 Hi everybody, welcome back to Las Vegas. My name is Dave Vellante. We're here at the Mandalay Bay at Clickworld 2023. Click was a long time supporter of theCUBE, back in the big data days. But we saw them at Hadoop World. They were at our big data events, big data NYC. So we're really excited to be here. We're at the peak of day one. Molly Burns is here. Molly, we met at AWS re-invent last year. She's the VP of sales and North America partners at Click. She's my co-host, okay? We're going to knock it down. We're really excited to have Annette Yonker here. She's the head of data analytics at Harman and Dr. Jay Ganesh, she's the chief product officer at Harman. Harman Kardon, my Benz is in the shop. I can't wait to get it back, the cranks of tunes. Guys, thanks for coming to theCUBE. Molly, set up this story here because we've been hearing about Harman and the keynotes and this is an awesome case study. But take us through sort of the top level. Yeah, absolutely. We're excited to be here today with Harman as our customer, which you're going to hear a little bit about today, but also in the spirit of announcing our official partnership with Harman DTS also this week. So what we're going to hear a little bit about today from the rest of the team is about Click and the impact that we're making at Harman and about the future of our partnership with Harman DTS and some of the innovations that brings to customers and the Click ecosystem. Annette, tell us a little bit more about how all this started. Great, thanks, Molly. Pleased to be here. The journey really started in 2014 for us with Click when we did our first ClickView purchase. The business used it and we actually had multiple pockets of ClickView implemented across the organization. Fast forward, we had a lot of success with ClickView and actually decided that that would become our enterprise or global analytics enterprise platform. So we selected it for the global platform. And then in 2019, we brought analytics together into its own group for Harman Analytics. At that point in time, we actually converted everything from ClickView to ClickSense. So that's where we are today. In terms of the way that we really handle analytics within Harman, it's a little bit unique in that we work across multiple areas of the business. So you'd mentioned your BMW being in the shop. My Benz. Oh, your Benz. My BMW, I might go back to BMW, but yeah. My Benz has an awesome Harman car stereo. It's phenomenal. Great, so actually one of our first use cases was around warranty. Right, so when it got checked out, there's a lot of records that actually get put together on what happened during the service incident and all of that data actually comes to us. And we look at it in aggregate and understand what we can do better within our product. So it's all within the quality sphere. So as you know, Harman Kardon or Harman International is a quality sound organization. Yeah, amazing. So that's one of the applications that we use Click for. So Jay, your role as Chief Product Officer, I guess you are a technology company, of course every company is, but talk a little bit about your role and how you use Click in your products. Sure. So within Harman Group, I am heading the product for Harman DTS, which is our services and consulting division. So we work with some of the largest global customers. We help them in their digital transformation journeys. A lot of our work which we do for our customers is in the area of data, analytics, AI, machine learning and cloud computing. And now what we learn from what Annette explained in terms of our deep expertise in Click is what we are bringing to our customers through this Click partnership is about pushing the boundaries, use some of the great technology which Click brings to the table and combine that with a lot of AI ML, particularly some of the emerging trends around generative AI, building solutions to help customers make better sense out of the data. And most customers are large data shops. Many of the times this data is not integrated with the workflows, they're not able to make better decisions. So what we are doing is, we are building AI based approaches for making better sense out of the data. For example, we're using generative AI and you'll see in the demo session tomorrow, is how can an analyst run completely unstructured queries on a Click dashboard and get very good answers, not just answers, but also prompts in terms of what other analysis he or she needs to run, again in the form of prompts so that they get the results in the three or four questions rather than going through a long process. So that's what we're trying to do as one of the early projects around Click. We are also doing some interesting stuff on the ML ops, automating the entire machine learning workflows, how can you shorten the time from concept to a production model ready for deployment? It was interesting to hear Ford this morning talk about how most of their dashboards are built by business people, right? And so it sounds like you're taking it to the next level where I can use human language. You know, Amazon turned the data center into an API, it's like ChatGPT is going to turn technology into a natural language interface. Is that really where we're headed here? So that is one of the advantages because we're not forcing everyone to shift towards the chat window, right? If there are people comfortable with the traditional approach, all the better. But if you're looking for something beyond what you know as a set approach for getting your answers, we give the option of having a completely unstructured query which becomes much more interesting because it's not just the answer you get, you also get further recommendations as what you need to do next. You know, Molly, not only sales in your title, but you got partners in your title. And so I'm sure that you don't wake up in the morning looking for a transaction, right? You want partnership. So maybe you could tell the audience a little bit about this partnership and you could dig into that a little bit. Yeah, absolutely. We definitely are looking for depth in our partner ecosystem. And one of the things that really attracted us to Harman DTS is the innovation, the artificial intelligence center of excellence that they have. And they're also able to assess and arm and enable customers throughout the journey, whether it be system dependencies, security concerns, new ideas, how that may be applicable within click, outside of click, or in the generative AI space