280,000 Employees, 200+ AI Agents, and One Thing AI Still Can't Do

Vídeo em Inglês, publicado pelo canal Lyzr AI.

Video transcript

Transcript from available captions; it may contain errors.

economist had the famous quote 3 years ago is that the threat is not AI but your threat as a professional is going to be another professional that knows AI. I have with Patrick today at the Howi Asian podcast. He's a VP and global head of marketing and external relations at Hitachi Digital Services. Patrick, thank you so much for being here with us today. >> Great to be here. I appreciate the opportunity. >> Hitachi has obviously been announced in 1910. massive company 280,000 employees 618 subsidiaries globally and Hitachi digital services is essentially a gsi which is fully owned by Hitachi you've been playing a central role in talking to a lot of analysts really amplifying the brand for Hitachi and yeah I feel like you've been in the space for multiple decades so Patrick that's on the work front tell me more about you outside of work what's Patrick like >> well when you can find the time to get outside of work I like to golf and I like to boat. I'm very fortunate that I live on Long Island, which is, you know, the the waterfront strip of of New York State. And most importantly, I am a member of a civic organization known as the Knights of Columbus, which has been around a little longer than Hitachi, 1885, but they are a Catholic fraternity that focuses on charitable work. So I have the pleasure of leading the local council here where I live and uh that occupies a lot of my personal time fortunate enough to be traveling a lot around from you know from a Hitachi perspective and being with my colleagues in the seats because you know how it is remote work is tough sometimes. So, I actually look at traveling and being with my colleagues as as sort of nonwork because you actually get to be with people that you talk to every day in person. Makes a huge difference, especially as you said in the beginning in the AI era. So, it's it all it all it's all woven together in my opinion. It's really hard to have separation in my in my world. >> Yeah, agreed. I feel like now it's work life harmony. I'd love to ask you also talk about tell me where you were during co and what were you doing pre AI and then finally I would say the last four years. >> So you know to level set Hitachi has been doing AI for over 25 years and so then the question always becomes well what do you mean I didn't hear about physical AI you know in 2011 I didn't know about chat GPT was that stuff still around then? Well, no, but you had machine learning, you had computer vision, right? Even things like IoT, remember SCADA, all that big comparison. So, all these technologies that have a bit of a cutting edge element to them, their earlier forms have always been there, right? And they've been being used because Hitachi has grown up in the OT space with lots of factories, lots of focus on energy and mobility. So you need things like SCADA, you need things like machine learning and computer vision to be efficient in those areas. So for us, you know, from a Hitachi side, it's been around uh and now it's, you know, evolving and that's, you know, through Lumata, which I know we'll talk about. So it's been there. Now from my personal experience having worked at DXC and prior to DXC Luxoft which was one of the earlier versions of the software engineering firm that leveraged nearshore when that was a big thing you know we were doing work in the automotive industry in the telecom industry in the manufacturing industry that was super cutting edge leveraging things like machine learning computer vision augmented reality you know aspects of the general AI umbrella of things that came together. It just wasn't wasn't really called that as a strategic goto market pillar at the time. >> Right? >> So, I've seen this stuff, you know, over the last decade just thinking about it from an industry perspective. Now, advancing uh Luxoft got acquired by DXC in uh 2019. And so, you know, we had a we had a strong run there in terms of uh leading up to the famous March 2020, which is when really co shut the world down. And that's when virtual world really came to life, I think. Right. I was in Germany. I was in London in February of 20. I was at an event on like March 12th and 13th in New York City. And that was kind of my last time out with with everybody. And that was when, you know, we were pushing DXC's workplace stuff. That weekend, to bring this full circle, I joined the Knights of Columbus on March 15th, 2020. I had everybody at my house that day after that ceremony and next thing you know every television in my house is on with President Trump, >> Governor Cuomo, everybody cuz I had my cousins are teachers and you know different public sector jobs. Everyone is worrying about work and that afternoon that's when it all hit us. So for me it's kind of interesting how everything always comes full circle in life with co spent two years at DXC really sort of under the restrictions and I couldn't have asked for better opportunity to join Hitachi in August of 2021 when things were starting to open up and really hit hit the ground running. So but DXC also had a had a strong pivot into the AI space especially in some of their core sectors like insurance. So for me, you know, I've been around it, right? It just hasn't been a strategic pillar of