Mitr Phol Group: Orchestrating Digital Transformation from Farm to Factory

In this episode, we are joined by Khun Athikom Kanchanavibhu, Executive Vice President of Digital & Technology Transformation, Chief Information Security Officer (CISO) Mitr Phol Group, Asia’s largest sugar producer and a leading diversified agribusiness. With a background that bridges consulting, business operations, and IT leadership, Khun Athikom shared how Mitr Phol is executing enterprise-wide digital transformation across its operations. We explored the group’s two-pronged AI strategy—combining low-code, organization-wide adoption with in-house development of vertical AI models for farming and manufacturing. Khun Athikom also discussed how the company embeds AI into sugarcane yield prediction and factory safety systems, leverages risk-based cybersecurity governance, and transforms traditional IT teams into agile innovation hubs. This podcast was recorded live at the AIBP #AskMeAnything session.

Founded in Thailand, Mitr Phol Group spans businesses in sugar, biomass, bioplastics, wood substitutes, logistics, and construction materials. Its transformation journey is rooted in aligning digital tools with measurable business impact, workforce enablement, and long-term competitive advantage in both agriculture and industry.

Transcript

AIBP: 0:00

Innovation means making AI practical, accessible and measurable, connecting it to clear KPI and business outcomes.

Athikom Mitr Phol: 0:06

So for me, point we do Pyro one is the low hanging fruit, and now we have around 255, 1000 employees using the AI. So they feel like they feel like this is the tool that they can use, and they start to feel familiar with it. And then another part, we continue to build the vertical AI, and then start to make the impact.

AIBP: 0:31

To move fast with AI, enterprises must balance flexibility, ownership of data and modular system design.

Unknown: 0:37

Apple's

Athikom Mitr Phol: 0:37

Approach that we use is build something that we can plug and play. So we might build a platform for farmers that level it generative AI to explain data into the natural language. So right now, we may consume open AI, but we build a platform in a way that it can service the models between Gemini, IBM, Deepseek or whatever. So I believe that approach from enterprise, we start to consider Enterprise Architect, especially in AI area, that we will build a core that when we consume Gen AI, we make it flexible enough that we can switch between different modules.

AIBP: 1:22

AI success depends not only on algorithms, but also on strong data governance, people capability and continuous iteration.

Athikom Mitr Phol: 1:29

And if organization don't have the good data governance and jump into the AI part, they will have a lot of problem to to resolve the technical debt around the data that they need to make it good enough to make AI reliable, and that will add up cost and time to the projects, and eventually people will give up because they never get the good result from the AI models. Even the model itself is really good, but the data and everything else is not ready to consume.

AIBP Intro: 2:04

The AIBP ASEAN B2B growth podcast is a series of fireside chats with business leaders in Southeast Asia focused on growth in the region. Topics discussed include business strategy, sales and marketing, enterprise technology and innovation.

AIBP: 2:29

Please enjoy this episode from the aibp ASEAN B2B Growth podcast.

Sam: 2:35

Well, everyone settle down. I think we have quite a number of like, quite people that we already know. There are also a lot of people that we might might not be familiar with us. So just a very quick introduction. We have been around for 13 years, focusing on driving growth innovation. Throughout this past 13 years, you will know that we have a lot as for every consulting business, we have a lot of contextual data workflows. And today, actually, in preparation for this session, I kind of query our chat bot that has some history, history about our relationship with articon. Actually, I think let me just look through that. What the chat bot say? Ah, we havestarted working together since 2019 that time you are leading the it for share human material company in Thailand. And then in 2020, 22, COVID times you were in Zurich Pharma. We were doing some supply chain activities. And then now you are at Mitr Phol. And I think maybe it's interesting for our audience. They're all here to listen to you, your journey to today and your role in this organization.

