00:00And we're at a pivotal moment in enterprise technology.
00:03Just a few years ago, digital transformation was the dominant narrative.
00:07Mode, renewing legacy systems, automating workflows and streamlining operations.
00:12But now, a new paradigm is emerging, agentic transformation.
00:16According to GodNow, by 2028, 33% of enterprise software will include agentic AI capabilities.
00:23Up from less than 1% in 2024, that means 15% of routine business decisions could be made autonomously without human intervention.
00:32In Afsiapas, 70% of organizations expect agentic AI to disrupt their business models within the next 18 months.
00:40And here in Malaysia, the government is finalizing the National Artificial Intelligence Technology Action Plan.
00:45Building on the 2021-2025 AI roadmap to ensure AI adoption is not only efficient, but also ethical and equitable.
00:54Joining us this morning is David Ayraki, Chief Technology Officer for Asia, Paris, and Japan, APJ Bumi.
01:00I want to say thank you very much, David, for joining me.
01:02Firstly, let's begin with the big picture.
01:05What exactly is agentic transformation?
01:07And how is it reshaping the way enterprises think about automations, decision-making, and leadership?
01:12So digital transformation was largely about improving efficiency.
01:19So at Bumi, we define agentic transformation as the next chapter in enterprise automation.
01:24So if you think about digital, it built the highways, but agentic AI is the self-driving fleet using them.
01:32And this is because these agents can reason, make decisions, act autonomously, and it's a shift from AI as a tool to a teammate.
01:40So these agentic AIs are adaptable and collaborative, but they require connectivity, clean data, and strong governance.
01:49And that's an area where Bumi excels them in connecting, orchestrating, and managing these intelligent agents.
01:55And as you mentioned, overall, it's to improve efficiency.
01:59And Gartner predicts a sharp rise in agentic AI adoption by 2028.
02:04In Asia Pacific, 70% of organizations expect disruption within 18 months.
02:10What's driving this agency, especially in our region?
02:14There's a wide variety of different use cases.
02:17So for Malaysia particularly, as you mentioned, you have a very strong national AI roadmap.
02:23And with the national AI office, the timing's ideal.
02:25There's a lot of immediate potential in government services and healthcare and education.
02:33So, for example, agents could help automate document reviews in the public sector, or they could potentially personalize learning in schools.
02:42But the key with any of these use cases is to start with well-scoped ideas that lower the risk to the business, but have a high return on investment.
02:54And as Malaysia is finalizing the national AI roadmap, from your perspective, David, how does this align with the broader agentic transformation trend?
03:06It's an interesting question, because if you look not just in Malaysia, but across the ASEAN and AIPAC region globally,
03:14governments and businesses are approaching this topic in a wide variety of different ways.
03:19We have the EU AI Act, as an example, which provides a very formal path and penalties, should data sovereignty and other things not be met.
03:29Whereas for many other countries, it's more around frameworks.
03:33How do we get started with AI and, to your point, do it in a very responsible and ethical way that reduces risk to the business?
03:40And equally, it's not just about governments proposing some of these frameworks, but also higher education and businesses themselves are key actors in this discussion.
03:50And as you mentioned, we're also looking at a very multiple dimension about this, we can see approach from Asia and Asia and even AIPAC.
04:01How can we find, and most of them is, how can we find a better and the best framework that is very suitable with this and in a very ethical way and also reduce the risk of business?
04:11And how is a genetic transformation fundamentally different from digital transformation, especially what new capabilities and challenges does it introduce?
04:22Well, I think one of the big key things in this discussion is trust and trust has to be earned and not assumed.
04:29And the trust comes from transparency, traceability and control, particularly in data.
04:35So where digital transformation is more around automating some of those manual processes within your business,
04:42agentic transformation is more around further automating and having some of those processes be fully self-sufficient.
04:51But without data, you are going to have some problems.
04:56And so inaccurate and inaccessible data erodes that trust and layering AI on top of it only amplifies the problem.
05:04So starting with your data, understanding what liquidity you have, what quality you have, provides you a good foundation to implement AI for your business
05:14and provide further efficiencies and automations in a lot of those digital transformation programs you've already completed.
05:21And putting trust, building trust in agentic AI system to make decisions independently,
05:28what are some of the implications that we cannot see of the astronomy for business operation and leadership accountability?
05:37Yeah, it's interesting because we have this topic of human in the loop as an example,
05:43because with a lot of these processes, you have things like model drift where the AI's answers shift over time
05:51because the data hasn't been updated or it hasn't been trained on new processes or new initiatives.
