 Machine intelligence is everywhere. AI is being embedded into our everyday lives through applications, process automation, social media, ad tech, and it's permeating virtually every industry and touching everyone. Now a major issue with machine learning and deep learning is trust in the outcome. That is the black box problem. What is that? Well, the black box issue arises when we can see the input and the output of the data but we don't know what happens in the middle. Take a simple example of a picture of a cat or a hot dog for you Silicon Valley fans. The machine analyzes the picture and determines it's a cat but we really don't know exactly how the machine determined that. Why is it a problem? Well, if it's a cat on social media, maybe it isn't so onerous. But what if it's a medical diagnosis facilitated by a machine and what if that diagnosis is wrong? Well, what if the machine is using deep learning to qualify an individual for a home loan and that person applying for the loan gets rejected? Was that decision based on bias? If the technology to produce that result is opaque? Well, you get the point. There are serious implications of not understanding how decisions are made with AI. So we're going to dig into the issue and the topic of how to make AI explainable and operationalize AI. And with me are two guests today, Shay Subhiki who's the co-founder and COO of Cognitive Scale and longtime friend of theCUBE and newly minted CEO of Cognitive Scale, Bob Picciano. Gents, welcome to theCUBE. Bob, good to see you again. Welcome back on. Good to see you, Dave. Yeah, thanks for having us. Thanks for being here. Shay, let me start with you. Why did you start the company? I think you started the company in 2013. Give us a little history and the why behind Cognitive Scale. Sure, Dave. So look, I spent some time, you know, through multiple startups, but I ended up at IBM, which is where I met Bob. And one of the things that we did was the commercialization of IBM Watson initially. And that led to a thinking about how do you operationalize this? Because look, there are a lot of people thinking about data science, machine learning and isolation, building models, you know, trying to come up with better ways to deliver some kind of a prediction. But if you truly want to operationalize it, you need to think about the scale that enterprises need. So, you know, we were in the early days enamored by ways. I'm still enamored by ways, the application that takes me from point A to point B. And our view was, look, as you go from point A to point B, but if you happen to be, let's say a patient or a financial services customer, imagine if you could have a ways like application giving you all the insights that you needed, telling you at the right moment, you know, what was needed with the right explanation so that it could guide you through the journey. So that was really the sort of the thesis behind Cognitive Scale is how do you supply AI to solve problems like that in regulated industries like healthcare financial services, but do it in a way that it's done at scale where you can get bring the output of the data scientists, application developers, and then those insights that can be powered into those end applications like CRM systems, mobile applications, web applications, applications that consumers like us, whether we're in a healthcare setting or a financial services setting can get the benefit of those insights, but have the appropriate sort of evidence and transparency behind it. So that was the thesis for Cognitive Scale. Got it, thank you for that. Now, Bob, I got to ask you, I knew you couldn't stay in the sidelines, my friend. So what was it that you saw in the marketplace that lured you back in to take on the CEO role? Yeah, so Dave, it is an exciting space and you're right, I couldn't stay on the sidelines. So look, I always felt that enterprise AI had a promise to keep. And I don't think that many enterprises would say, you know, were there experience that yeah, we're getting the value that we wanted out of it, we're getting the scale that we wanted out of it, and we're really satisfied with what it's delivered to us so far. So I felt there was a gap in keeping that promise. And I saw Cognitive Scale as an important company in being able to fill that gap. And the reason that gap exists is that, you know, enterprise AI, unlike AI that relates to one particular conversational service or one particular small narrow domain application is really a team sport. You know, it involves all sorts of roles and all sorts of aspects of a working enterprise that's already scaled with systems of engagement and systems of record. And we show up with the ability to actually help put all of that together. It's a brown field, so to speak, not a green field. And where Shay and Matt and Minoji and the team really focused was on one of the important last mile problems that an enterprise needs to address that aren't necessarily addressed with any one tool that might serve some members of that team because there are a lot of great tools out there in the space of AI or machine learning or deep learning, but they don't necessarily help come together to deliver the outcomes that an enterprise wants. So what are those important aspects? And then also where do we apply AI inside of our platform and our capabilities to kind of take that operationalization to the next level with, you know, very specific insights and to take that journey and make it highly personalized while also making it more transparent and explainable. So what's the ICP, the ideal customer profile? Is it highly regulated industries? Is it developers? Maybe you could parse that a little bit for us. Yeah, so we do focus in healthcare and in financial services. And part of the reason for that is the problem's very difficult for them. You know, you're working in a space where, you know, you have rules and regulations about when and how you need to engage with that