Athikom Mitr Phol: 3:43

Okay. First of all, thank you so much for having me here today. I'm really nervous. I'm really nervous, actually. So anyway, glad to be here so my career as soon maintenance cup, first half of my career has been on consulting side. So I start my career reading St Joe's, and has been on consulting track for half of my career. And second half of my career I like so mentioned before that part, I joined yesterday to lead business applications. So my background is more or less have technology side and have business side. So my role clearly is to drive the transformation program for Mitr Phol Group. So the group is conglomerate from from Thailand, but we clearly operate in nine countries. So we started business from sugar that expanding the business to bio renewable energies, external logistic, wood subsidy material, construction material, including bio plastic, bio chemical and related business to the groups. So the road is quite interesting, and it's kind of use different bit of my background. So it used a bit of the consulting part, because they need to drive transformation and deliver the top line impact, bottom line impact, to the CEO and business. And also I need to manage the change in organizations, and also lead the digital data AI and security part. So it's like orchestated that connect between different stakeholder in the in the company, both in Thailand and different countries in our subsidiaries. Yes, yes. Did I miss anything?

Sam: 5:37

I'm not sure, but I think like he sounds, he goes on, he went on. And I was thinking, does he ever sleep? Do you sleep? You sleep? Right? Okay, at least he's in the plane just now. So I think connecticon kind of highlighted a few things, right? The digital part, the risk management part, the cyber security pillar that you're focusing on, maybe walk through a bit more about like, what are some the because today's session, we are focused on AI talk. A little bit more about, like, what are the AI initiators within midpoint group?

Athikom Mitr Phol: 6:07

Yes. For Mitr Phol Group, it has been like initially doing digital transformations like most organizations. So we started doing that since 2018 so so we run a big digital transformation ideation workshop with all leadership teams and have the roadmap. And then we have been doing that for, like in the past four to five years. And then once we digitize and start to have data in the data lake, then we start to move on to AI initiative in summaries. The high level is that we do the farm to table transformation from from upstream, the farming activities to cooperation, the manufacturing, the marketing and B2B and B2C part, maybe the AI initiative. Aside from what we try to do from upstream to downstream, we mainly focus now on the upstream, the farming activities and also the manufacturing side for the farming area is like as we are in agriculture business. So the AI that we doing is more on the unstructured data, data with the farming satellite image processing. So we operate the sugar key land with the farmers around 400 hectares in Thailand, and not including the others land in Indonesia, Australia and China. So with that, the leverage AI to drive the the operation from the plantation planning and also monitoring the progress of the sugar cane, because in the past, we need to have like hundreds of people to engage with few 1000 of farmers to monitor the progress, engage them and do the suggestion how to improve the yield, what to use the fertilizer and all that, and that take a lot of effort, and we have a lot of challenge because of the seasonality of the La Nina and in your that have a lot of impact, impact to the supply of sugar into our sugar news. So we explore the AI, how it can improve the operation, similar to how Brazil has been doing in this area. So we have the in-house team that does the MC scene. Why you are smiling?

