Keeping the human touch in an AI-driven future

Like all technology, all new technology there, people who will try to exploit analytical uses with that technology and I'm sure generative is no exception and as the technology becomes more powerful, fully seeing it and controlling it probably requires more powerful tools as well. It's an area of concern overall. Hello and welcome back to Conversations with ZenBesk where we explore how new trends and technology in customer experience are shaping the way that businesses connect with customers. Each week we speak with customer experience innovators and experts to hear their thoughts and ideas on the future of CX. I'm your host Nicole Saunders. Last time we heard from Christina Funseka, VP of Product for ZenDesk's AI Division and learned all about how our tech is evolving. This week we turn our attention to how some other companies are developing in the space of generative AI. I speak with Deepam Misra, a senior advisor to AI and machine learning startups and the leader of several advisory programs for startup scaling generative AI and joint innovation at Amazon. In this role Deepam advises some of the world's leading AI and ML startup founders as well as enterprise AI leaders on how to bring artificial intelligence into production. Deepam has over 15 years of leadership experience in startups and large enterprises. He has founded four AI and ML startups, was the CEO founder of Venture Factory and has had several leadership roles in enterprise innovation. Our conversation is full of all sorts of interesting insights and ideas about generative AI. So let's dive in. Ready to take your customer experiences to the next level? Build lasting relationships, the ZenDesk complete customer service solution so that you can exceed every customer's expectations. Sign up for a free trial at zendesk.com. Welcome to Conversations with Zendesk. How are you today? Great. Thanks for inviting me. Very happy to be here. Well, I'm so excited to speak with you because you've actually been thinking about AI and working in this space for a long time, probably longer than a lot of people who are getting on the AI hype train more recently. I would love to have you share a little bit more context about your experience and your background with AI. I have spent way too many years around the AI that I get to remember, but it's still a passion for me and it's funny that now people are saying it's the age of AI even though I've been doing machine learning for a long time. So I started my career as a research scientist at Stanford Research. I was working in machine learning computer vision. I have had various roles creating my own startups and founded four different machine learning startups and as well as a small early stage venture fund. And for the last few years, I've been working with big technology firms like Microsoft and now with Amazon primarily focused on AI machine learning startups. So I saw that you recently announced the launch of the AWS Generative AI Accelerator. Tell us about that. What is that all about and what kinds of things are you looking into? So we were very excited. We launched the Accelerator just literally yesterday or day before. We were really overwhelmed with the response. The intent was to find the world's best machine learning startups in Generative AI and work with them to both help them scale and accelerate fast while we also learn to see what are the challenges through the eyes of these amazing pioneers. What AWS and Amazon can help to solve as you probably know in Amazon more than 90% of what we build is based on customer feedback. And this is one of our mechanisms to get customer feedback by being part of the journey with them. So the Accelerator offers a bit of financial support, a fairly generous $300,000 Amazon credits for building on AWS as well as it offers within my opinion a lot more value. The reason why so many people have applied is we provide bespoke technical and business mentorship and support to all these startups to help them leverage the Amazon ecosystem. And we have as I mentioned a tremendous response in terms of applications even with our best of estimates we were way over subscribed. But we were very excited that we've chosen about 20 or 21 startups who are participating in all different areas of technology, different industries, different parts of the world. I would even say diversity of founders. It's an amazing cord and very excited about working with them. Everybody has been hearing a lot about what open AI is working on and what some of the bigger players are. But it sounds like you're seeing a really diverse field of people developing in this space. What are some of the more interesting problems that people are trying to tackle or applications that they're working on? My work is a lot focused around startups. So I can certainly say that from the startups perspective the focus has been very wide and pretty much across the board. Both in terms of diversity of applications, the kinds of technology, focus areas as well as possibly business applications and consumer applications. It's very, very diverse. So I wouldn't say even in our own accelerator which I was referring to, we were just discussing this over a dinner and we feel like this almost every facet of human life as well as businesses, all functions of businesses which are virtually touched by even these 20 or startups that we're working with. So it's kind of hard to bracket them. The current hype as well as excitement is around doing enterprise search and document analysis in terms of analyzing your data and being able to generate more value that especially unstructured information like text files, all kinds of logs, even for example in customer service, you all probably record and save a lot of customer feedback. A much of that information can be analyzed for more information. So there's a lot of work happening there. Obviously in consumer areas, there's also a lot of work in image and video generation for entertainment or graphics for game development. It's really fascinating to see the wide areas going on. I think the patterns are still emerging in probably next six or 12 months we can see what the highest and updated applications are. We talk a lot about obviously customer service, customer support, how people are finding