client. So the bar for trust is very, very high and everything that we do is around trusted AI, which means, you know, thinking about using the data platforms and the model platforms in a way to create marketplaces where being able to utilize that data is something that's provisioned and permissioned before we go out and do that assembly. So the target customer really is somebody who's driving digital transformation in those regulated industries. It might be a chief digital officer, it might be a chief client officer, customer officer, somebody who's really trying to understand, I have a very fragmented view of my member or of my patient or of my client. And I want to be able to utilize AI to help that client get better outcomes or to make sure that they're not lost in the system by understanding them more holistically, understanding them in a more personalized way, but while always maintaining, you know, that chain of trust. Got it. So can we get into the product like a little bit more about what the product is and maybe Shay, you can give us a census to kind of where you started and the evolution of the portfolio. Look, where we started, Dave, is truly the application of AI, right? So, look, the product and the platform was all being developed, but our biggest sort of view from the start had been, how do you get into the trenches and apply this to solve problems? And as Bob pointed out, one of the areas we picked was healthcare because it is a tough industry. There's a lot of data, but there's a lot of regulation and it's truly where you need the notion of being able to explain your decision at a really granular level because those decisions have some serious consequences. So, you know, we started building a platform out and our core product is called Cortex. It's a software platform on top of which these applications are built. But through our engagements over the last six, seven years, working with customers in healthcare and financial services, some of the largest banks, largest healthcare pair organizations, we have developed the software product to essentially help you scale enterprise AI, but it starts with how do you build these systems? Now, building these systems requires us to provide tooling that can help developers take models, data that exists within the enterprise, bring it together rapidly, assemble this, orchestrate these different components, stand up these systems, deploy these systems, again, in a very complex environment that includes on-prem systems as well as on the cloud, and then be able to turn on APIs that can plug into your end applications. So we had to essentially think of this entire problem end-to-end and that's what Cortex does. But an extremely important part of Cortex that didn't start off initially, we certainly had all the makings of trusted AI, but we found that the industry wasn't quite ready. Over time, we've developed capabilities around explainability, being able to detect bias. So not only are you building these end-to-end systems, assembling them and deploying them, you have as a first class citizen built into this product, the notion of being able to understand bias, being able to detect whether there's the appropriate level of explainability to make a decision, and all of that's embedded within the Cortex platform. So that's what the platform does. And it's now in its sixth generation as we speak. Yeah, so Dave, if you think about the platform, it really has three primary components. One is this application development or assembly platform that fits between existing AI tools and models and data and systems of engagement. And that allows for those AI developers to rapidly visualize and orchestrate those aspects. And in that regard, we're tremendous partners with people like IBM, Microsoft, H2O, people that provide aspects that are helping develop the data platform, the data fabric, things like the data science tools to be able to then feed this platform. And then on the front end, really helping transform those systems of engagement into things that are more personalized with better recommendations in a more targeted space with explainable decisions. So that's one element, that's called Cortex Fabric. There's another component called Cortex Certify. And that capability is largely around the model intelligence, model introspection. It works across things that are of course model driven, but other things that are based on deterministic algorithms as well as rule-based algorithms to provide that explainability of decisions that are made upstream before they get to the black box model. Because organizations are discovering that many times the data has aspects of dimensions to it and biases to it before it gets to the model. So they want to understand that entire chain of decisioning before it gets there. And then there's the notion of some pre-curated applications and blueprints to rapidly deliver outcomes in some key repeating areas like customer experience or like lead generation. Those elements where almost every customer we engage with who's thinking about digital transformation wants to start by providing better client experience. They want to reduce costs. They want to have operational savings while driving up things like NPS and improving the outcomes for the people they're serving. So we have those sets of applications that we've built over time that imagine that being that first use application that starter set that also trains the customer on how to utilize this operational platform. And then they're off to the races building out those next use cases. So what we see is one typical insertion place play that returns value and then they're scaling rapidly. Now I want to cover some secret sauce inside of the platform. Yeah, so if before you do I think I just want to clarify. So the Cortex fabric because that's really where I wanted to go next but the Cortex fabric it seems like that's the way in which you're helping people operationalize inject