Sam: 8:31

Wondering whether it's time for promotion to support on your

Athikom Mitr Phol: 8:35

Okay, okay, yeah, so we forgot already in vendors. house team, so we built in house team to do the So initially, we try to learn from what other leading vendors has been doing in Brazil, Israel and different countries. But after we do a lot of evaluation, we design, we're going to build mostly the AI model in house, so we start to build the AI team to develop the models, but we also explore the option to partners with the vendors as well. So we develop the models and start to deploy and use it to the crop predictions and also harvest monitoring that helped the sugar cane team to improve the operation, especially during the highway season. Usually the highway season for Thailand will be from November to February, and this is like Lazada and chop a period. So yeah, so you can imagine, right? So every second count and like, we need to really work, allow the clock to have a planning, understand situation in the in the field, which is really vast. So the AI help provide the leaders and the model help feed information to the team to do a better planning, and that helped us win the IDC award last year. So we won the IDC future enterprise award with this AI operation. So it really helped to make business and bu feel more confident in AI that we develop and build that it can help complement the business operation. So that is one area. There are a number of other modules, in fact, in farming, but there's maybe a few of the highlight from the farming area. We also try to push the model to the farmers by embedding the models, like you prediction into the web apps that we put on the live officials and then let the farmer use them. But we don't, don't tell them that this is AI model, but this is just the tools that we have people to explain and then let them use it. Another area is factory. So the factory is like quality control, safety environment. So there are, like 15 modules that we are building in Sugar Factory energies and also eternal. So they are both those modules that we use to automate the operation, because many of the operation in factory need a lot of human to monitor. Maybe in Singapore, it's not like that, mostly automate. But in Thailand, especially the traditional factory, we still have a lot of human they need to monitor the production line, monitor the camera and all that, and all those operations starting to be automated by the AI. So we work with engineer team and then build the modules based on the engineering knowledge, and put those knowledge into the modules start to deploy. So that's what's happening now for the manufacturing side. Thank you. I

Sam: 12:00

Thank you. I actually got a lot of follow up question, but you can questions, but you covered all of that. I'm wondering whether anyone on the on the side questions.

Athikom Mitr Phol: 12:10

Okay, thank you.

Sam: 12:10

So later, any questions from the floor

Speaker 1: 12:15

As organizations begin exploring AI adoption, What advice would you give on choosing between off-the-shelf solutions and building custom models in house?

Athikom Mitr Phol: 12:22

I think when we talk about AI, classic AI and generative AI, I believe my based on what I experienced, what I would like to comment is that there should be two prong approach. The first approach is try to make the low hanging fruit happen in organization. So you gotta make the design whether you're gonna adopt the out of out of the box solution to use it in organization, or you're gonna build something simple, but make it quick, like GPT for internal organization, that you provide the infrastructure and manage by yourself, so that you have confidence that is secure and then roll out. So I think this is this approach make management and employee feel AI is more accessible, and then let them use often, so that they feel like AI is like nothing too complex, and then they feel comfortable. And another part is the more longer journeys, the vertical AI that we need to build using the AI scientists, data scientists. So this part take longer journeys, and it need a lot of prerequisite like data and digitalization part before we can move to AI training, because it needs a lot of data to do the training. So for me, point we do parallel one is the low hanging fruit, and now we have around 255, 1000 employees using the AI so they feel like they feel like this is the tool that they can use, and they start to feel familiar with it. And then another part, we continue to build the vertical AI, and then start to make the impact. So to approach that we use now,

Speaker 1: 14:16

And what changes or improvements do you wish to see from Ai vendors and partners to better support innovation and co-creation?

Athikom Mitr Phol: 14:23

Okay, I think when, when we talk about AI related in in the past, when we work with partner and vendors, because they used to be in vendor side as well, the approach that we use in the past is we need to establish a project, the budget, scope of work, and try to make the AI project become like one project. So you need a clear to scope budget. That approach is quite difficult now in this day, especially for like for us. So because many of the AI project are like many of them still trial and error, meaning we know what we want to do, but we are not sure yet what is the end state look like. So we cannot really tell how much effort will be required for the projects. So we cannot say, like, Okay, for this project, this is the budget. So what I think will be useful if the partner have a new business model that is more flexible to engage with the customer, for example, you may partner with the another, vendors, cloud providers, that may have funding for the innovation projects, and then do the MVP or pilot programs without the need for customer to pay per budget, and then get it approved maybe 1 million USD, for example, which sometimes can be difficult to justify when we don't have the end pictures yet. So this is getting the flexibility around how we build a business case, business model and scope of work, maybe something which is useful for the vendor, I think because, for example, we are now running many AI programs, and we explore the option to use the students and university. So we do a lot of MOU now with university, because a lot of talent AI when they are juniors the course is really low comparing to the top tier vendors. So we leverage a lot of the resource from university now to work on the AI project with us. But I believe if the vendor has more flexibility to deliver without a clear score, that will be that will make things happen easier. Maybe, if you can make it like based on the business impact, if you can save cost this much, then you have the performance sharing return to the business. That is something probably, if possible, it will make a lot of CO innovation happen easier,

Speaker 1: 17:17

Given the rising complexity in both cybersecurity and AI deployment, like, what cybersecurity consideration should both AI developers and enterprises keep in mind when building or like deploying AI solutions? And second part is like that today's saturated AI market, like where large companies often face tool duplication, like, how can your organizations effectively manage vendor consolidation and platform governance while ensuring security and innovation?

