theCUBE + NYSE Wired: CMO Leaders | Kazuki Ohta, Treasure Data

Clips
More from theCUBE + NYSE Wired: CMO Leaders

Kazuki Ohta

Co-founder & CEO

Treasure Data

play_circle_outline Unlocking Enterprise Success: The Evolution of Treasure Data From Hadoop to Customer Data Platform

play_circle_outline AI transforming marketing, sales, and support processes

play_circle_outline Potential for automating marketing stack operations with AI

play_circle_outline Real-time data quality diagnosis using AI agents

play_circle_outline Democratization of entrepreneurship and innovation with AI advancements

Info
Transcript

Kazuki Ohta, Treasure Data

Kazuki Ohta

Co-founder & CEO Treasure Data

explore Keep Exploring

What prompted your company's decision to rebrand as a customer data platform (CDP) in 2017? add

What is the importance of combining foundational models with customer data in order to create a more effective AI agent for customer support in various industries? add

What is the potential opportunity for disrupting the use of AI in managing Customer Data Platforms (CDP)? add

What is the possibility of improving data quality by having the CMO inform AI agent teammates to diagnose data quality daily and suggest edits before making fixes? add

What impact does the implementation of AI in mundane tasks have on human capital and entrepreneurship? add

bolt Powered by CUBE AI

Kazuki Ohta, Treasure Data

search

>> Welcome back everyone to theCUBE's coverage here of the CMO leaders. I'm John Furrier, host of theCUBE in Palo Alto Studios, part of theCUBE and the new NYSE Wired community. Founder, Brian Baumann and his team really creating this new open community around people coming together and celebrating and contributing content, sharing knowledge, and of course feeding the AI systems. And that's really what it's all about. Kazuki Ohta is here, co-founder and CEO of Treasure Data, friend of theCUBE, many times on theCUBE. You team, going back to 2011, 2012, and you guys been on in 2015? It feels like It is a decade ago.

>> Wow. It's a decade. Oh, yeah. John, thanks for having me today. I'm really excited. We're coming back to the show.

>> Congratulations on all your success. And again, you guys have been on the journey. You've seen the movie, you remember the big data craze with Hadoop. Got everyone's attention. Okay, data's the new oil. Remember back in the day, okay, blah, blah, blah.

>> Yup.

>> Now fast-forward, okay, Hadoop had issues, but data warehouse never went away. In comes the cloud, in comes all those innovations. So people now are front and center staring at the value proposition that over a decade later, that's now mainstream. And CMOs represent the biggest opportunity, one of the biggest opportunities for leveraging data, but transforming their, I guess, vertical specialization of talking to customers, capturing leads, engaging, using data, not as a looking in the rear-view mirror dashboard, but actually instrumenting their business. That's a huge part of it, and you're in the middle of it.

>> Yes. Yeah. So when we were first interviewed, we were Hadoop on the cloud, but since then we actually evolved a lot. So around 2016 to '17, we just realized that majority of our customer is using our Hadoop on the cloud service to analyze consumer data. And then we also saw customer data is one of the most important precious asset of the data for any of the enterprise. So around '17, we decided to call ourselves as what we call CDP, customer data platform, which was just a little bit of maybe seven to eight vendors, little category. But now that we have maybe 250 vendors in this category, but again, we were named as a leader by Forester, IDC and Gartner. They did the magic cold run for the first time ever last year. We're the only independent vendor who are leader in those three category analysts, and we're really, really proud of where we're right now.

>> If you remember those conversations and where you kind of adjusted your business model. And the customer focus really was also talking about personalization. If you look at AI right now, its hyper personalization seems to be the value. And then the data models at this customer data platform layer is harmonized a little bit differently than say the old school data warehouses in the cloud, which were very mechanical in the sense of they had to be programmed. But now with generative AI coming down super fast, the application developers want to build apps with data, but also managed how they're used. So you have two views of this. Can you talk about those two levers? Deploying applications with customer data being used securely and effectively, and the other lever of how it's being used.