use familiar tooling. It sounds like am I correct that the Cortex certify is where you're kind of peeling the onion of that complicated whether it's deep learning or neural networks, which is that's where the black box exists. Maybe you could tell us, you know is that where the secret sauce lives? If not, where is it? And if it's all about. It actually is in all places, right? So there's some really important introductions of capabilities because like I mentioned many times these regulated industries have been developed in highly fragmented pillars. Just think about the insurance companies between property casualty and personal lines. Many times they've grown through acquisitions. So they have these systems of record that are that are really delivering the operational aspects of the company's products but the customers are sometimes lost in the scenes. And so they've built master data management capabilities and data warehouse capabilities to try to serve that but they find that when they then go to apply AI across some of those curated data environments it's still not sufficient. So we developed an element of being able to rapidly assemble what we call a profile of one. It's a very, very intimate profile around declared data sources that relate to a key business entity. In most cases it's a person, it's a member, it's a patient, it's a client but it can be a product for some of our clients it's real estate, it's a listing. It can be somebody who's enjoying a theme park. It can be someone who's a shop or in a grocery store. It can be a region. So it's any key business entity. And one of the places where we applied our AI knowledge is by being able to extract key information out of these declared systems and then start to make longitudinal observations about those systems and to learn about them and then line those up with prediction engines that both we supply as well as third parties and the customers themselves supply them. So in this theme of operationalization they're constantly coming up with new innovations or a new model that they might want to interject into that engagement application. Our platform with this profile of one allows them to align that model directly into that profile and get the benefits of what we've already done but then also continue to enhance, differentiate and provide even greater value to that client. IBM is providing aspects of those models that we can plug in and many of our clients are. That's one element. Well, that's interesting. So that profile of one is kind of the instantiation of that secret sauce, but you mentioned like master data management, data warehousing, you know, as well as I do, Bob, we've decades of failures, trying to get a 360 degree view, for example, of the customer. It's just not real time. It's not as current as we would want it to be. The quality is not necessarily there. It's a very asynchronous process. Things have changed. The processing power, you and I have talked about this a lot. We have much more data now. So it's that profile I want. So, but also you mentioned curated apps, customer experience and lead gen, you mentioned those too. And you've also talked about digital transformation. So it sounds like you're supporting, and maybe this is not necessarily the case, but I'm curious as to what's going on here. Maybe supporting more revenue generation in the early phases than say privacy or compliance, or is it actually, do you have use cases for both? It's all of it. And Shay can really talk passionately about some of the things we've helped clients do. Like for instance, Shay, Mani, why don't you talk about the hospital discharge process? Absolutely. So just to make this a bit more real, Dave, when you talk about Profile 1, it's about understanding a patient, as I said earlier, but it's trying to bring this notion of not just the things that you know about the patient, you call that declared information, you can find this information in traditional EMR systems, but imagine bringing in observed information, things that you observed to an interaction of the patient, and then bring in inferences that you can then start drawing on top of that. So to bring this to a live example, imagine at the point of care, knowing when all the conditions are right for the patient to be discharged after a surgery. And oftentimes, as you know, if all the different evidence or the different elements don't come together, you can make some really serious mistakes in terms of patient discharge, bad things can happen. Patient could be readmitted, or even worse, there could be a serious outcome. Now, how do you bring that information at the point of care for the person making a decision by not just looking at the information you know, but also understanding not just the clinical information, but the social, the socioeconomic information, and then making sure that that decision has the appropriate evidence behind it, so that when you do make that decision, you have the appropriate sort of guidance behind it for audit reasons, but also for ensuring that you don't have a bad outcome. So that's the example Bob's talking about where we have applied this in real settings in healthcare, but also in financial services and other industries where you can make these decisions based on the machine telling you with a lot of detail behind it, whether this is the right decision to be made because of the sort of explainability and the evidence that's needed. You know, that's interesting. I'm imagining a use case in my mind where after a patient leaves, so often there's just a complete disconnect with the patient unless that patient has problems and goes back, but that patient might have some problems, but they forget it's too much of a pain in the neck to go back, but the system can now track this and we could get much more accurate information and that could help in future diagnoses and also decision making for a patient in terms of outcomes and probability of