Sam: 17:45

Do you remember the first question about to ask you? I was about to say that I forgot to take note, about the security

Athikom Mitr Phol: 17:53

For the first part, Security, yes, but, but I can commend is that, from my point of view, we drive security based on the corporate risk. So it didn't drive from the security for the security, but we drive from the corporate risk. So we do risk assessment and see what are the area that we have risk because different industry, different business, have different risk profile when we talk about our data and cyber security. So that is the approach that we use, and we combine them with NISD and all that ISO, and then we do the assessment. So the approach is that for the basic until Foundation, we make sure that we cover all the core components of this cyber security, from the protection, detection and respond and all that. But if we talk about a higher tier, like advanced cyber security, AI or data related, we do risk assessment, and based on that, we invest based on the area that have higher risk. So on annual basis, we do the risk assessment of the cyber security across different information, asset and system or business process, and then based on that, we do a mitigation plan or cyber security investment based on the area that have higher risk that needs to be managed. But personally, I use the hybrid approach, meaning 80% of cyber security are invest in the same field platforms, meaning only few vendors that cover like 70 or 80% of the cyber security domain so that we can minimize the investment while trying to cover as much area as possible. And then for the remaining 20% we level as brace upgrade solution to close the gap. So this mix of the hybrid seems to have an optimum PCO that does not require you to invest in so many technologies, but still have the optimum level of protection and risk in terms of the cyber security, I believe this is the cyber approach that we use. But in terms of the AI chairman, many to two side. I think from the vendor side and from the corporate customer side. I think from the customer and corporate side. I think more and more we treasure data a lot, and AI has a lot of dependency on data. So I think one thing as for us, we always talk with vendors that allow us to have control of the data that relate to the AI projects that we work, for example, for farming area, we cannot work with vendors that they provide AI platforms, and all the data need to be hosted on their platforms, and we will lose all the access to our data, because we believe data is a business asset and it leads to competitiveness of organization if the vendor can have the AI models and data that we use, they can provide similar service to other competitors. So we really keen on access data and protect the data. So I believe this is one thing that vendor site should consider when, when they provide AI solution, maybe the option to leverage AI to API. Other approach beyond just hosting AI, everything, data, everything on your platforms, and another approach that we use is that, as you mentioned, many AI providers have similar models, a lot of duplicate similar models. Approach that we use is build something that we can plug and play. So we might build a platform for farmers that level it generative AI to explain data into the natural language. So right now, we may consume open AI, but we build a platform in a way that it can service the modules between Geminis, IBM, deep sea, or whatever. So I believe that apple from enterprise, we start to consider Enterprise Architect, especially in AI area, that we will build a core that when we consume Gen AI, we make it flexible enough that we can switch between different modules.

Speaker 1: 22:53

What are the key success factors and common failure points - none you've experienced in managing enterprise AI projects?