>> Right. Yeah, it makes sense. So I think I became a CEO three years ago. I was a CTO, and when I became CEO, I created this vision called connected customer experiences. Customer data, this is the CMO conference by, but when I talk with a lot of enterprise, yes, marketing is the easiest way to get the ROI, meaning turning data to the dollar. But at the same time, we've seen a lot of auto companies using customer data to empower seller at the dealership. So by helping those decisions and how they sell and recommend the product, we increase the dealership conversion rate from 17% to 23%. And John, that's a lot of money, given it's a car. And then we're also using CDP, it's customer seeing CDP for tackling with compliance, GDPR, CCPR, CCPA and improving customer support and on even improving logistics, right? Trying not to overcook some chickens based on the customer data and demand.

>> I love it.

>> So the opportunity of using customer data, I'm seeing a lot start from marketing, but I just don't want to limit it because every business has the customer and then the only thing you know about the customer, it was face-to-face, but then customer journey becoming now 60 to 70% in digital, data is the way to know your customer.

>> I love that digital aspect of it because the crown jewels or the main value of the customer data is the customer themselves. Okay, yeah. There's low-hanging fruit use cases on marketing, content, call center. Okay, check. Okay, that's clever. Where it goes into is the value, and again, enterprises right now is where all the money is. Everyone sees that, but people are stalled right now, the new entrants, but you guys have been there for a while. Customer data is complex, it's invaluable, it's super valuable. And also the insights into it horizontally, but yet in the areas they need to be focused on with AI manages the domain expertise. So talk about that dynamic between domain expertise, the vertical stripe, and then you got the horizontal harmonization that's needed. Why is that important and why can't the traditional governance warehouses do this?

>> Right. I think of CDP as more like brain or memory for the AI agent. So if you look at top five gen AI use cases or big opportunity, one is security. The second is coding. Those use cases, you can use foundational models because you want a vast amount of knowledge. The third use case is marketing. Fourth use case, sales. Fifth use case is support. So if you look at those customer-facing use cases, foundational model doesn't know about your business, doesn't know about your customer. So it really is about how do you combine foundation models with your own data? So for me as a consumer, when I go to let's say airline company calling, I will try to avoid AI conversation as much as possible. You know why? They don't understand me. That's probably the biggest gap we can bring in for the society, which is how do you connect all these AI agents getting built for specific multiple purposes, but without them knowing customers? That's a no go, right? So at Treasure Data, we're investing a lot on AI, LLM, how foundational model and then customer data working together to better build a better, clever agent. And when you think about that horizontally harmonized platform, the importance of building agent on top of CDP is one, obviously data has to be unified, but then it's all about, let's say permission. You have to treat this AI agent almost as if it's an employee. So how do you control the access to that PII? Permission, role-based access control. So there will be a lot of AI agent, this is a copywriting agent, this is an email marketer agent, this is customer support agent. Each agent has its own role. How do you control PII permission? And also a lot of company right now is talking about AI to AI collaboration. But what I'm thinking about what's important as of today is AI to human collaboration. What if you can actually AI as almost like a teammate where you know what? I'm the CEO, I want to run this campaign analysis for this specific campaign I run. Create a task on Monday.com or Asana, assign that to your AI teammate, and then boom, when you get up, you get some report. And then at that point, you don't even realize it's actually done by human or AI. It's just a workflow between the human and AI. So I guess a lot of key, what I'm hearing from the enterprise is how AI, which is really strong as math and coding and research, copywriting and analytics, more importantly into the daily workflow, that's what we're trying to solve right now.

>> Kazuki, talk about that, you just mentioned something super important. You're collecting, you're unifying, you're analyzing customer data across various channels. Remember back in the day, the omnichannel? Now real time is super valuable. So delegation, trust. Trust is huge. So when you have these agents working on your behalf as a marketer, that's a big piece. Talk about the trust aspect.

