.

    1. We can build trusted data.

    2. We can build secure infrastructure.

    3. We can build capable organisations.

    4. We can nurture people who know how to learn and adapt.

43. Because the biggest advantage may not belong to those who predict the future the best. It may belong to those who are best prepared to adapt when the future arrives.

44. That is how we make AI real.

45. Thank you very much.