that we're talking about today. So for us, it's very important to have strategic partners that can have bigger conversations and also be a trusted voice to the customers and to the community. And that's one of the reasons we're so excited about the partnership kicking off this week with Harman DTS. So Harman International has a lot of data all over the place, right? I mean, the consumers, I know, I turn it up loud. Maybe others don't. Maybe the type of music, I don't know. But so Annette, what's the general trend around, there's a big discussion around, shove everything into a single data store or distributed data, data mesh, data fabric. How are you guys dealing with that discussion? There's a lot of organizational conversations around who owns the data, right? Who should be responsible for the data lifecycle, the data quality, the data governance. We'll get into some of that in GPT as well. But what's the prevailing sentiment right now inside the organization? Yeah, so at the end of the day, what it comes down to is speed to insight, right? So that is really the first iteration of any application is how quickly can I actually get the insights to the business. With Click, we actually had an interesting place where you can build an application and then retrofit it to wherever the data source is or should be. So we're not blocking our speed to insight by saying, hey, everything has to be within the data lake. What we're doing is saying, hey, let's build the application, get the insights, and then productionize those insights in some cases, which really takes away a lot of those initial questions and initial feel of fears. Now, that also leads us to our data strategy. So for us, data mesh seems to be the most applicable. Because we're focusing on use cases again, right? We're not trying to boil the ocean or create this very large data warehouse where everything lives. We're only bringing the data that we're actually going to use into our data mart or data warehouse, which lives within Snowflake today. So it's not limiting the business user while ensuring that these applications are supported within production. So one of the other reasons to bring data in from its source to a platform is so that it becomes more reliable and resilient, right? And you're not impacting the source system's operations. And logically, that data source can live anywhere. Live in the cloud, well, it's got to be in the cloud, I get it, but it could be in Azure, it could be Google, it could be GCP. Really can't be on-prem. It kind of can, but there's some, I know. I know, I know, right? So that's kind of cool. I like the data mesh approach. You know, it's funny, because Snowflake, to get the advantage of Snowflake, it's everything's got to be in Snowflake, and it's awesome, right? You get the governance and the security and the data sharing, but they have a data mesh-like philosophy, don't they? It's kind of interesting to see. And so I'm curious as to you're saying, you're kind of leading into the data mesh and Snowflake, because they have this kind of distributed, but logically centralized architecture. So it's like the best of both worlds. It is, right? So everything doesn't have to live in Snowflake, right? That's not a decision that you have to make immediately, right? However, yes, they do kind of go towards the data mesh in that what we're creating is data as a service. So this is really kind of a step towards. And then being able to aggregate data at the same level of aggregation and making it applicable, you know, being able to use data sources that are at the same level of granularity from different places, right? So how does click fit into that sort of value flow? Like, you're using it for data integration, data quality, increasingly, I guess, some parts of governance, is that right? Yeah, so we're actually using it very much and very effectively as the front end for all of our business users to the data, right? So, and one of the things that we found is there's such a demand for data and analytics that we really have to, just like your Ford use case, the business users should be enabled and able to create their own applications. Because we cannot really keep up with the demand, right? So we have an enablement program that walks people through from how to use click to how to extend click, which is really unique to click, right, being able to add charts on, publish them out into the central hub or really create their applications from start to finish. And that is where we feel the biggest value is, right? And then that enterprise architecture with secure data sources that allows us to create one application and distribute it to multiple different people or different groups through security, right? So, yeah. So, Jay, I saw something on your LinkedIn and it resonated with me because I was interviewing Nir Zook, who was the CTO and founder of Palo Alto Networks. He said, within five years, every job is going to be powered by AI. I think your post on LinkedIn said 80%. Okay, maybe not every, but who knows, right? So, how should we think about foundation models in terms of, or how are you applying them specifically? And what does that mean for individuals and how are they receiving it? Yeah, so I think that was a study published by OpenAI and so the way I look at it is that a human plus AI is a better decision maker than just a human or just the AI, right? So, this is where we advocate something called a human in the loop approach, right? It's a simple concept of not letting the AI run wild and take decisions which can be questionable. Neither let the human function without the efficiencies, without tapping into the efficiencies of AI, right? So, the simple concept is every single job in the future is going to be augmented by AI, right? So, there was this old study published, old study as in four or five years back, there was a study published by a few faculty members from University of Oxford where they looked at the future of work and they shortlisted every single job out there and they gave a probability of getting disrupted by AI. By disruption is a simple word to say, getting eliminated, right? So, that's a very interesting study to look at. So, and the interesting thing is a lot of the jobs they had listed those days five or six years back as difficult to be disrupted today with all the generative AI tech, they're getting disrupted. So, the way forward is to embrace AI at the same time follow a human in the loop approach where there's a human checking a decision, critical