Athikom Mitr Phol: 23:00

Yeah, I will say it will also happen in in our case as well. I believe there are part that we can manage and part that we cannot manage, meaning, even we manage it, we will fail somehow. The part that we manage, I believe, is clear, objective. How do you measure the success of the AI projects? For example, if we are building the project for the manufacturing AI, we are aiming to reduce the manual effort by 90% we might say we need accuracy higher than 98% for example, that is the objective that we need to make sure business and technology side set that target and Ally together. Or even better, set as a KPI or ok for both side, so that will make sure that there is a commitment from business and technology side. Second part is around people and skills. Most organization rely on, I'm sorry, but most organizations rely on vendors, and when the vendor complete AI project cannot continue to evolve, because usually AI need to evolve over times. So when we have the first module completed, we need to continue to improve. So we need to build the internal capability as well to sustain the longer development of the AI. So we, that is what we try to build internal team so the people, both the technical team and also the business user, make sure that the business feel comfortable to use. And management need to make sure that they are not using AI to reduce people. So that is. The message that management, the HR and people need to align in terms of the implication to people, and then for the overall cost management and data, especially for data governance and data management, because I believe many organizations, when they build the AI projects the accuracy and the power of the AI rely on the data, and if organization don't have the good data governance and jump into the AI part, they will have a lot of problem to to resolve the technical data around the data that they Need to make it good enough to make AI reliable, and that will add up cost and time to the projects, and eventually people will give up because they never get the good result from the AI models. Even the model itself is really good, but the data and everything else is not ready to consume. But for the failure part, I believe many of the business cares that we aim. Could it serve by AI or not? We try them, but some really cannot solve. For example, among farming projects, there are many AI models that we develop. There are some area that there are technical limitation that we already try, and we cannot get accuracy high enough because of the data. Again, maybe not good enough data from satellite image. If you need to get it good enough, it can be super expensive, and it will not justify the investment, for example. So there will be some area that I technically may not ready yet because of the cost to build it, the readiness of the technologies. So there will be two part, I think, but in chat, there will be people data and what else

Sam: 27:02

People data and what do you want me to do in the end, what you're saying is that you need accuracy. Technologies may not be ready, so there are some challenges that you face and correct. I here,

Speaker 1: 27:23

with your With your journey through your current work, has there been organizational or operating model changes because of solar technology?

Athikom Mitr Phol: 27:42

Yes, I think there are two part one organization chair is the technology teams. Another part is the business team. I think first of all, the technology team. In the past, go back like seven years ago. The team is like IT organization, the classic IT organization, business, application, team, SAP, infra and blah, blah, the classic IT team. What happened is that there are a lot of new competency and organization that really structures. At the moment. We structure the team as like Transformation Office, so meaning there is a business transformation office that connects between leadership teams from CEO and our head of functions to align business strategies and technology roadmap and make sure that kind of alignment. So this is a new area that we built. So it's kind of internal consulting business that need to develop the roadmap and alignment with the business. So this is the first team. Second new organization is data data scientists, AI engineers and agile DevOp team, digital product teams. So in addition to the standard software like SAP, Salesforce and all that, there are a lot more digital product concept that we introduce around any application beyond the core SAP or basic IT solution. So this is a newer team, so digital product management, UX, UI and all the agile team, DevOp team, data science, AI scientist. So these are the new team that we set up, and we try to split them between delivery teams and after go live team. So once the this innovation, agile team complete delivery, then we hand over them to the classic it operation. So that team mostly still operate the classic it is. And all that. But they also start to support the new digital products that have the SLA and all that basic governance that this team don't like to do, so we hand over them to the classic IT team to operate so that it can sustain the new solution. And the last area is around government risk and compliance. So this is like the internal audit, make sure that everyone is a good people, that they don't cut corner here and there, and make sure that we deliver things that we propose or commit with the business. So there is another GRC team that ensure everyone has a good standard, and then report when they report something, we check if they really did that properly or not, something like that. So that is how we change in terms of the digital and technology team. So it changed from classic it to like transformation unit that drive the business impact, and then focus on new innovation and delivery. But in terms of the business side, we kind of work a lot with the HR that drive the digital capability for the business. So what we focus is around sales service approach so that we empower the business to do local development, workflow, automation, dashboard, AI, by themselves as much as possible. So this is where we work with Asia to develop the programs, provide the tools, and then onboard them to like, innovate by themselves. So we have around 600 business user now that do a low code, automation and workflow by themselves. So that part also helped, because the workload, I mean, the expectation from business now very high. They always say we are too slow to deliver, so really need to turn the research user to become technologist and do things by themselves and help the team focus on the complex and big impact area. That is what happens.