information. What are some of the other realms that people are thinking about with this? You mentioned a little bit around like AI image and video generation. Are there any other unique spaces that you think aren't getting talked about as much that are really interesting? Yeah, it's almost you can imagine what generative AI does. It just stretches our imagination boundary. So yes, there are the first few applications. I think we will see people are saying, okay, what we used to do with conventionally, I can be try to do that better with generative AI. So for example, if you had like a chatbot for customer conversation or you had even some kinds of bots to understand customer feedback or you had some image analytics bots, which could allow you to do kinds of interpretation of different types of image data, you can do a lot better with generative AI. Well, that is true. I think the real interesting part is net new applications of things which we could not do before, which we could not even imagine before. That's really the fascinating part about generative AI. So like point you mentioned text to image generation. For example, imagine I show this video in one of my presentations where we able to generate a simulated video sequence of a water droplet dropping on a surface of water showing that in slow motion with a background of sunset with a certain color use and the color tones around it. All of that can be done just by somebody's imagination and typing a text script. And you can generate very sophisticated video and image week from that we birth with one startup example runway ML, which has just finished running. I think it's the international film festival for generated images or videos. It's amazing to see the kind of diversity of what that people presented in that. And it's just a start. I think we get a lot more exciting applications from generative AI in areas that we probably can't even imagine right now. That's amazing. I didn't realize that there was enough of it out there right now to actually have a film festival around it. That sounds like it's way further along than maybe people are expecting or aware of right now. There's still a lot to be done, but yes, certainly a lot of experimentation already happening. There are gaps in the technology as we all know, but folks are doing amazing things and just I would say it's an advanced experimentation stage, but there's a lot of work happening in pretty much all of these areas. So that brings to mind an interesting point. Right now we're in the middle of the writer's strike. And they're very concerned about AI replacing them in terms of writing the words and the scripts and the jokes and all of that, which I think is a legitimate concern hearing about video getting produced. It sounds like this is something that we're probably going to have to tackle in a variety of fields. Certainly one of the concerns with AI is how do you keep the human touch? We talk about it in customer support of how do you make a bot empathetic, but not too empathetic, right? You still want to know that it's a robot and not a person. What are your thoughts on how as companies are approaching these new technologies, they need to balance this amazing innovation? and with still keeping it human, and what is that role of the human interaction with it? - I wouldn't say if I am probably more of an expert on the technology side of things. I'll say, for as technologists, we will want to make sure that we provide both the tools to allow creators to experiment and arrive at the right mix of human and computer generated content, but also help them address one of the bigger concerns that most companies developing in this or businesses trying to adopt this is not understanding how the technology works. So being able to understand a bit more about the nature of generative AI tools, how they produce, what they produce, and hence, once you understand better, you can control better, you can optimize better. So things like explainability in generative AI is an industry challenge. Being able to control the kind of content that you can produce both from a perspective of what you mentioned as human versus technology generated content, but also from perspectives of opirides, of original content, these are all very important questions. And as technologists, what we are looking at solving is building those guardrails and those types of tools that allow developers and generators to have more control of these systems. Ultimately, I think people who are far more experienced in each of those areas of creation will decide and probably some industry groups and others will come together to decide what's the right balance. And my own personal opinion is human creativity keeps getting better. Every time a new technology comes around, just that humans tend to want to create more things or maybe a higher value added things than they were doing maybe a few years ago. So I tend to take that perspective, but from a technology perspective, I'm very keen on helping creators get control and more insights into what they're developing. - I love that you bring up how creative humans are and how human creativity keeps evolving. It feels like there's a lot of fear about how these technologies are going to take over what we do, but I think some of the real excitement is in what it's going to enable us to do that we can't even imagine today. For example, when Twitter first was a thing and people had to start composing in these short snippets. And there were all sorts of new memes for one, but ways that people communicated things and creative things people did with those limitations. So it'll be really interesting to see where generative AI takes us and what new things come out of that as well. I think it's important that we balance having an awareness about the concerns with the really cool things that could happen. - So to that end, I know something that you have thought a lot about is what is just hype and what is actually really interesting and really the meat of what is going on in AI right now. So what if is your take on that? - Wouldn't call myself as a sweet share and be able to predict things will be in this technology is moving so fast and every day we are learning. So you've got to be humble about where the direction the technology can take. But I would certainly say some