marketing the way that we see it today. >> Yeah. So maybe let's talk about since since you since you got into I feel like Hitachi just around the time of like OpenAI and like obviously anthropic so the world was slowly picking up what what do you feel like your role initially was while you joined and maybe the role hasn't changed as much but how is that flavor of AI kicked into your as a pillar is what you brought up earlier. >> Yeah. So from a professional perspective, you know, I leverage Gen AI every day for one reason or another. And our design team here, you know, has various tools that they use to help support them, you know, working with you guys and building agents to support content development and and even some of the analyst relation database building has been tremendous. Right? There's no doubt about it that leveraging, you know, uh, your tools and and just in general from a pure professional perspective is something you need to understand, right? And I think the economist had the famous quote three years ago when things uh started to open up is that the threat is not AI, but your threat as a professional is going to be another professional that knows AI, right? And I think that's why I've been pushing on my team, leverage the tools, learn the tools. And even here at Hit Itachi, you know, globally, there's been a very prescriptive approach to learning and development and certification for everybody. No matter what you're doing, you have to understand AI. Now, that's part one. Part two is, you know, obviously you it's hard to talk about any of your offerings without having a subset discussion about the role that AI plays. Now, as always, when that happens, it has it sometimes gets a little bit here we go, here's the next AI piece, and you almost celebrate those discussions when you're not talking about AI because you really want to focus on something where you know, okay, we know it's happening. We're going to leverage it. It's sort of table stakes, right? The more important discussion is about the industry, the expertise that's needed, the certain level of engineers that are going to be needed to be able to do that work, you know, or the forward deployment approach, right? So it's becoming a part of everyday life in terms of professional usage and image in the market. But I also think that there is a little bit of push back happening that we can't let this take over every LinkedIn post that's written, you know, every every flyer that's drawn up by chat GPT. >> Yeah. >> You know, every strategic idea and what to do next. So the humanization part will always happen and I think it's I think it's starting to develop a little bit now. >> Do you feel like you're in a place where you're driving where humans the alpha and AI is the beta and or is it vice versa and what I mean by that is do you feel like internally or with your team or with you are you mostly driving starting conversations with AI or is the flip slowly happening where AI is telling us what to do and then from there you're deciding what are the next steps. So it's a it's a balance for sure. There are times where you want to leverage it to get nonbiased feedback on your idea where you want to be able to get a logical algorithm-driven evidence-based approach that you just don't have time to do the research and then you get the feedback. And that's tremendously helpful because the iterations that you used to have to do before you went to market in planning have totally reduced which allows you now to be creative which allows you to be more aggressive which allows you to spend time and money on actually getting to the people you want to get to because we all know that's even harder than it's ever been before as well. So that so that's the part where I think it's been tremendously helpful is really shortening that planning cycle. On the on the other hand, I don't want us to stop being able to write. I don't want us to stop being able to have creativity and design, right? The creative element, and I mean that in a broad sense, not just building powerpoints and putting together social media tiles. That is a human trait that the computer will never have no matter what they say. Okay? And we've had this discussion time after time through every type of every world that happens from the enlightenment, from the Renaissance, the Middle Ages, whatever it might be. That element will never get beaten out by a machine. But you have to understand how to tailor the machine to help >> expand, augment, support, and build upon that. I'm I'm in a course right now with Kellogg for accelerated marketing. And one of the things I love most about it is that AI is not a part of every element of that discussion. It's a piece of it. how do you leverage it? But each of the tiers that we've gone through, I think it's been like 13 weeks now, it just focuses on how to improve each specific part of what you do as a marketeteer. And it doesn't always say rely on AI to do it, right? It makes you understand the discipline. So you have to have a balance, you know, in how you use it, when you use it, why you use it, and and what's what's the outcome you're looking for. >> Yeah. You know, one of the things we're seeing in the market right now is like we have, let's say, an agent, we call it Scott. It's like a marketing agent we use internally. We've added it to Slack. So, anybody should be able to ask Scott for things. So, it's