Sam: 32:12

Maybe a follow up question to that is that the team that are new, the jowl team, the GRC team, are they typically moved out from IT Team. Or are they new hires, new members?

Athikom Mitr Phol: 32:33

Okay, okay, oh yeah. Another

Sam: 32:37

Answer me first, then

Athikom Mitr Phol: 32:41

Okay, I personally always use high hybrid, but when to answer your end your question, no, for like GRC, we have a mix of the old employee, any long year of service, employee, and the new one, because the one that stay longer with organization, they tend to be good in operation, so meaning they understand the business. They know what happening for like over 10 years. So and they are good in operation, they probably not good in innovation, but usually they are good in operation. Follow the standard make sure that nothing forgot, comparing to the new one. Demo, we cut corner here and there because they want to finish work fast. So we mix them together so that someone will lead the new process that need to return like a governance, data governance, we need someone who can like lead new area. But at the same time, when we enforce, we need someone to like, have the ruler and we follow people, have you do this or that, and not get bored about doing this kind of thing. So so we mix them and for but for the digital team, mostly. I mean, people it

Sam: 34:05

first, yeah, which got it first, yeah,

Speaker 1: 34:09

which capabilities or competencies are required to hire right now hiring somebody who can haven't?

Sam: 34:22

Are you volunteering yourself?

Athikom Mitr Phol: 34:25

We are hiring VP it now. Anyway, I think what we are looking is like those who have the T shaped competencies, so meaning someone who has good technical capability but also able to connect with business. Yeah, so this is the area that they need to speak business understand technical good enough to connect between business and technical. I think this is the area that. Um, has a lot of value for us, which we don't have many but, but we try Yes.

Speaker 1: 35:15

Could you share more how you measure ROI of your internal digital transformation initiatives, particularly across financial, operational and people need development metric and like how your hybrid cloud infrastructure strategy supports these transformation goals?

Athikom Mitr Phol: 35:29

For the first part, we use the transformation balance scorecard.

Speaker 1: 35:37

Oh yes, the team, the transformation team that you just the team, the transformation team.

Athikom Mitr Phol: 35:44

But we kind of at first, we have the business strategies, business goals, and then we use balance scorecard concept to link business goals to the measure. So under the measure that we use the four categories. The first one is the financial impact. So this is the top line impact, cost impact, and all that lie in the balance sheet and profit and loss. So these are the first group of the KPI that we used to measure. So if something can be linked to the first group, it will be very good. So we try to put them in this area. And usually the team that I manage there will be cascade with this first financial impact target. They need to somehow think, because in the past, they believe their work is only the system, no impact to the financial but they need to think a lot more, and then eventually they will figure out the impact to financial somehow. So this is the first area of impact. Second area is around the experience management. So it could be farmers, supplier, employee, customer and consumer. So this is a second measurements that we use whether we do this and will it impact to some stakeholder in our solution or not. The third one is

Sam: 37:09

scorecard for the second part. How do you measure that impact? There are hundreds of so you do surveys. How do okay?

Athikom Mitr Phol: 37:18

How do we survey? We do two dimension. One is the footprint monitoring. So how much we onboard farmers to the digital platform. So we track the bridge, the acquisition of how many we get, and then second passes, rather NPS farmers, NPs of customer NPS of employees. So we do surveys, and basically we use MPs to collect how much they love us, yes, so the NPS isokay. The third metric is around operation improvement. So this is where we track how much we lean operations, how much we automate the operation, number of hours we can reduce or save from adopting the technologies. And the last part either are the people, how much we develop people capabilities. So they are a number of approach we use now. First is around number of AI, literally employee. So this is what we track, how much employee is able to use AI in organization. Another part is we use the black back billable build clean bill, get low Bill concept to track the number of employee who can do the digital tools like local development workflow automations across different views, and then work with HR to help the program to develop higher Build in organization across different teams. So this is the measure that we use to track the impact and people aspect. And the second question is