of the fears are little or blown in terms of not being able to control the output and outcomes of these technologies. Like always, some concerns are real about certain types of jobs getting impacted. I would think like always technology drives to first automate and eliminate those tasks which are lesser value add are probably more repeatable. So that I think is generative AI is probably no exception likely to happen in some areas. But I also feel some of the things that people are expecting to happen are probably too complex for technologies to be able to deliver automatically right now. So for example, even customer service jobs are providing that human touch in content creation or any form of artistic expression or any kind of human to human engagement. What makes us human is a bit of that touch which I think we will always want to retain as my personal perspective. I think what's real and there's a lot of debate and discussion around that. And of course, we'll see how the job impact is in the near future. But I'm also more excited about seeing there's lots of new types of jobs and opportunities getting created. For example, I've seen some startups just recently. One startup is helping you create content on demand or like for example, this is the next level up of streaming content and entertainment where you have so much to choose from. You're probably overwhelmed today if you're looking at Netflix trying to decide what to watch this evening. You could probably now think about content creation happening which is very customized to your liking. Just as some can put together a milieu of content which is customized just for your watching or entertainment or depending on what mood you are in, what time of the day you are in, and just different ways of organizing content which could not be done before, very customized for each individual user of you. So the kind of personalization and the kind of possible content generation for that kind of deep personalization is a completely new set of experiences, startups and other companies are going to try to create. On terms of the reality where I think a lot of business investment is going on, I think there is part too much focus perhaps on the entertainment side because most people understand it. But in terms of the businesses, what I'm seeing is businesses have so much of data and content which has not been used in the past. I see a lot of investment going in, a lot of interest going in trying to understand how to get value out of that data. And just a very simple statistic is no way to verify these numbers, or somebody in a panel said roughly 70%, I think is probably closer to 80% or 90% of information is in most companies is in the form of text and other documents. Only 10, 15% is in terms of numbers which we can actually analyze in your Excel spreadsheets and other kinds of spreadsheets. And right now there's literally no way to analyze that 70 to 90% of information easily without generative AI. So we have information still sitting unused in businesses and corporations, which I think is the real hot topic right now in terms of business investments. And I'm seeing that like you to be one of the first and the bigger adopters of generative technology in the near future. - I am all for anything that reduces the number of hours we spend scrolling through Netflix menus and who many is a fellow of those kinds of things on the weekends. I would love something that helps us find things that we want more. It is interesting to think about how jobs will be evolving and I think there's been a lot of talk about how there's gonna be new skill sets brought into jobs, right? Those of us in the marketing space, learning how to write prompts and work with AI. But you're right, there's also gonna be a whole new category of jobs for people working in this space. Your point about text analysis is really interesting too. My career has been largely in the online community space and one of the challenges that has always faced us is how do we take 10,000 community posts and summarize the themes of those, right? And it's mostly happened through like meta tags or the literally team sitting down with a big box of donuts and reading through 1,000 posts and writing down what they think it's about and sort of doing again audits to get those themes. And so you write that there's so much valuable information that we will be able to better leverage and better utilize once we're able to use these technologies to help us identify those trends, understand it, do all of that reading for us and help us figure out the themes that we can then take action on. If there's anything we've seen with development of technologies in the past 10 years, it can go in a wildly different direction than we expect. So then let's take a look backwards in time because you've been in this space so long, let's just even time box to about a year or two because and I know that AI has come into popular awareness really this year, like the release of CHAPGPT and where were we just a year or two ago in this space and what's different now? What's the evolution that you've seen in the recent past? >> It's a great question and we tend to forget, yeah, it has been around, even Amazon has been doing machine learning for 20 or years. In fact, I would say almost everything that Amazon does there is machine learning involved in it. Machine learning as a, again, I don't want to get too much in the technology side of the discussion, but generative AI is really a subset of machine learning or even larger beyond that as artificial intelligence. So we've been talking artificial intelligence in 1950s, but as you rightly said, the hype cycle or rather the interest of generative AI has captured the average user across all age spectrums across the globe. And that's because in my opinion, it has become easier for people to use AI in the last maybe six to nine months. The more power that we have been able to put behind these machine learning models, it has made the front end and the engagement model with human operators that much simpler. Now, it comes with its own set of challenges which we can talk about separately, but the big change has been just the size of the machine learning models, just the raw horsepower of computes and the amount of data that we are now using to train these large generative