Member of the Management Board and Chief Product & Technology Officer (CPTO), TeamViewer SE
Unlocking efficiency with AI - Panel @HumanX 2025 in Las Vegas with TeamViewer's Mei Dent
AI is transforming how organizations collaborate, mine processes, and improve customer satisfaction. How can connected ...
Transcript (45 segments)
M
Moderator0:10
Good afternoon everybody. I'm too to actually sit in the chair. So today we're going to be talking about how AI can help businesses become more efficient and help people collaborate better. I think each of the people on the stage is coming at that from a different and very interesting angle. So I guess first of all I wanted to start by going around and having each of them explain what that term from the session title, connected intelligence, means in each of their businesses and each of their contexts. So Robin, why don't we start with you?
R
Robin Daniels0:42
Sure. Hi, I'm Robin Daniels. I am the chief business officer at Sensei. We are an HR tech platform that's all about human success. And the way we think about connected intelligence is that when it comes down to it, an organization is just a group of people trying to do great things together. Sometimes those things are small and solve small challenges. Sometimes those are really big. What we're trying to do is by having a platform that can ease the teamwork, the collaboration, the communication between team members, we can actually get much more efficient in an enterprise because those are usually the issues that stop companies from doing great work. It's like those little interpersonal things, miscommunications, misunderstanding of goals, misunderstanding of where you're going. So we think of it as can we take the friction out of how teams and individuals collaborate together to do great work.
M
Mei Dent1:33
How many of you already know TeamViewer and what it does? Great, thank you. I really like the title because connectivity is the heart of what TeamViewer does. What we have really been working on for 20 years in business is working with our customers on this digital transformation information journey, right from connectivity, I think that's at the core, to workflows to auto co-pilot and now to autopilot all the way to AI. So I'm really looking forward to this. It's a graveyard shift with lunch being served everywhere around you. Thank you very much for making it. I hope you find some insight into the use cases and how we work with our customers and the return on investment that we're able to deliver.
G
Greg Fallon2:17
Yeah, thanks. So my name is Greg Fallon. I'm the CEO of Geminis. We're coming at this panel a little bit different from my peers. We deal with large cyber physical infrastructure, things like power plants, oil refineries, and we think about connectivity in multiple ways. One is between the people and the machines. That's where we spend a lot of time. How do you help these human beings make these machines that are infinitely complicated run more efficiently or more productively? But then also the people that run these machines also work in silos. These are teams of physical engineers, people that deal with engineering in the physical world. They have multiple computational systems that are connected to each other. They have individual disciplines, and bringing those together is a huge factor in trying to make the overall industrial plant more efficient.
M
Moderator3:17
Great.
D
Dan Brown3:17
Um, I'm Dan Brown, not the author, and I'm from Salonus. We provide a process intelligence tool. What that means is most companies have a number of systems. They do things, they store data, but almost always they don't really know how the company operates, how business processes work, how their interdependency happens. Our technology allows you to infer that and allows you to infer what systems are doing the work, how information flows, who's working on stuff, how multiple processes interact. So when we think about connectivity, we think about how those business processes come together to deliver value. You can think about an example. If there's a customer order that is delayed, you want to ask the question why. There are a lot of connections inside your company and there's a lot of information that is outside that come together to explain that and then ultimately to help that person who's picking up the phone or an agent that is responding go and help out that customer, maybe reroute, maybe explain the situation, maybe provide an alternative.
M
Moderator4:31
And then just to further set the stage a little bit here because I know these terms can kind of get mixed up, how much of what each of you do would you call traditional machine learning and how much is generative AI? How are those two things working together? This time Dan, we'll start with you.
D
Dan Brown4:50
Okay. So traditional machine learning, it feels funny saying that already, and generative AI. The answer is both. I would be surprised if most technology companies have not already done quite a bit in machine learning but are using Gen AI to augment that strategy. We do both. We use traditional machine learning to provide insights on your process intelligence, understand the distribution of variation, understand predictions, throughput time, and so on. But we also use generative AI to put in the hands of human beings how things are happening, what might be a good way of responding. The other thing that I think is really interesting and we're seeing this in agents is those worlds are really coming together. Transformers, for example, have traditionally been for large language models, but they're actually being used for different types of tokens, for example activities to infer when things will happen. I think that's going to continue to happen with large language and foundation models being intermixed or used for different use cases outside of just NLP interaction.
M
Moderator6:08
And Greg.
