How AI Is Transforming Business Intelligence | Utilizing AI Ep. 9

It's always been a challenge for business people to extract insights from business data, but AI is changing this. Companies like Lytic are making analytics and business intelligence more visible and friendly to people who aren't ready to write database queries or build dashboards. A great dashboard provides continuing value that makes people want to revisit it, and the best are proactive.

Can AI help companies derive real value from their data? That's the topic on this week's episode of Utilizing ai. Welcome to utilizing ai, the podcast focused on practical applications of artificial intelligence from the Futurum group.

Each episode brings together diverse perspectives to explore news and use cases in the ways in which AI is transforming enterprise IT and the industries it serves. I'm your host, Stephen Foskett, president of the Tech Field Day Business Unit here at the Futurum Group. Before we dive into today's discussion, let's meet who's on the panel today.

Brad. Hi everyone. I'm Brad Shiman.

I'm the VP and practice lead at futurum for data intelligence, analytics, and infrastructure. And of the three analytics is my favorite. So I'm, I'm quite happy that we're having this discussion today.

And Paul, I'm Paul Blankly. I'm the CTO and one of the founders of Lytics Lytics. It is an analytics agent that helps answer questions, uh, that people will need to ask in their day to day to make better decisions in their work.

This could be anything from an executive making a decision about, uh, how to better allocate resources, a marketing person allocating ad spend or procurement, making better decisions on, on how they, uh, buy product for a manufacturing company. And as I mentioned, I'm Steven Foskett. I've been hosting the utilizing tech, uh, podcast and utilizing AI podcasts, uh, for, uh, five years now.

And we've watched as AI has grown in usefulness and practical applications. Uh, we launched the utilizing AI series with the Futurum group analysts last year and all the time the plan was that we would eventually be inviting on guests to join us in the conversation who are doing interesting and relevant things with ai. I got an a briefing with Lytic right about the time that we were planning this series.

And I have to say, it really, um, appealed to me because both of us had the same idea, which is basically, AI is cool, but it's also useful. So let's talk about how it can be used for practical business use cases, and that's really what Lytic is doing. So before we begin, Paul, let's kind of dive in.

What was the question that was being asked that you decided to apply AI to answer? I think some of the best ones are, are questions that you can't easily answer by yourself. So a good one, that was where I used our own product, uh, to analyze our own users of our product was I had this theory that people were using the product in different ways.

Most of the people chat with the product, but I, I went in and I asked Zoe, our, our agent, that same question. I was like, Hey, can you break out how people use the product based on their usage of different features into different clusters? And then, you know, gimme a sample from each so that I can go have some 15 minute conversations with them and better understand their use cases, what they're trying to solve.

So Zoe went in and helped me, you know, find basically three buckets of people. There's like the people who use, uh, chat the most. They're talking with the AI agent, definitely the pro, like the bulk of our, of our users.

Then there's people who use that and then also use dashboards and other sort of more traditional reporting mechanisms a lot. That also makes sense. Then we had a few people down there who kind of just use the dashboards and like, almost never the thoughts to the AI agent.

So I was able to find these three clusters and then go and have conversations with each of those to better understand their use cases, where we're meeting their needs and what we could do better as a, as a product. So that's an example from my just day-to-day of how I use, uh, how I use agents to, to better help my analytics needs. Sorry, Brad, I was waiting for you to just dive in.

Uh, we'll, we'll Out. Yeah, I, I, I apologize guys. And dive and Corey.

Yeah. Apologies, man. I, I didn't know if you were doing this more formally, Steven.

Yeah, No. Yeah, I might have, but I didn't. So you, I suspected you could get my way, So, okay.

Let me just start like, I'm, I'm just responding. Yeah, I I like very much what you're, you're saying, Paul, and, uh, you know, I see this in our research, uh, quite frequently that, um, as data professionals work through their DA daily tasks, which are many and varied, um, that they are, have found a great utility in using, uh, ai. I, I think initially, um, in the market, especially after we invented Transformers, there was this, uh, sort of idea or a thought that, well, they can't handle structured data.