success. Question, what do you actually sell? So it's a middleware product. It's a, how do I license it? Talk more about that. It's a software platform. So we sell software and it is deployed in the customer's cloud environment of choice. Of course we support complete hybrid cloud capabilities. We support native cloud deployments on top of Microsoft and Amazon and Google. And we support IBM's hybrid cloud initiative with Red Hat OpenShift as well, which also puts us in a position to both support those public cloud environments as well as the customer's private cloud environments. So we're constructed with Kubernetes in that environment, which helps the customer also realize the value of that operationalization because they can modify those applications and then redeploy them directly into their cloud environment and start to see those spaces. Now I want to cover a couple of the other components of the secret sauce, if I could Dave, to make sure that you got a couple other elements where some real breakthroughs are occurring in these spaces. So Dave, you and I, we're passionate about the semiconductor industry. And we know what is happening with regard to innovation and broadening of people who are now siliconizing their intellectual property. And a lot of that's happening because those companies who have been able to figure out how to manufacture or how to design those semiconductors are operationalizing those platforms with our customers. So you have people like Apple who are able to really break out of the scene and do things by utilizing utilities and macros, their own knowledge about how things need to work. And it's just, it's very similar to what we're talking about doing here for Enterprise AI. They're operationalizing that construction, but none of those companies would actually start creating that the actual devices until they go through simulation and design, correct? Yeah, we are. Well, when you think about most enterprises and how they develop software, they just immediately start to develop the code. And they're going through A, B testing, but they're all writing code, they're developing those assets, they're creating many, many models. Some organizations say 90% of the models they create, they never use some say 50, and they think that's good. But when you think about that in terms of, the capital that's being deployed both on the resources as well as the infrastructure, that's potentially a lot of waste as well. So one of the breakthroughs is the creation of what we call synthetic data and simulations inside of our operational platform. So Cortex Fabric allows someone to actually say, look, this is my data pattern. And because it's sensitive data, it might be PII, we can help them by saying, okay, what is the pattern of that data? And then we can create synthetic data off of that pattern for someone to experiment with how a model might function or how that might work in the application context. And then to run that through a set of simulations, if they want to bring a new model into an application and say, what will the outcomes of this model be before I deploy it into production, we allow them to drive simulations across millions or billions of interactions to understand, was that model going to be effective? Was it going to make a difference for that individual or for this application or for the cost savings goal and outcomes that I'm trying to drive? So just think about what that means in terms of that digital transformation officers having the great idea, being in the C-suite and saying, I want to do this with my business. Oftentimes they have to turn around to the CIO or the chief data officer and say, when can you get me that data? And we all know the answer to that question. They go like this, like the... Yeah, I've got a couple of other things on the plate and I'll get to that as soon as I can. Now we're able to liberate that. Now we're able to say, look, what's the concept that you're trying to develop? Let's create the synthetic data off of that environment. We have a corpus of data that we have collected through various client directions that many times gets that bootstrapped and then drive that through a simulation. So we're able to drive from imagination of what could be the outcome to really getting high confidence that this initiative is going to have a meaningful value for the enterprise. And then that stimulates the right kind of following and the right kind of endorsement throughout really driving that change to the enterprise. And that aspect of the simulations, the ability to plan out what that looks like and develop those synthetic aspects is another important element of the secret sauce inside of Cortex-Fabric. And tying back to the semiconductor innovation, I could do that very cheaply. I think I'm thinking AWS cloud. I could experiment using Graviton or maybe do a little bit of training with some new processors and then containerize it, bring it back to my on-premise state and apply it. So just as you say, a much more agile environment. Yeah, speed efficiency and the ability to validate the hypothesis that started the process. How do you guys think about the TAM, the total available market? Can we have that discussion? How big? Yeah, I mean, if you think about the spend across the healthcare space and financial services, we're talking about hundreds of billions in that in terms of what the enterprise AI opportunity is in just those spaces. And remember, financial services is a broad spectrum. So one of the things that we're actually starting to roll out today, in fact, is a SaaS service that we develop that's based on top of our offerings called Truststar, truststar.ai. And Truststar is a set of personalized insights that get delivered directly to the loan officer inside of an institution who's trying to really match lending to someone who wants to buy a property. And when you think about many of those organizations, they have very, very high demand. They've got a