Sam: 39:11

about on prem and hybrid,

Athikom Mitr Phol: 39:14

We clearly use hybrid. Yes, I really love hybrid. So everything is hybrid cloud as well. So we are running mainly 1/3 on Azure, 1/3 on AWS, and 1/3 on pirate cloud, which we use, HPE, glid, Pirate cloud, and then we combine together become a hybrid cloud management. So yeah, because we sometimes we need cloud agility. Is sometimes we need on prem or private cloud latencies, more control over our resources. So yes, we use hybrid cloud between a sure AWS and delete for the pad with Cloud part. So, yeah, I really love hybrid Yes. I guess hybrid car for everything.

Sam: 40:06

Yeah, hybrid car as well. Yes. Last question,

Speaker 1: 40:12

How does your firm manage collaboration with external partners for AI development, particularly in terms of like, data governance and deployment environments and what factors like determine whether midpoint builds AI solutions in house versus adopting off the shelf platforms or services?

Athikom Mitr Phol: 40:31

The first part when we were AI with external window, assuming most of the AI project running from internal environments, so meaning all the AI models are deployed on our environments, but the partner that we work with, they provide expertise to develop the model with US. So like, when we work for the farming area, we work with vendors who have, like, agriculture expertise. We work with Professor in the University who research on the agri tech, like how they use satellite image to track the growth of the different crop type. How do they use the sensor technique to scan to the cloud to get the pictures like that, So I think most expertise partnership that we are working at the moment, but when we explore the platforms, mostly off the chill consumption to API. So it's like they provide a standard API, and we consume that AI service to the API, so that we still level some of the module from external but data are still mostly with us. So this is how we operate in terms of AI with the external partner. Second question is

Sam: 42:06

to AI system in house?

Athikom Mitr Phol: 42:08

Ah, yes, we separate between if it is something that create business competitiveness, we will do it in house. If it is something that everyone just have it like productivity improvement, we will not build in house. We will just use mostly off the shelf solution. So this is something that we need to make it quick, so we just buy and then acquire and get people on board. But if something, they create a competitive advantage that our competitor may take time to develop the we will make this in house. So it depends on the the business implication,

Sam: 42:54

yeah, because I'm very thirsty, so I have to end this session soon. Maybe last question for you, right? What'smost exciting for you in the next one year in terms of technology? Chai

Athikom Mitr Phol: 43:10

yes, it will be boring, because it's about AI and technology Yeah, of course, everyone will be like

Sam: 43:15

yeah, that's what everyone agentic AI, definitely. So the team need to, because I asked all my team to really explore agentic AI now, and believe it will have a lot of implication, because the management the board, also have a lot of hide from their peers, I believe, because when the board and CEO talk to their peers, they always like got the with each other. They say that they have this right so they have a lot of expectation and need speed for it to be delivered. But I believe, agent, think AI will be a key change that will happen this year and next one or two years with technology probably mature in the next year, but this year, we really need to explore and understand implication, both the standard agent and those that we evolve from the RPA tonow people call it intelligent people how a genetic process automation, so APA, so that is something that the team are exploring, how they can make APA become APA and also level as agentic AI across different areas. Okay. Thank you very much. I think I'll leave more questions to the networking. You can try to ask the less boring questions, like, where's the best from Yang Gong and Van Gogh and all I know Kun adiko, he rejected an IDC event to be with us, but because I told him you all would be a fun people. So please leave us to his standard and that. Thank you, Paul.

AIBP: 45:03

We hope you've enjoyed the episode. For more information about business growth in the ASEAN region, please visit our website, www.iotbusiness-platform.com.

Unknown: 45:16

you.