They can't work with a CSV file, they'll just fall over. Or, you know, we, we can use them, but, um, we can't trust what they say. And I, I find that ironic for a couple of reasons, because, um, in, in my, I'm also a practitioner as well as an industry analyst, and in my, um, job as a practitioner practitioner, I, I find it invaluable for, um, almost daily, I'm, I'm finding new ways to use this, uh, in, in making my job easier, whether it's like you're talking about doing some basic segmentation or if it's just data cleaning and prep, uh, or if it's something, you know, more deeper like data mining, like, like you're doing, you know, it, it is tremendously, uh, I I would say useful and trustworthy to a degree.

But I, I say it's ironic because when I look at the industry and I look at how, you know, companies are using BI in particular and they're looking at dashboards, I feel like they, uh, have more trust, uh, you know, in, in just their gut instinct or in the dashboards themselves. And they, they put more emphasis on that than they do on, on ai. So we seem to demand more transparency and accuracy from AI than we do for bi.

We kind of say, yeah, it's a black box, but I trust it. What's, what's your, um, experience Paul at, at your, at alytics in terms of how your customers are looking at their dashboards versus ai? Oh, yeah, I think that's a, that's a really good point about how a lot of times people trust their gut instead of, you know, what's ever whatever's on their dashboards.

I think there's, there's a really good reason for that. Actually, the reason for that is, is maybe best personified in, um, in, in a conversation I had with one of our, with one of our pilot customers. Uh, so I was talking to this, to this woman who's a, who's a PM on, in a major public company, and we were going through, she was asking some questions, but only we'd only imported a few data sets so far.

And she was like, Hey, well, I'm, I'm not really getting that much outta this. Like, what's, what's going on? And I was like, well, you can ask about all the stuff you could ask about on these, you know, that's on these two dashboards.

And she was like, well, I don't want that. It's already on the dashboards. And the, the dashboards don't tell the full story.

That's why if you're a PM trying to make a nuanced decision about how people move through your login flow, knowing some basic title level numbers doesn't really give you the information you need, doesn't really answer the questions you care about and is therefore just not that relevant for you. It's maybe useful to monitor, to understand kind of the, the basic parameters of what you're looking at, but it's not gonna actually impact your decision. You need way more sophisticated analysis than you can get out of a, out of a dashboard to actually impact a decision.

And that's why for these, these analytics agents, uh, what we're building at Lytic to be impactful, they, they've gotta be able to go way beyond what you can do in dashboards. The PM has to be able to come in and ask really, really nuanced questions, which are not only the valuable questions for him or her, but they're also the ones that are way harder to explain your methodology. We explain what you're doing.

Um, and that's what, that's one of the things that the agent has to really excel at to make sure that, that, that that and business person can trust and can actually act on confidently on the data that they got back. It does seem like people are, are trusting AI already in many cases to give them truthful answers, even as there's this sort of understanding that, uh, generative ai, you know, for example, the search box, you know, Google type generative AI isn't always trustworthy. It seems almost a paradox that people trust AI more than they themselves would say that they trust ai.

And at the same time, I definitely feel like people trust AI more than they would trust people, you know? And and that's a, you know, even people who would know things, right? I mean, and I wonder if, if, um, you know, kind of looking at my history with, uh, data professionals and how they've interacted with the business, um, to Brad's point, it Does, does seem like sometimes data professionals have labored long and hard to come up with a dashboard or to come up with a metric and, and, and only to have somebody say, well, yeah, no, and, and I wonder is, is is AI going to change that a little bit?

Are people going to trust AI more than a per more than a data pro? I think, yeah. Were they even use dashboards in the future?

Sorry, sorry Paul. But, uh, yeah. Will we even have dashboards?

Are they, are they necessary? Is, you know, are we moving toward an interface that is the human voice and nothing more? And, uh, you know, will we have what a lot of people in the industry would just call headless bi, that's ag agentic that would, you know, find the answer to your question and do, like, Paul, you're talking about with digging deep into the meaning behind the data and the context for that meaning, and putting that into perspective and bringing back like good answers.

I would love that. Yep. I, I actually think there's two different use cases in data that are, that are often conflated, and that's why we get the, like, dashboards are gonna die.

I don't think dashboards are gonna go in anywhere. Um, dashboards solve this monitoring use case, which is like, I need to see the same stuff every week. For me that's like, Hey, what are our usage in the different model providers?

How many people are using Claw now versus OpenAI? How many loggings did we have? How is that trending?

Like, which customers are active, which customers decrease their activities? Like that's just monitoring stuff I wanna see every week regardless. Um, but then the, the really valuable use cases usually aren't monitoring.