lot of information. They've got a lot of regulation they need to adhere to. But many times they're very analytically challenged in terms of the tools they have to be able to serve those needs. So what's happening with new listings? What's happening with my competitors? What's happening as people move from high tax states where they want to potentially leave into new, more attractive, you know, tax and opportunity-based environments where they're not known to those lending institutions that maybe they're trying to be married up with. So we've developed a set of insights that is a subscription service, Truststar.ai, which goes directly to the loan officer. And then we use our platform behind the scenes to use things like the Home Mortgage Disclosure Act data, MLS data, other data that is typically exogenous to those sources and providing very customized insights to help that buyer journey. And of course, along the way, we can identify things like, are some of the decisions more difficult to explain? Are there potential biases that might be involved in that environment as people are applying for mortgages? And we can really drive growth through inclusion for those lending institutions because they might just not understand that potential client well enough and we can identify the kind of things that they can do to know them better. Yeah, and the benefit is really two-fold there, right? And Shai, I'll let you jump in, but to me it's two-fold there. One is, you want to have accurate decision, you want to have low-risk decisions and you want to be able to explain that to an individual that may get rejected. Here's why. And it wasn't because of bias, it was because of X, Y, Z and you need to work on these things. But go ahead, Shai. Now, I was going to add a point here, Dave, which is a developed point on the dam. One of the things that, and the reason why industries like Healthcare Financial Services are spending billions, it's not because they look at AI and isolation, they're actually looking at the existing processes. So established disciplines like CRM or supply chain, procurement, whether it is contact center and so on. And the examples that we gave you earlier, it's about infusing AI into those existing applications, existing systems. And that's what's creating the left because what's been missing so far is the silos of data and your traditional transaction systems, but this notion of intelligence that can be infused into the systems and that's what's creating this massive market opportunity for us. Yeah, and I think a lot of people just misunderstood in the early days of the AI, new AI, when we came out of the AI winter, if you will, people thought, okay, the incumbents are in big trouble now because they're not AI developers, but really what you guys are showing is it's not about building your own AI, it's about applying AI and having the tools to do so. That's right. The incumbents actually have a huge advantage because they've got these systems in place, if they're smart, they can infuse AI and then extract value out of that for their customers. Right, and that's why companies like IBM are an investor and a great partner in this space, Anthem is an investor of the company, but also someone who can utilize the capabilities, Microsoft, Intel, we've been really blessed with great backing, Northwest Venture Partners obviously is an investor in us as well. So we've seen the ability to really help those organizations think about where that future lies, but one of the things that is also one of the gaps in the promise is when a C-suite executive like a digital transformation officer, chief digital chief customer officer, they're having their idea, they want to be accountable to that idea. They're having that idea in the boardroom and they're saying, look, I think I can improve my customer satisfaction by 20 points and decrease the cost of my call center by 20 or 30 or 50 points, but they need to be able to measure that. So one of the other things that we've done at cognitive scale is help them understand the progress that they're making across those business goals. Now, when you think about this, people like Andrew Ning are just really talking about this aspect of goal-oriented AI. Don't start with the problem, start with what your business goal is, start with what outcome you're trying to drive and then think about how AI helps you along that goal. We're delivering this now in our product, our version six products. So while some people are saying, yeah, this is really the right way to potentially do it, we have those capabilities in the product. And what we do is we identify this notion of the campaign, an AI campaign. So in the case that I just gave you where the chief digital officer is saying, I want to drive customer satisfaction up, I want to have more explainable decisions and I want to drive cost down, maybe I want to drive call avoidance, and I want to be able to reduce handling time to drive those costs down, that is a campaign. And then underneath that campaign, there's all sorts of missions that support that campaign. Some of them are very long running, some of them are very ephemeral, some of them are cyclical. And we have this notion of the campaign and then a mission planner that supports the goals of that campaign, showing that leader how they're doing against that goal by measuring the outcomes of every interaction against that mission and all the missions against the campaign. So we think accountability is an important part of that process as well. And we've never engaged an executive, it says I want to do this, but I don't want to be accountable to the result, but they're having a hard time identifying, I'm spending this money, how do I ensure that I'm getting the return? And so we've put our secret sauce into that space as well. And that includes the information around the trustworthiness of those capabilities. And I should mention as well, when we think about that aspect of