like part of a conversation. But the next step is can Scott automatically join conversation. So, let's say Patrick's talking to one of his colleagues. It has enough context and it automatically starts suggesting. So, that's your that's the hope that becomes truly a co-pilot experience versus you asking it, right? So, and I think that's interesting because it'll keep it keeps expanding on context of of the organization as it proceeds. So, cool. Maybe switching gears a little bit, Patrick. Uh, tell me some of the use cases that you're excited about from a marketing standpoint. Just some of the things that you're working on versus even what you've seen out there of how AI is being applied in marketing. >> So, it definitely helps from a production perspective. I think you know improvement in productivity is is and to me that's really the baseline for everyone that's using AI but for us it helps a lot in the planning stages. It helps a lot through administrative defeating all of that boring work keeping yourself organized but really productivity and execution I think is the baseline for which marketing can benefit off all of us. And it doesn't it doesn't necessarily need to be agents doing it for you, but it could be a combination of an agent, your use of Gen AI and uh and in some areas where you have lots of budget, maybe there's an applied AI element to it as well. But I would argue most of that's probably in the BTOC space, you know, attracting for for that type of sales. You know, the second piece I see a big piece in, the second area I see a big piece in is the market intelligence world. Now having managed and engaged with analysts now for over 13 years you know certainly their roles are changing and a lot of firms have adapted some faster than others as to how they're going to use AI whether it be things like AI is going to schedule the briefing all the way through to AI is going to synthesize a report about a particular provider that you need on demand and you necessarily can't wait 3 days to get somebody on the phone to to do some discussion with. Now on their end that should be helping shortening that timeline where you could talk to an analyst much faster engage quicker in a workshop and so on so forth. So marketing intelligence in general whether it's done internally by uh you know in our in our case we have a knowledge hub team or it's being used externally is is really the second piece and then to what we were doing together really around content right I think that thought leadership and um you know I would I would say to to some extent your engagement in the media always can help clean up your responses and ensure that there's consistency and making sure what you're talking about is aligned to the facts right? Not just some of the opinions that are out there. So, it's going to play a big role in content development and elevation of of thought leadership. And what we have to be careful of, what everybody has to be careful of is making sure, you know, that that what's being produced has primary and secondary sources attached to, right? Like the old school journals, articles that were written, you know, peer-reviewed stuff. >> So, and I don't think we've got there yet, right? There's a lot of people writing stuff on LinkedIn that you could tell is AI generated that that some is AI edited and so on. This is I think is going to be a big test for us is how do we make the next world the next uh iteration of this you know stick to the human element and then there are things around governance quality and you know other areas within marketing but but those are really my three you know top top three areas I think. >> Sweet. Cool. So Patrick, I think we were we were talking about and and this came up on the research as well. This the HRC agents is do you have like do you say it as HC or do you see it as HARK? >> Hark. >> Okay. Hark. Okay. >> Yeah. >> So it looks like Yeah. Hitachi kind of launch. It's I'm assuming it's like an agent framework. It's an agentic platform. There's an R2 as well which I think is making sure you can build agents that are reliable and responsible. You also have I'm assuming a forward deploy team which is part of the whole managed services and then there's this multiple pre-built agents across different enterprise domains right so and I'm seeing that happen across a lot of global system integrators as well they have an agent builder you have a managed services team you essentially also have an Asian library how how is the team internally thought about this practice given given there's a big demand for not just like maybe productionization of services but it's also like hey I just don't want to depend on your services model I also want services plus agents for lack of a better word so how's has how's the model changed I guess is the question so to to level set the audience because we hit them with a lot of acronyms there about hark and R202 and and I think this is something that we as the marketing team are always thinking about right cuz sometimes when you get when you build an internal acronym or you build a platform internally your mind everybody knows knows what it is and you sometimes operate like well everybody else must know what it is too right so you always have to take that step back and not over AI yourself when you're going to market about it hark hc which is Hitachi application reliability center is a concept a platform that