It's usually a question that you're like, oh, okay, I really need to deeply understand how people are using this. I need to understand where this login flow is breaking. I need to understand, you know, what's going on in this area of the business.

And those are those deep questions that are just not served well by dashboards, and that's what the, what the analytics agents are gonna are gonna take over. Yeah. Do you think, Paul, that, um, through eight analytic agents, um, that we can get around, what, what seems to be, to my, to my mind, one of the biggest problems with, uh, dashboarding and traditional bi, and that is utilization across the business and what that means in terms of, you know, how a business actually runs.

We, we've been laboring for decades now, trying to, you know, have data professionals build beautiful dashboards that get used. Uh, but unfortunately, you know, it, it typically ends up where, you know, the analyst spends weeks or longer building a dashboard that nobody uses. Uh, and there's no true understanding of why it isn't getting utilized.

But I, I would love to, to think about, you know, uh, as, you know, an industry watcher, the idea as you just talked about, of being able to balance those two, you know, the two sides of that coin, the, I wanna just have the same data all the time, and I wanna be able to dig into what I want and to have that available to everybody in the business, you know, is that possible? Do you, do you think from, you know, analytics perspective that we can change the business as we've been trying to do for decades now? I think, I think there's, there's two ways to think about that.

It's one, I mean that the, the business teams, obviously, they're not getting what they need out of the, these dashboards, even if it takes a very long time to develop, um, or the dashboards answer a one time question, which I think is probably most of the pattern where a business person has a need that need is for a specific one volume stone, a dashboard gets built, it takes a long time. By that time, it's almost rear mirror. It comes in, the person looks at it, they, they understand the one thing they needed to understand, and then they never revisit it again because it was never a recurring mood.

It was never something they needed to monitor in the first place. And dashboard needs to buy a vehicle to, to, to provide that answer. The, um, the other one though is like, how do you increase utilization?

Like, how do you get business people using data more often? There's certainly the, you make it really easy to ask questions. Anyone can come in and ask an analytics agent a question, get a good answer back, and, and that's a great thing.

That's a great utility. That was, I think, actually only the first step in this because there's only so often you have a question, you think, oh, this is actually something that would be really helpful to have some data to like back up or support my decision here. That that only really crosses your mind, not that often in the workday.

So you wouldn't expect a, a massive amount of questions, but are just people coming inbound. Um, to really solve that, you need to do what the, what the absolute best in, in our industry do, which is you're proactive. You're, you're not just, whenever you're asked a question, you're reactive and you're coming back and giving an answer that's good.

Um, you, you need to say, okay, well, if I was a VP over, you know, our, like procurement in a manufacturing organization, what are the things that I would be most concerned about? How can I proactively go? And before my VP comes to me with questions, I, I come to them and I'm like, Hey, you know, I've, I saw this trend in, you know, these park suppliers.

I know that we have rebates that we can claim on suppliers once we increase X volume. I found all those suppliers. Here they are.

I've drafted emails even, like, just gimme the, okay. And I can send those for you. It's like, that's the level of correct proactivity than the absolute best people in organizations have.

And I think the agents can absolutely emulate that behavior. And that's, and that's where you get the utilization that's really impactful. It's not just people coming in and asking the agent questions.

The agent needs to be proactive. The agent needs to be able to go to them and say, I thought this would be relevant for you, and it needs to be relevant. And that's how you build that kind of, uh, continuing value.

And you get people to really invest in a tool, I think, is to have them have it be something that is not just, um, not just there for them, but like you said, kind of providing proactive value. It is exciting to think that that is possible in this BI space because again, you know, like we said, BI has always been a very reactive, you know, uh, kind of plotting process where you basically, you know, you're almost requiring someone to know what they want before you build it for them, and then they, then you go and do it, and then it, you know, they go back and forth and, you know, I mean, Brad, we, we, we experience that sometimes internally with our dashboards. And it's frustrating because, um, what you really want is something that, that just gives you that information.

Think about the metaphor, you know, you look at the vehicle dashboard in your car, right? You know, it's providing you the key information right. When you need it.

That's really what we want from our business analytics, right? Yeah. You want your car to have a dashboard that, that shows the things that are relevant to you at that moment in time.