AI models which are also known as foundation models, that's just unprecedented. And it's grown by orders of magnitude. Just by, if you look at class three or four years, the first, what I would call large models, large language models, even four years ago, were about 300, when the birth models were probably around several tens of millions, or hundreds of millions of parameters. And in the last year and a half, we've gone on the GPT-3, which was about 175 billion, to some models, ladies this year, I think from the open source community, as well as probably from some private providers, which are more than a trillion parameters. So the size of these models has grown exponentially. Yeah, that's huge. Yeah, that's huge. And what it does is, the way I explain to people is the number of parameters is just directly proportional to how many patterns the system can understand. And at the end of the day, machine learning is primarily a technology of understanding patterns between input data and output data. And then you're able to match those patterns on unseen data. This is how you solve problems, which are not easy to solve using simple mathematics and simple reasoning. So as the number of these parameters has grown, so has been the ability to see patterns, which are sophisticated and complicated, all the way to the point that I believe the systems are now beginning to understand the patterns of how humans think, which is ultimately what generalized intelligence is going to be sometime in the future. But how points write poetry, how artists actually create art, how software programmers create software, it's not just what they create, but how they create. Those kinds of deep seated patterns have now become within the realm of reach of these technologies. So that's the valve factor as well. People are saying, wow, I could not imagine you can ask the system to write a poetry in a certain genre or a certain language in such a short period of time. And that's primarily because these systems have become so much more powerful. And I think that that idea of these machines understanding how human thought processes work, that's both the valve factor and the thing that makes everybody's skin crawl a little bit, right? Things that we go, oh, I don't know how I feel about a machine understanding how I think and how I make decisions. But of course, that is the power. It's that double edged blade, right? Yes. So you mentioned that there are some challenges with so many more people having access to these tools. Let's unpack that a little bit. What are some of the things that you're seeing, what are those challenges that people are facing with that? Yeah. So first is the ability to understand the way humans think. That's not quite perfect at all. In fact, far from it. I think we're just starting seeing the early signs of it. These systems are still at some level are imitating information patterns that they've been trained on. And they're not yet generating completely new patterns, at least in my opinion. And there are lots of imperfections in terms of the patterns that they measure as well. So just because a machine suggests that a certain type of pattern doesn't mean the pattern is true or real. So there are lots of, which is the famous hallucination issue that people talk about with these large models, is that some of these answers and responses look completely made up, sometimes they are. And hence, they're not accurate for certain types of business applications. So there's a lot of accuracy challenges. I would say that's probably one of the biggest challenges in these systems is how do you control and understand what is accurate. And then there are other issues from a business perspective, of course, which will include things like data privacy. And we've heard about people inadvertently leaking corporate data, not knowing that these systems are able to do that. Also, we talked a little bit earlier about intellectual property rights. And there are open questions about who owns the intellectual property. If the system was trained on proprietary knowledge and it generates new knowledge or new information, who owns that new information. So these are human questions that we have to also understand before commerce can actually flourish around these technologies. But I would say those are the major issues about accuracy and data and probably a bit around ownership of information. Those are the big sticking points. I think the only one that I might add to that is also that question about the data sets that these things are getting trained on, right? And we've seen those examples of the Twitter AI that became racist really quickly, because it was being trained on-- Yes. --the whole of what was out there on Twitter already. And so I'm really interested in how can we start to train these models in a way that helps us ideally work some of those things out of the system and does try to create a more objective language model and things like that. I don't know if you've seen anything going on in that space, but I think that that's a big challenge too. No, I've seen-- you're absolutely right. I think like all technology, all new technology there, people who will try to exploit unethical uses with that technology. And I'm sure, generative is no exception. And as the technology becomes more powerful, policing it and controlling it probably requires more powerful tools as well. As technologists, again, from a innovator's perspective, what I'm excited to see, there are-- so for example, even in our accelerator recently, we have one or two startups, which are primarily focusing on predicting on being able to measure which content is generated versus original human created. And so you'll see technology-- it's a cat and mouse game technology, being the police off the technology itself. This has been going on in AI for a while. It's not new. AI solutions and systems for doing what we call AI monitoring or AI explainability or AI predictability are a very important area of research and development for the last several years. And I think we will continue to see a lot of that. But of course, like many things, the solution, the not just technological require maybe a combination of us coming together as a society to figure out ways to control it and putting some processes and checks and balances and how