came out man I think about four years ago which is really focused around taking an engineering approach to efforts and modernization and and your cloud work uh you know versus old school body CAP approach and HARK continues to evolve to where it is now where we talk about HARK agents which is essentially our enterprise platform for all of AI agentic work. when we had issued that press release launching it coming up on a year now you know we had over 200 agents in this library that were ready to get deployed into into an enterprise you know depending on what they were looking for and one of the benefits a side comment I'll make of having Hitachi as such a group company is embracing this customer zero approach meaning we can practice it internally with our other sister companies >> and be able to get internal feedback as we're going to market with it that's a huge benefit benefit that not not every service provider has because some of them are purely gsis and that's all they do. You know, for us having energy and rail and and manufacturing work and so on. That's been a big benefit. But what the agentic hardcark agents allow you to do is really build and orchestrate and govern do all the agentic work across the across the business hitting all the processes and so on. And then when you bring in R202 which is a framework really around security compliance governance and observability that's where it brings in the full 360 approach to having our value ad which we see in the market of HARC agents. So it evolved from really a engineering attitude against development to where it is now by hosting agents and making sure that everything is underpinned by some of the key areas. That was a lot right. Uh, so we're always working to make sure that when we tell these stories in the market, we're being as efficient and quick as we can to show the business. But in some cases, you know, you have to be you have to explain these things in detail because everyone is still getting used to it, you know, and I I would assume from your end as well, you know, you're walking into enterprises, some have an understanding of AI at a high school level and some have it at a PhD level, right? You're always you always have to consistently adjust. So, so for us, HARK is a is a real backbone to everything that we do AI related. >> Okay. And then you you spoke about physical AI as well. Tell me tell me more about are they pre-built agents around the physical world you're seeing? By the way, I don't know if you I heard about this today. There's a company in New York called Shiftapp.nyc. >> Okay. What they've started doing, Patrick, is I mean I I wish I got paid to to show these guys out, but they clean their homes for free, but it comes with they all the all the cleaners basically wear cameras. So, they're basically collecting data about your home like how Google Maps went and collected data about streets in the world and then they're going to use that and give it to AI labs who want to train on that data, right? Uh and actually we saw another company which is similar to this in India that is already doing like cleaning services but they're adding this layer of like hey this is like a verified service which is like a premium service over and above. So I this is this is crazy right because they can license the data they can essentially make monetize this. How are you looking at physically or how's Hitachi kind of looking at physical as well? You know, I'm debating if I'd want somebody to clean my house for free and give them that that architectural data. Interesting. >> If you're 24, you you might consider it, I think, right? Like, because you're like, "Yeah, I maybe know I want somebody to tune in my house. I don't want to be spending $200 $300 on it." But >> yeah. Well, from from our perspective, we are we're definitely not thinking about physical AI in the context of of of cleaning homes and collecting data. You know, for us having the IT in the OT background, meaning you know, as you mentioned, Hitachi's 110 year history, 110 years of that has been in the OT side and, you know, about 60 years of that has been on the IT side. And again because of the the internal relationship between all these different companies there's been a huge resurgence of industry 5.0 Um this is an area where that historically has been sort of growth in project work meaning you need a sensor here or you need a you know connector here and it's been very deploy it and leave and what we're seeing now is the need to for example one best use case is make my factory smarter right it's not just about deploying a robotic arm and then sort of leaving now now it's creating managed service capacity around that because these things are becoming so critical to operations now. our you know our IT practice you know has been shooting for the moon in terms of the growth of everything it's doing and so when we think about physical AI you know it's all about how do we take the AI that we always often think of and apply it to the machines and the industrial equipment and the vehicles and and most importantly the mission critical infrastructure >> right >> which allows you know all of the digital and the physical assets to come together so it's moving from chat bots into how do we make the field technicians make better decisions? How do we make the command centers that are monitoring the operations, you know, improve their resilience around