Like your, your left tire has less traction, uh, coming into this turn. So I'm going to turn on four wheel drive for you so you don't crash. That.

That's what we're talking about here. And we've seen for, you know, quite a few years now, especially after the, uh, chat GPT moment that, um, you know, companies that were building bi tooling were, were trying to do that with SQL text to sql, voice to sql, and with, uh, things like explaining dashboards and, uh, tooling that, that could automate or at least augment some of the workflows that data professionals went through so that they could maybe have more reactive and more immediate, timely access to data. But, you know, I, I feel like we're still a long ways away from that overall as, as, you know, an industry just because the inertia that companies have, uh, in, in getting to that and that mindset, uh, is something that I, I think is still gonna take some time to shed, even if we have the tech in hand to do that.

Paul, do you see in your customers any, any kind of like, um, do, do they seem like they, they understand and grasp how they can make that leap to that, that real time responsive dashboard for their, for their car? I think, I think, I think a lot of people do, but it's not, it's not evenly distributed. It's kind like the future's here, but it's not evenly distributed.

Uh, what, what we see happen most, which y all actually heard is from a lot of other companies that are, that are sort of AI natives like us, is you'll have these power users that just do incredible things. They're usually the people who are, you know, more, they're interested in ai, they wanna push these tools to their, to their limit. Um, one of these users in, in j Crew, uh, human, and he did a deep dive on, sorry, to we better set up the assortment in our Upper East Side store, some of the most expensive real estate we have in the country.

Um, and what she said is, Hey, we've always put stuff, we always put clothes in this store based on historical purchases in this store. What if instead we looked at the humans who are shopping in this store and picked what they buy through all the channels that we know about all the other stores we shop in everything, and we, we tailor the store for the persona that most often shops there. So we took that approach, which is way more complicated, but with, with Zoe, uh, you would be able to do it.

And they actually redesigned the Upper East Side store and in, and increased dramatically the amount of children's flows that were there, because that's where they found out that that, that that persona wanted to buy. And that's the kind of thing that you have this power user that just has outsized impact on the organization, um, because of that. At the same time, you have a lot of other people coming in, getting answers to questions, getting answers to questions like, how many transactions did we have that came in through such and such campaign?

And, and those are fine, those aren't bad questions, but those aren't at the same level of impact. So I think the thing that we'll increasingly see as people get comfortable with AI products and, and these, uh, these power users can kind of influence other people in the organization, is that more people feel comfortable coming in and asking those big questions. Like, try it, try to get it to do something you don't think it can do.

Maybe it'll surprise you, Right? Break, break it. Yeah, exactly.

And those are the kind of, um, sort of unexpected insights I think that, that companies wish they could get from their data, right? I mean, that's what, that's what they really want. That's why they've been doing this all this time and trying to build these, you know, data warehouses and building analytics and building dashboards and trying to get business intelligence meetings right there in the name.

Uh, but unfortunately, I feel like the whole process has been just so reactive instead of proactive, like you're describing that they just haven't been able to do it. Is that your experience, Brad? It's, and I, I would say that it, it's, it's funny because, uh, when you talk to data professionals, uh, and it across the spectrum, so every, everyone from a casual business user that's using data to the, the person that's in charge of maintaining, you know, realtime data pipelines, um, or, or the data scientists working with the data, um, you know, in, in very select deep ways, uh, and all of them have, like, you know, said, you know, we are shifting toward the business.

We want to be able to focus less on the syntax and more on, you know, the intent. And one of our, one of our, um, uh, we, we do our yearly prognostications, and one of mine this year for this re this area is the rise of the AI Shepherd. I'm, I'm calling them.

And it's, it's basically to say that, you know, we, we've tried augmentation, that's old news. New interface is natural language. Uh, people don't want to write SQL to, to get something done.

They want to just state their intent. And how you do that, whether it's in an interactive dashboard that has Zoe, let's say, as your co-pilot to work with you on that, or whether that is just in a chat interface or whether that's in a line of business app, uh, that you have, you know, your traditional pull downs for what you're working on for that workflow is irrelevant. They, they want to just be that, they want to be the person that has AI with them to, to make these decisions, to bring back these insights.

So we're seeing, you know, the traditional data, uh, professionals shifting toward this. I'm not a technician anymore. I, I'm more of the validator, curator, adjudicator, uh, facilitator to, to, you know, innovation built on top of data.