>> So first of all, real time is a really good point because we're hearing a lot of CMOs and customers saying, any products, the buying journey is becoming shorter and shorter. You see TikTok ad and purchase it, that's becoming the behavior. That real time aspect is important. But at the same time, looking at all of the historical data that on the batch side is also important because to build the intelligence and machine learning models, you need to look back everything. So at Treasure Data, we actually unified both batch and real time infrastructure together. So you can actually build real time actions on top of the batch learnings. So that's important. And the trust piece is also interesting. There will be a lot of future about AI making billions of decisions every single day, right? What interactions, what answers to do, what personalization. Maybe you can build landing page generated for every single page view. But at the same time, as of today, I think this AI to human collaboration, the previous point about how human can just do, let's say approvals and reviews, and that's more important right now. So when I hear about requests coming from as of today, it's all about how human and AI collaborating together, but then over time, how we as a human can delegate decision to AI. That's much more becoming more important.

>> CMOs that we talked to here, there's ones that lean towards, "I do a lot of the stuff around demand gen, lead gen," they're like marketing machine. They have the MarTech stacks. And then on the other side, you've got the CMOs who had to deal with this existing stuff and the large enterprises, but they're leaning more towards the CEO, the CFO, the CISO saying, "We want to transform business transformation for customer experiences." You mentioned your experience with chatbots. Nobody wants to talk to a chatbot that's not tapping another database they should be tapping to get better experience. Okay, great. Those CMOs that want to innovate there see a huge gap between the complexity of pulling that off databases, harmonization layers, semantic web, what do you want to call it? Whatever that layer is with the available skillset. So most CMOs don't have developers on, they don't have platform engineers, they don't have data engineers. So you have this SRE kind of role where it's like, "Hey, CMOs need a team to fill the gap between complexity and the skills required." What's your take on that? What's your reaction?

>> Yeah, so right now, let's say when you are trying to implement and operate and manage Treasure Data, let's say from data perspective, you pay $1 for Treasure Data license, but then maybe you are paying $4, $5, $10 to operate and implement, manage it. I think the biggest opportunity where we can disrupt with this power of AI is what if we actually build AI to manage CDP, which finds all the data from disparate data sources, configure this identification, create segments, suggest segments automatically optimize A-B testing every single day, and then actually analyzing it and then optimizing every day. So the future I'm looking for is, yes, we're CDP vendor right now, but with the power of AI agent, how can we actually make this entire process of implementing, unifying and use of the customer data maybe one third of the cost with the AI agent? Or ideally, you have this one big red button, turn data into the dollar and then the AI will take care of it. That's kind of the product direction we're actually building towards right now.

>> I mean, basically the CMO's opportunities to have a team of people behind them that superpowers them into having real time, highly agile execution, built in agents doing A-B testing. So I see that vision very clearly, and I think that's why people are leaning hard towards that use case, and they're excited about agents. But you got to do the prep work. In the old days, it was data prep, like wrangling data. So now that's going to happen faster. You guys are doing a great job there. The question for you is what you're seeing now, given what you've seen in the past 15 years or so, you're starting to see signs of real examples where breakthroughs can happen that were not possible. I've been using the example of enterprise search. Very hard problem, if you know search and how data is structured, it's a lot on different databases, stovepipes, unstructured, structured schema. I mean, by the time you end up doing it, if you end up even doing, it's mutated from -

>> That's a nice way to say it, yes....

>> I'm obsolete. But now you're seeing people nail the search category by using RAG. Okay, just abstract the way the complexity essentially keeping everything hidden. So the question for you is what use cases do you see that are breakthrough opportunities for marketing infrastructure, marketing stacks, marketing executives to solve now that they couldn't do before, that was a heavy lift? What search was in enterprise search. Hard before AI, now we're seeing easily good retrieval, great check. What on the marketing side do you see that a CMO can go back to and say, "With Treasure Data and with AI, I can do that breakthrough. It would've been a huge transformation project. Now it's gettable."

>> Yeah. I think for any technology, it's just a mean to the end. It's like how? So I always believe that when technology is used by the business people, that drives faster business result in cycle. So what I'm most excited about is maybe this time, CMO can actually tell AI agent teammates on having data quality diagnosed every single day. We actually have a prototype agent collect the data, unify it, and every day, they actually diagnose your data quality about your customer data and then trying to fix it. But before you fix, you actually suggest, "Okay, here are the edit," or suggest we want to make for this IDU algorithm, so I want to get some approval and review. So they're asking it right now, but we always can make it more automated.