we can not fall prey to every piece of information that we get. Today, even now on internet, people ask me, how you're going to control all this misinformation that could be generated rapidly. And one of my sponsors is, even today on internet, there's a lot of information which is not real. And we can search and find all kinds of information. If you keep digging deeper in your search list, you will find information. But as humans, we've learned to trust and not trust. I think all that experience is what we will develop on the next several years. And how do you put those kinds of experience, guardrails, and technology guardrails around generated data? Well, I'm very glad to hear that there are companies that are working on these monitoring systems and tools to manage the tools. I've been thinking about this a lot, of course, with my background in community management, which obviously there's a heavy bent of that, which is content moderation. One of the questions that's come up for me is, what is it going to mean if I'm in an online community where we can't tell what content is generated by a human and what is generated by AI. And so I've been like, I need water marks on that content or some kind of tag so I can tell, I also need moderation tools that can help me identify what is legitimate and what maybe is-- I mean, we started with that today with spam bots. And I can only imagine that getting so much more complex. And so I think it will be really interesting to see how those kinds of regulatory and moderating and the learning tools come into play in a whole variety of spaces. On the flip side, always bringing it back to the positive, I think that there's some real opportunities for generative AI to help conversation. For example, I see work-lobe company and I see customers who may be English isn't their first language. Sometimes struggles who engage in those conversations as effectively as they could if they could write a little prompt and say, OK, chat GPT expand on this or help me describe this idea to people who are native English speakers. And so I think that there's actually a real opportunity for increased engagement that is facilitated through some of those tools in AI. So it's not all bad for sure. Absolutely. I would say there are far more positives and interesting areas than probably people have yet imagined. And for example, just in the communication side, I think what generative technology is especially large language models for text. In my opinion, other big things they do is they help create a revolution of sorts in communication, the depth of communication and engagement add on that support system. So like, for example, if we have a call, a lot of the time we spend is just making notes or trying to remember what we talked about. And that distracts from the conversation itself. Also, sometimes we come prepared for conversations and we are listening to a few things. But the other person may have some other areas of interest and he or she may express. And we may completely miss that because of our limited ability to focus on a few things at a time or make notes and remember a few things at a time. I think the fascinating part about that, in my opinion, is take a customer conversation or take a human to human conversation, for example. These types of support systems can help enrich and unburden us from a lot of the things that we probably end up spending time instead of just focusing on the human interaction. Of course, if we have to be disciplined about how we use these tools, but I can certainly, for example, a startup that I recently met is solving the problem. I think you alluded to earlier, they look at your thousands and thousands of social media feedback for product users and comments on the internet just to analyze patterns on what is the next set of features that the product manager should provide or if they have changed a feature which is being appreciated the market or not really going down the same way they plan. And those types of signals would require, I mean, so long just to read those comments and frankly, most of the information is often not reviewed and analyzed. Now you can imagine the job of these product managers, even the social media managers, significantly more fun and being able to find patterns instead of just reading documents and putting through outlets hours and this. So yeah, I think you're right. If they think very many new innovative opportunities and more fun opportunities are likely to emerge as well. It is so important to remember that these technologies are going to enable a lot of us to hopefully get rid of or minimize the mundane parts of the job and really dig into the parts that are more meaningful. What do you think is unique to the startups that are developing as compared to some of the big companies? Because I think everybody's heard about some of these larger organizations are an element of this stuff, but what is unique to the startups base and what capabilities and opportunities do they have? The big difference in startups and the large companies is a of course speed, but also risk-taking ability bigger company have a lot more fiduciary responsibility around data. Not that startups are reckless, but they have less worries about losing data because many of them don't have proprietary data to begin with. They are just trying to start a new process and new application. They also much more disruptive in their thinking about changing ways so the startups are faster to come up with ideas and experiment ideas, which probably larger companies will evaluate a little bit more longer and assess. Although I would say that the pace of innovation with generative AI in general has really accelerated everywhere. It's not just the startups. I think that's one of the challenges even startups have is even bigger companies are able to move almost as fast if they're focused in catching up with any areas. But what I'm seeing in terms of startups is very many creative experiments on very point specific problems and sharp specific areas. A lot of the old SaaS product models are being experimented with very rapidly. So if you had an email tool or a chat tool or a video calling tool or even in types of a business application, you had something to manage your finances or accounting or HR communications. A lot of startups are