05:58And so it's not just a technological discussion, but it's also about a leadership discussion.
06:04And leadership's critical here because it's not just an IT initiative.
06:08It's also a very large strategic shift in Boomi.
06:12And as an example here at Boomi, we created a head of enterprise AI role,
06:16and they're meant to align innovation and governance within the business.
06:20And we're seeing that across APJ.
06:22We're seeing companies appoint AI officers or form councils because without oversight,
06:28agents risk bias or flawed decisions.
06:31And it's interesting if I was talking to someone the other day and they liken these new leadership roles,
06:36especially that of a chief AI officer to being the new chief HR officer for AI,
06:42because you wouldn't hire a human without vetting them.
06:45So why are we letting ungoverned AI agents run our workflows?
06:51Definitely, it's not easy as we have to balance all the trust within the system, AI, technology,
06:57and especially when that system adapts poor or insufficient data.
07:04So in this context, agentic AI, what happens to agent AI that acts on poor or biased data?
07:12How can organizations safeguard against this risk, considering that we have to make all of the IT department all aware of this risk,
07:21and then we have to maybe upgrade the system?
07:24So what are the possible scenarios that we are looking at?
07:28Yeah, you've definitely outlined a few of them there.
07:30It's not just a technological shift, it's also a cultural shift in business.
07:36We have a term shadow AI, which we're seeing become more prevalent in the industry today,
07:41because lines of business, users like ourselves, bringing models and agentic AI systems off the shelf into our daily work.
07:49IT may not be across those, so it's getting the skill set and the knowledge across your employee base
07:57to understand what they should be doing with AI and how they should be using it.
08:01But equally on the IT and the technology side, it's putting in those governance capabilities,
08:08so being able to provide explainability and observability to those models that you're using within your organization,
08:13and equally in that layer down, having a look at how you've connected your systems together,
08:20what the quality of that data is, and how you're sharing that data.
08:24And typically that's done via APIs, and APIs are that means for AI agents to securely share information between each other
08:32and connect back into your business.
08:34And definitely everyone, the person who is responsible to overview all of this decision-making anonymously or known,
08:42we're seeing also new executive roles like Chief AI Officer emerge.
08:47What does this tell us about the shift, and what kind of organizational culture and skill sets are needed to support agentic transformation?
08:55How do leaders prepare their teams?
08:58Yeah, definitely. As we talked about, those visionary leaders are treating the shift as a strategic overhaul,
09:05not just a tech refresh, so those leaders are driving that oversight, ensuring that they don't make biased decisions or flawed decisions.
09:16But equally, to your point, across the organization, we have to look at that skills gap,
09:21and we have to understand how do we get our workforce to enable to use AI as part of their day-to-day job,
09:27because it is about AI being that teammate, making you a lot more efficient in your day-to-day role
09:33to enable you to work on higher-value tasks for the business.
09:37And that may come through initiatives within the business themselves,
09:40that may come through higher education, providing courses,
09:44or even through government initiatives to skill up the workforce at a national level.
09:49And David, maybe we can check moving forward, especially for the business.
09:55How can they ensure that agent AI systems remain explainable and auditable,
10:02especially at a scale with the framework that is done or being executed by the government?
10:08Is it enough?
10:11Yeah, I think that's where, when you have those multi-agent ecosystems,
10:14you have to have guardrails across those workflows, not just single models.
10:19Because if you think of use cases such as the patient journey in healthcare,
10:25agentic AI will be embedded to different parts to improve that patient experience
10:30as they enter the hospital or the doctors, you know,
10:34understand what's happened with their x-rays, as an example,
10:37their CT scans to what medication may be assigned to them.
10:41So having agent behavior that's auditable, it's contextual and correctable is going to be key,
10:49especially in industries like healthcare.
10:51If we're talking about retail and, you know,
10:54I'm accidentally suggested or sent the wrong piece of clothing,
10:58that isn't as big of a risk to the business.
11:01And as we know, as many are transitioning to transformation of agent, agent AI,
11:06we hope that all businesses in every category can stay compliant with regulations and ethical standards.
11:14And again, I will say thank you very much to our guest,
11:16David Araki, Chief Technology Officer for Asia Biosphere and Japan APJ,
11:20for helping us understand how can the agent AI and the future of enterprise decision.
11:26And again, definitely all of our discussion here will be featured in astronomy.com
11:30and across all social media platform.