So they want it, they, they really, you know, the data professionals really want it. The business users, I think, you know, need to, to sort of see that. And as Paul you mentioned, you know, if you have people in the organization that can trumpet that and, and promote that as, yes, this is here today, we can do this today, that that's what we really need for everyone in the, in, in the organization to be able to say, you know what?

I don't have to wait two weeks for somebody working in SPSS to build a dashboard for me. I could get, I get an answer like right now. Totally.

I think the, I think your, your way of framing it as, as AI shepherds is really good, because when you think about what are the barriers to, to these agents actually working inside of a complicated enterprise, there's like two facets of it. There's context and there's sort of like explainability. It's like, do I know what you did?

Um, the context part is straightforward, but still complicated, which is there's a lot of historical patterns you have. There's a lot of facet knowledge hidden in dashboards, hidden in databases, hidden in the, just the head of, of the data, people who have been having to answer these questions and accommodate all the weirdness in their data, uh, because everybody's data is weird to some extent. And, and getting, getting all of that information and making it really easy for, as you use the product and as you get value to add that context in as you go, instead of having to kind of put it all together and boil the ocean before you can start asking any questions.

So, so that's absolutely key. That's like gathering the context. The other one is the explainability, and like, why do I say explainability instead of trust?

Because any really good analyst, the really elite analysts are gonna come back and not just explain to you like, Hey, I got this such and such number. They're gonna say, Hey, I fo this is my methodology. Like this is, this is what I followed to get to this.

And they're gonna actually show their work. And what that does is, instead of the VP or whoever coming back and being like, that number's wrong, obviously, like, um, have you guys even looked at this before? Um, they're, they're gonna say, oh, okay, I see you followed this methodology when I wanted you to do this, but I didn't actually say that you should do this other approach instead.

And that's the same way that an AI agent builds the, the trust through that explainability and understanding where that end business person can understand, Hey, this is the approach you took. Great. That's what I wanted.

Or actually I wanted to see it a different way. I wanted you to approach this differently. And, and those are the two sort of fundamental things.

It's like that context of like, what's going on in the business, how do I answer questions correctly for this business? And then the showing your work, the, like, how did the end user know what I did and can be confident in the, in the end result? You know, what AI is really good at right now, when you, when you look at, um, areas like, uh, software development in particular is in documenting, um, that methodology, um, and in turning it, we, we've had some tooling now that just has swarms of sub subagents that all will go toward tackling a, a given problem and they'll come up with a plan called a spec, spec driven development.

And I think that we should start seeing if we're not already the data professionals picking up that same methodology for how they do it, because AI is terrific at building that, that taking that knowledge that usually just sits in somebody's head and operationalizing it. Yeah. You know, that's, that's a really good point.

And I think that that's maybe something that we can kind of step back from the specific conversation about business intelligence and, and, and talking to the folks who are listening to this, who are not maybe in this field, but are wondering, you know, what, what, what could their takeaway be? The idea that AI can help them, um, I guess understand the whole process, you know, talk to us a little bit more about that, Paul. Yeah, I think that's, I think that's maybe the most important aspect because what you, what you want is you want a virtuous loop where the person can ask a question, get an answer, they really understand how the AI went about getting that answer.

Then they're more equipped to ask a better follow-up question. They're more equipped the next time they have a question. They're like, I actually know a bit about how this, this works now.

And then they're, they're able to ask better and better questions as you go. And you get this virtuous cycle where you ask a question, you learn a little bit more about how this is structured, you get the answer you need, then you can ask better questions as you go. And that you get this virtuous cycle where people are asking better questions, getting better answers that are more relevant to them and all because they actually understand what approach to AI is taken.

And that's the, that's the just so crucially important. Yeah, I, I'd love that because, um, as we were just talking about with, with documenting all of that institutional knowledge and not losing out on, on tribal knowledge, um, is one of the biggest challenges that companies face. And if you can introduce a technology, whatever kind of technology that preserves those and, and honors them and puts them to work in improving that cycle of question and answer, uh, whatever domain you're working in, whether it's bi or, you know, sales enablement or, or you know, any particular area in the business that, you know, an investment in that is an investment in the future of the company.