decisions at least, the common decision, the level one, level zeroes can be automated have a human in the loop in for your decisioning. So, you know, we've all used chat GPT, right? And you put in a prompt, okay, not quite do another one or you might have the manual multi-step prompts. We already want to automate this. Well, lo and behold, here comes auto GPT, there's sort of a multi-sequencing and it happens so fast, right? So now I could say, okay, I'm going to New York for the weekend, my family, we got six of us, do an itinerary, we need a hotel. Tell me the best way to get there, what's the hotel? What about a kid-friendly activity on and on and on? And it will create that multi-step, it'll feed itself the prompts. I mean, it's just mind-blowing, it is. So, someone like you, Dr. Jay, you see that, you got to start thinking about, okay, how can I apply this? So, how can you apply that? A simple analogy is that if you want to call GPT as a pearl in a necklace, you can call auto GPT as a complete necklace. It's like steroids on steroids. Exactly. It's unbelievable. And where this is going to fit in from an enterprise scenario is that it is going to help automate workflows. GPT is, chat GPT or any of the GPT mechanisms are good at automating or giving you great results for pointed problems. What auto GPT can let you do is that it can connect these pointed problems and complete your workflow. It can run these individual instances from multiple systems, get the output from each system and give you the most optimal output. It can automate complete workflows and that is the beauty of the whole thing. I get these inbounds all the time, Molly, from people saying, hey, do you need more leads? I'm like, we're drowning in leads. No, actually, good leads, okay, fine. But, you know, you have organizations, they have BDRs that they're out, hey, you know, they downloaded the Gartner White Paper, and they go, AI could do, it could identify, you know, the right target can follow up with and suggest, actually, the right document to send them. Yeah, determine the next best action. We're not that far away from that. We're kind of here, right, today, so. And I think the evolution's just beginning as we're starting to unlock different versions around generative AI and chat GPT. And when you think about automating workflows and how that also can fit in with CLIC, there's a lot of power to be harnessed there as well around application automation, around AI, operationalizing machine learning, and making ML ops part of that journey. So I really think we're on the precipice of an evolution and kind of a new world, so to speak. You know, there's a lot of talk, too, about how social platforms, you know, discussion of, okay, are they responsible for the content that's published, because it's user-generated content, and there's laws that protect them. Well, this is, this foundation model is, it's all user-generated. It's, you know, the prompts are coming from users, so every company is a software company, every company's a technology company, so it brings about some really interesting and challenging ethical considerations, privacy, issues of bias, not to go negative, but those are things that we all now have to think about. Right, so how have you thought about that? So that is something which is the top of the mind for many of these enterprises, particularly for enterprises where the content is their core business model. And that's when the challenges happen, because they have paid for this, and let's say that there's an image company which is making a living out of selling images, certainly you have an engine which comes in and learns from that and creates better images. Who's IP is that? Who's IP is that? So there are court cases going on, there are companies who have sued, for example, Stable Diffusion and a few others, saying that you learned, your model's learned from my data, right? So you can't, that's my copyright. But, you know, I think this, you know, in the next one year, we would see some resolution in terms of what the legal fraternity has to say, but in my opinion the solution lies somewhere in between. You know, you learn from this data, you're generating revenues, you have a revenue sharing mechanism or something like that, because otherwise it is very difficult for the data owner to claim complete ownership because, you know, the next version would make it even more difficult for them to identify where they copied it from. But that's going to be a tricky one. Yeah, where do you stand on the slowdown or not? So the market slowdown as such? Yeah, you know how, I guess Musk came out with, hey, I always thought, when I first started, he said, oh, he's trying to slow down the market so he could start his own company, of course he did. But it's not, it's a hard question, right? Because there's competitiveness for U.S. companies, there's, well, there's the ethical considerations, so. Yeah, see, first of all, I fail to understand what does a slowdown for six months mean for this? Why six months? Why not? What's it do for you? Why not one month? So it's sort of a weird scenario. And to keep in mind the fact that if companies in one geographies slowdown, it doesn't mean everyone else is slowing down, right? So people are not slowing down across the world. So if you're slowing down, you're part of the industry, you're losing your competitiveness or your country's competitiveness. I'm not in favor of slowdown. Let the market take its course. It's hard for public policy to adjudicate this. All right, Annette, we'll give you the last word, sort of the future of data in Harman and click the partnership where you guys want to take this partnership and your vision. Yeah, so I think, you know, for us, the future of analytics is bright. There's a lot of more data sources, there's a lot more use cases. And we're an innovative company, right? So we do things differently and interesting and we come up with solutions. In terms of the partnership, it's taking the solutions that we've developed and then bringing them to our customers. So we are our best, our reference customer, really. Great. I wish we had more time. Jay, I'd ask you if Quantum's going to help me mine Bitcoin. I saw you posted about that too. Guys, thanks so much for coming to theCUBE. It was a great conversation, really appreciate it. It was. You too. Thank you. Thank you for having us. Right there, Lisa Martin and I will be right back. We're here live at Clickworld 2023, day one. Right back, we're watching theCUBE.