>> Take your headaches away.

>> Yeah. To join your point, I think this time the hard part is about having really high quality data. AI is all about data. How can I feed this high quality data? And that was really messy and then a lot of technical work. But with this wave of technology at announcement right now, I see this real opportunity, you of what? Maybe 70 to 80% this could be automated at some point.

>> I have to ask you, now that you're into it, into your journey, you got a lot of history, what are you most excited about right now in terms of the technology theater around the opportunity, around having fun with the tech? And then the customer side, the value capabilities you're bringing to the table for the customer are highly accelerated. Do you see the needle moving faster now than ever before? And what areas are you excited about the most?

>> Yeah. When I started programming, what excited me the most was I built this little shooting game on my mobile app and then open it up, put it into app store, and half a million people download it. It's about how little this random high school student can make an impact to the world. So when you think about right now, there are 8 billion people on the Earth, 5 billion people internet connected. We're actually managing almost 3 billion people's data. That's about 60% of entire internet connected population, so we have a lot of social responsibilities. And when I take a look at how AI can disrupt right now, let's say if we can actually optimize 1% of sales and marketing labor market, that's about $100 billion efficiency increase. And that's, you can have three more sales force gets born in the software industry. I already feel like ChatGPT already optimize 1% of my work every single day. It's just some of the customer use case, again, they don't connect with customer data. But if I can solve that, I can optimize for the entire society, maybe 5%, 10%, all of the sales and marketing economy. That's just a huge contribution to the entire society.

>> That's a great point.

>> Yeah. I'm also coming from Japan where it's actually Asian country. When I talk with a lot of executive there, maybe in the next 20 years, 20% of workforce will be gone because we're Asian. So the only way the country can grow or the economy can grow, is how do you leverage this technology like AI so that you grow the business, but with maybe 80% less people, that's just naturally happening with this age distribution?

>> And with theCUBE global initiative that we're executing on and with the NYSE Wired, it's a global network, so it's an idea network too as well. So when AI starts implementing some of these kind of mundane tasks, toil work, that's hard, whether it's QA or cleaning the data, semi-autonomous to autonomous. Whatever phase it is, the human capital becomes super valuable and entrepreneurship becomes democratized because from the dorm room to the boardroom, anyone could innovate a category. Right?

>> That's exactly right.

>> And so the data knowledge of the domain expertise becomes very clever opportunity for anyone to move the -

>> So I get the execution, right? Yeah. It's fastest than any age, so I'm actually really glad we were born in this age.

>> I loved your app store story because that is what we see coming. I'm curious, what was the reaction when you saw all those downloads? You were like, "Wow, what's going on? This is great."

>> Yeah. I was like, okay. Random high school student built that game and half a million. And then when I actually came back from school, the next lady I was sitting was playing my game. So I was just blown away.

>> Were you on a train? Were you on a plane? You like, "Hey, that's my game"?

>> Yeah, I didn't say that, but at least after that, I was into this technology world for almost 25 years.

>> Well, we're a big fan of your work, and hey, congratulations on the continued success. Again, we've been following your journey and your team.

>> Thank you.

>> Again, big data world from 15 years ago, that's when we started theCUBE. We were on the same journey.

>> Yes.

>> We're getting older, so you look great.

>> We can do 100X more from here.

>> Well, I really appreciate taking the time. Thank you for coming and being part of our CMO series. I think the CMO opportunity can be a very strategic, above the line boardroom level conversation because the customer relationship isn't a CRM software package. It's a data platform that will have to behave differently and be ready for these agents.

>> It's becoming an AI problem.

>> Yeah.

>> Right.

>> Kazuki, thank you so much for coming in. I appreciate your time.

>> Thank you, John. I appreciate too.

>> I'm John Furrier here on theCUBE. We are here as part of theCUBE in the NYSE Wired community. Brian Baumann, the entire team is here in Palo Alto. And of course, I'm John Furrier, host of theCUBE. Thanks for watching.