the responsible AI capabilities, it's really important the partnerships that we're driving across that space. No one company is going to have the perfect model intelligence tool to be able to address an enterprise's needs. It's much like cybersecurity, right? People thought initially, well, I'll do it myself, I'll just turn up my firewall, make my applications, roll access much more granular. I'll turn down the permissions on the database and I'll be safe from cybersecurity. And then they realized, no, that's not how it was going to work. And by the way, the threat's already inside and there's long-term persistent code running and you have to be able to scan it, have intelligence around it. And there are different capabilities that are specialized for different components of that problem. The same is going to be turned around responsible and trustworthy AI. So we're partnered with people like IBM, people like Microsoft and others to really understand how we take the best of what it is that they're doing, partner it with the best that they're doing and make those outcomes better for clients. And then there's also leaders like the Responsible AI Institute, which is a nonprofit independent organization who were thinking about a new rating systems for the space of responsible and trusted AI, thinking about things like certifications for professionals that really drive that notion of education, which is an important component of addressing the problem. And we're providing the integration of our tools directly with those assessments and those certifications. So if someone gets started with our platform, they're already using an ecosystem that includes independent thinkers from across the entire industry, including public sector as well as the private sector to be able to be on the cutting edge of what it's going to take to really step up to the challenge in that space. Yeah, you guys got a lot going on. I mean, you're eight years in now and you got now an executive to really drive the next scale. You mentioned, Bob, some of your investors, Anthem, IBM, Norwest, is CrunchBase right? It says you've raised 40 million. Is that the right number? Where are you in fundraising? What can you tell us? They're a little behind where we are, but we're stage B and we're looking forward to now really driving that growth. We're past that startup phase, and now we're into the growth phase. And we're seeing the focus that we've applied in the industries really starting to pay off. Initially, it would be a couple of months as a customer was starting to understand what to be able to do with our capabilities to address their challenges. Now we're seeing that happen in weeks. So now is the right time to be able to drive that scalability. So we'll be looking in the market at how we assemble that necessary capability to grow. Shay and I have worked in the past year with the board support of building out our go-to-market around that space. And in the first 100 days, it's all about alignment because when you're gonna go through that growth phase, you really have to make sure that things are pointed in the right direction and pointed together in the right direction. Simplifying what it is that we're doing for the market so people can really understand how unique we are in this space and what they can expect out of an engagement with us. And then really driving that aspect of designing to go to market and then scaling that. Yeah, I think it sounds like you've got, if you're down to days or weeks in terms of the ROI, sounds like you got product market fit nailed. Now it's about sort of the next phase is you're really driving your go-to-market and the science behind how your demand gen and your sales productivity and you can now codify what you've learned in that first phase. I like the approach a lot of times you see companies, of course this comes out of the West Coast, the East Coast guy, but you see the double, double, triple, triple, grow, grow, grow, grow, grow. And then churn becomes that silent killer of the software company. I think you guys, it sounds like you've taken a much more adult-like approach and now you're ready to really drive that scale. I think it's the new formula really for success for hitting escape velocity. Guys, we got to go, but thanks so much. Bob, I'll give you the last word. You mentioned some of your 100 day priorities, maybe you could summarize that and what should we be looking for as a market success? You know, look, I think that our measures of success are our clients' measures of success and the same for our partners. So we're not doing this alone, we're doing it with system integrator partners and we're doing it with great technology partners in the market as well. So this is a part about keeping that promise for enterprise AI. And one of the things that I'll say just in the last couple of minutes is, you know, this is not just a company with a great vision and great engineers to develop out this great portfolio, but it's a company with great values, great commitments to its employees and the marketplace and communities we serve. So I was attracted to the culture of this company as well as I was to the innovation and what they mean to the space of AI. And I said, I said, I'll give you the last word. Actually, I got a question for Shay. Shay, you're Austin based, is that correct? That's right. It's over here, but we have global presence, obviously operating out of Austin, other parts of the US, but offices in the UK as well as in India. So Bob, you're not moving to tax-free Texas, are you? Like everybody else? I've got an important home and life in Connecticut. So I'll be traveling back and forth between Connecticut and Austin, but I keep in my home there. Well guys, thanks for coming on. Best of luck. We want to follow your progress and really appreciate your time today. Good luck. Thank you, Dave. Thank you for watching this CUBE Conversation. This is Dave Vellante. We'll see you next time.