issues? But I let me let me tell you a very important story about this. Even though again AI is continuing to make a push into this into the ITOT space. We were at one of Hitachi Rails command centers in Lisbon. And what this is responsible for is signaling coordination and train movement both for safety reasons of the riders and also operational reasons. Now when you hear a story like that before you walk into a command center, you're thinking, okay, this is going to be very computed driven, right? There's going to be large mainframes and, you know, and LARs and and cameras and all that stuff monitoring all this, but there were human beings sitting at desks >> because this type of mission critical work cannot be perfected by only using the AI or the physical AI components. So when I look at my colleagues sitting there, you know, on their 8 to 10 hour shifts, making sure that everyone who steps on a train that is monitored through Hitachi rail is safe and every train is running operationally efficiently. This really said to the team that we were with on the visit, you know, the human piece plays such a critical element in all of the mission critical work we do. So that was that was really first of all it made me really proud that these guys are doing you know the men and women in there are doing this awesome work and I can call them colleagues because of the one Hitachi approach that we've always taken but it was a clear example of how physical AI uh is up and running but there's still that oversight from the human the human piece and I I could not ask for a better use case having just walked into an office and seeing all these men and women at work making sure everyone is safe and operating the way it Now for us, physical AI is big. It is going to continue to be big. So we see that as a big differentiation for us in the market. >> Solid. Cool. I'll maybe ask you one last question, Patrick. I'll let you go and start your morning or rather I'm sure you're halfway through your morning as well, but tell me this. I mean, I can tell you internally we were big users of charge GPD in 2025. We also launched a product which was called Lizo GPD. And then we realized that all the anthropic really picked up. Claude has almost become second nature for us internally are you seeing like a tussle in at Hitachi is are folks using both open AI and anthropic what's what's happening within Hitachi in that sense >> so we we do have at a Hitachi limited level which represents all of the companies with both anthropic and open AAI and we are uh you know we're looking to be able to maximize the benefits that both of the platforms and the systems offer obviously you You you have the personal use of AI and some people I think have used AI tools for a long time now and I mean long meaning 18 24 months >> right >> where they spend so much time with them that they feel like the AI knows them right and so if you if you're asking them for support on hey can you help me write a LinkedIn post about my recent experience there sort of there's sort of what's the word it's there sort of discipline that I can only ask the AI that knows or else I'm going to have to do several iterations if I suddenly switch tools >> out. >> Then you have on the enterprise side, right? There's certain and this is for everybody, not just Hitachi, right? There's certain licenses you have. There's security measures what you can use for internal versus external. So I I really I really think the balance has to happen from the users of this, right? There's always going to be a prescriptive element of what's a preferred tool to use at each company. There's going to be various levels of partnership, various levels of R&D engagement, and then there's going to be stuff that you as a personal, you know, you personally use on your own and you just feel comfortable with because it's sort of grown on you. I don't know the answer to how that gets solved. >> Sure. >> Right. It's there's a there's a there's a relationship that has been built with users with their Gen AI or their you know Gemini or Copilot or whatever you know what Claude whatever they're using that I think people are a bit afraid to leave because it knows so much >> about about you and it'll give you the best answer by virtue of that. So it it's I think it's it's definitely a big discussion that's going to keep happening. It's a big debate about, you know, do enterprises allow you to let use what you want to use? Do they, you know, do they take a survey internally of what you want to use? Do they want to drive it from the top down depending on where you are? But, but from our end, you know, we're using we're engaged with various organizations that makes sense as to what businesses is needing. So, I think it's an, you know, we could spend I think your next podcast needs to be about, you know, how how do we manage this business B2B and BTOC relationship between who's using what AI platform? M I'll I'll make sure to do an episode on that. But cool, Patrick, this was great. Thank you so much for your time today. Uh would love to obviously know how you're using AI in your life and obviously at Hitachi as well as we move forward. But but thanks for today. >> Absolutely. It's great to be here.

Assistir ao vídeo original no YouTube

Assuntos: Agentes de IA, ANI. Publicado no YouTube em 2026-09-12.

Seleção automática por assunto e idioma. O vídeo pertence ao canal de origem.