'cause as we all know, um, jobs are changing, roles are changing, but the business itself is, is changing. And so we need to, as an industry, invest in the technologies that can, you know, take the concept of this business, everything from the semantic layer up to the, you know, business decisions and the data points that sit on that interactive dashboard and to make that more resilient and more responsive to, uh, an ever-changing environment that we all live in right now. That's what we need.

Yeah. And the real, the really important part there is, it's gotta be really easy as you go, as you're asking questions, the agent is remembering things. You are able to say, Hey, great, that's that concept.

Remember that now it needs to be really, really easy as you're going. It can't be like a big, you know, months long project. Like set something up before you start using it.

You've gotta dive in and start using it now and have it learn with you as you go. That's the, that's the pattern That's called forward slash memory as you're, as you're typing. And, uh, they could tell it to remember anything you want.

Well, you know, it's funny that, um, you know, the most effective, um, you know, AI power users or generative AI chat chat power users that I've talked to, the, the way they do it is kind of flipping it on its head and having the, um, having it interview you instead of you interviewing it. So instead of saying like, Hey, give me an answer, you say, Hey, um, if I wanted to achieve an answer, what were the, what are the questions that you, you know, I should be asking what are the facts I should be bringing forward? Um, how should I be thinking about this?

And it helps you to organize your own thoughts and your own queries in a way that can help then the AI helping you and, and I guess Paul, that's, that's the kind of thing you're talking about, right? That, that, that the more questions, the more interaction you have, the better it gets instead of the worse. Yeah, Yeah, yeah.

It's, it's gotta improve. And then I think, like you said, the other, the other really component, the, the really important component there is asking the right questions and asking the right questions means, just like you said, you're not asking for a specific data point. You can do that and it'll give you an answer, but you're telling it the problem you wanna solve instead.

Yeah. Because almost inevitably when you do, it'll, it'll pick out that data point you had in mind for that problem. That'll, it might also add another one that is really relevant that you just didn't think about.

Humans are bad databases and the language models are actually pretty good databases, so much better than humans. So ask them the, ask them your goal question, ask them what you're going for, and they'll probably catch something that you just didn't think of in the, in the moment. It's a really helpful pattern.

Yeah. I like to have my agents, um, use the Socratic method to, to argue a point. Well, on that note, um, perhaps then, shall we continue the No, I I, I, I'm not even gonna try, I'm not even gonna play that game.

That Was cute. Yeah. So this, this has been a really interesting conversation and I really appreciate the fact that we were able to, you know, kind of go from specific to general here and that we're able to come up with some really interesting ideas for how people can make best use of this technology.

Um, before we go though, I do wanna give you a chance, uh, Paul, uh, tell us a little bit more about Lytic and, and where people can connect with you and find you and learn more. Yep. You can find me on LinkedIn.

com. If anything that I've talked about is interesting, please, you know, chat with us, book a demo. Um, I'll also be talking at day-to-day Texas in a couple weeks, so if you're there, gimme a pin.

I'd love to, love to chat, uh, while we're out there in Austin together. Great. Um, Austin, hey, that's where Futurum is.

Uh, Brad, how about yourself? I know you're not in Austin, but, uh, where are you gonna be lately? Um, I'm actually home for a couple of weeks, which is, which is nice.

Um, and I'll be working on a couple of, um, comparative reports that are themselves using generative AI for, uh, a project that we have internally, uh, here that, uh, we call Signal for Signal reports. And one of those that I'll be working on is around semantic bi tooling. And, uh, so analytics will be, will be featuring in that as well as many others, um, because this is a rich, rich marketplace with a lot of great people working on, on this solution we've been talk, talking about today.

So I'm, uh, I'm looking forward to, to getting that going. Yeah. And if, uh, people wanna see that report, uh, Brad, where can they find that?

com and they can find me on LinkedIn all the time at Brad Shiman. All one word. Excellent.

Well, thanks so much for this and, uh, thank you everyone for listening. I hope you found some, uh, interesting ideas here for your own use of ai, your own practical utilization of ai. Thank you for listening to the utilizing AI podcast this week and every week.

If you enjoyed the discussion, please do subscribe. Uh, you'll find us on YouTube as well as in your favorite podcast application, and of course, a rating and a review would be helpful. This podcast is brought to you by the analysts and experts from the RUM Group where insights meet ai.

ai, the utilizing AI YouTube channel or the Text Strong TV app. Thanks for listening and we will catch you next Wednesday.