just jumping in and saying can I make that better by adding generative AI to it and I at task so much faster and more fun or much more accurate or exciting. So I think we're seeing a flurry of startups were jumping in trying to take all of the SaaS type products that we had built in the last 10 to 20 years even longer and trying to reimagine them with generative AI applications. That's kind of one of the big things and then in the creative world, I'm seeing a lot of new startups coming out in various different areas of creative engagement, everything. Like we mentioned movie creation is just one but gaming is another very exciting area in a lot of startups. The amount of effort required to build these games, especially with different types of avatars and textures and context. It simplifies the task for individual game developers to build these types of advanced capabilities much faster. So we've seen a lot of focus going on there as well. So yeah, I would say in general is just faster experimentation and areas which have not yet been explored. I think startups are really swarming into those those areas. Well, it's going to be so exciting to see what they come up with and now you got me motivated. I want to talk to somebody in the gaming space and hear all about the development going on over there. All right, last question. What has been the most fun or interesting thing that you got into play with in the AI space so far? If you can pick one. I've seen many exciting and interesting entertainment applications which as a consumer I like I enjoy but one thing I feel is quite mind bending in my opinion is how we as users of technology is likely to fundamentally change in the sense that we used to have to be trained on using different software tools. And we had to learn how a particular software tool worked in order to become an expert on the tool and hence perhaps on the job or that kind of functional role inside a company. So even if I'm a creator, I have to learn maybe a design tool and it is a lot of work required and learning curves. I think that's being flipped around. We can have very simple chat interfaces with much like a search or a chatbot and we can just ask the chatbot to give us the answers or create whatever we want. We don't have to worry about going after menus and menus of options and learning the steps of running a particular software. How do you write a SQL query to generate a graph from a database? How do you you know produce software by writing learning Python and Java separately? You can do a lot of that just by simple English as an interface. So the power of human productivity and the ability to engage with technology is getting transformed. And to me, it's just probably a singularity moment in how I think humans have used technology that's likely to be something that we look back and say our user experiences and interfaces are so primitive going back a few years. I wish I had started with that question. There's so much in there. It is so cool to hear that because I know how many times I've looked at technology and been like, I don't want to learn this whole interface. I wish I could just enter into a chatbot or a search widget. Here's what I'm trying to do. I'm going to make that function go and it sounds like that is going to be become increasingly the way that we're able to interact with this technology and hopefully that will accelerate adoption rates of different kinds of software to you because people don't have to go through all of this. Absolutely. Extended learning to be able to utilize the really powerful tools that come to me. Absolutely. And think about all those non technical folks like doctors who like my dad was a doctor. He only used the minimum amount he can get away with. And we have we subject our physicians to writing notes and looking up medical records before they can see patients and they're being forced to learn tools that they don't really want to learn. And I'm sure that's the case with many, many areas of work and how much fun it is that technology disappears behind the chatbot. I think that's probably where we had it. I think that's ultimately what I feel is the communication revolution or the discussion. And the way we communicate with technology and communicate with each other is going to get so much simplified that we are not going to be slaves to a particular product feature or technology said, but we will have more control in a much more natural communication process that we're used to. Well, that shines a very bright light on the future. So I think that we should leave it there and end on a positive note. Thank you so much for joining me today. I really enjoyed this conversation and look forward to picking your brain more in the future and hearing about where things are at a year from now. Yes, I could tell you more, but I'm just excited to see where things will be in a year from now. And thanks again for having me. Appreciate it. I am so excited to hear what some of these startups that Deepam was talking about end up developing. It is such an exciting time to be innovating in the generative AI space. Next time we'll be talking with Ian Hunt, director of customer service, simple procurement for Liberty London, which is one of the oldest luxury retail stores in London. He'll talk with us about the evolution of going from being a brick and mortar retailer to moving into a combined physical and digital space. Should be some stories that we can all relate to as our worlds have moved ever more online. And there'll be some great takeaways about how to approach that evolution as a business. Until then, please be sure to like and subscribe to the podcast and tell a friend if you're enjoying it. If you want to connect with other Zendesk users, you can head to usergroups.zendesk.com to see about all of our upcoming user group events, community webinars, and more. I hope to see you around the Zendesk community soon. Until next time, I'm Nicole Saunders, for Zendesk, the Impeologant Heart of Customer Experience. Ready to explore how AI can help your business provide richer customer experiences? Zendesk enables you to instantly harness the power of AI to deliver scalable world-class customer experiences. In the era of Intelligence CX, Zendesk is here to help you find your way. Sign up for our free trial at Zendesk.com.