G
Greg Fallon6:12
When we talk about the mix of generative AI and machine learning, we do a lot of what I would call uncommon machine learning. Most of the work that we do, because we're dealing with large machines, requires very high precision. So what we think of as generative AI in terms of LLMs is something that we make use of but more as an agentive approach to enhance the human beings. But the real math gets done by a combination of machine learning algorithms and optimization routines which on their own are generative because they're actually creating control recommendations to these human beings. So I'd say it's a mix. In the hardcore industrial world where precision reigns, we're going to be a while before we can see generative AI in terms of LLM type technology emerge. But in the meantime, I think it offers loads of opportunities from an agent perspective.
M
Mei Dent7:12
I guess I'm switching to this. Hopefully you can hear me. I would say they are both useful. They're tools, right? I don't think you say this one wins over the other. It really depends on use cases. For us as well, we have a lot of traditional machine learning with structured data, semi-structured data. We use the more traditional techniques that are offering that AI capability, and that's been going on for the last 10 years at least. The recent adoption of the large language model is first and foremost to analyze the multimodal text that we deal with. In a typical TeamViewer connection, it could be IT, so it's chatty in nature. But when we go into a frontline support case, it becomes all of a sudden multimodal. It could be audio call, video, CAD images. All of these things come, and the large language model is just so much more capable to digest the information in large quantity and also the variety. So right now, there are different use cases focused on either the traditional semi-structured data versus the dialogue and interfaces. I think of it simply as machine learning is a fancy word for pattern recognition and pattern matching. It's like taking all these different data points, looking at them, and then coming up with an answer. True AI is really pattern recognition plus creative thinking, where you add that layer of creative thinking to understand the sentiment behind something and come up with something you hadn't expected would be the answer. For example, if you have an employee, many of us have been in this panel, and employees come and say I want to go this direction in my career. You use your pattern recognition to say based on what you're good at, you should go here. But that's different than if somebody comes to you and says I feel like I could learn analytics because it would take my career to the next level. Or if they use the word, I should learn analytics. Those are two very different things, and you need the creative thinking of somebody who understands the nuance between those different things to come up with a better answer and solution. One is just a fancy word for pattern recognition, and the other is that creative spark that gets added on top of it.
M
Moderator9:44
Got it. When you're designing AI tools for efficiency or collaboration, by nature you're going to have to design them for a lot of different roles. It has to be usable by HR departments or engineers or maybe systems of record. I'm curious how you think about making a tool that every different department or every different role can use. Is that more challenging? And May, why don't we start with you this time?
M
Mei Dent10:09
AI still has to solve a business problem at the end. Really, yes. No matter how fancy it gets and how much hype it gets, at the end of the day it's about return on investment and specific use cases. So for us, it's working with the customers to identify that. I can give an example. In the typical IT TeamViewer use case, we would measure on first ticket resolution, measure how much time you save in reducing ticket documentation, how much time you capture the knowledge so you can train the next new agent. Or when you go to an OT setting, a more production setting, it's a very different set of return on investment discussions and efficiency. In there, we have a picking solution using AR/VR solutions, and error rate is really important to reduce. The cost of wrong things shipped out is very high. So you define error rate and with labor shortage after COVID to get warehouse workers to come in and be productive right away, you define how many hours it takes to train them. Is it intuitive enough so they can just put on the AR glasses and go? These are different use cases, but it boils down to the same AI technology behind it.
M
Moderator11:45
Um, Greg, you want me to talk? Okay.
G
Greg Fallon11:50
We think about the business problem and efficiency. In our world, there's one thing that's really clear. We're all about increasing production, right? So it's relatively easy to measure. If you use AI to improve production of a power plant, you should be able to get more power out of the plant by spending less money or less carbon emissions. That's one way we think about efficiency and measurement. The intangibles are harder to measure. How do you measure what happens when you break down silos? You change a process, you get people working together. AI bridges a gap. Are you looking at hiring fewer workers? In most cases we've seen, no. In fact, a lot of what we're doing is seeing increased productivity from workers that used to be able to scale to like two people. These are experts and now they can scale to thousands of people. We haven't figured out how to measure that yet. It's very difficult. It's the type of thing where our customers see it when they believe it. Thank God for us, we have the industrial output that we can kind of pin a number on.
M
Moderator13:07
And Dan, what are you looking at to measure the efficiency that you're creating?
D
Dan Brown13:14
You had asked a question about designing, and I'll make that go into this because it's really important. When you're using any type of artificial intelligence, the semantics of what you feed it matter a great deal. If you leave out certain signal, it's obviously not going to make a prediction or classification based on that information because it doesn't have it. We talk a lot about process intelligence as what fuels artificial intelligence. When you understand what your company does, you have the semantics of what it actually does, then you can start to answer questions very well. For example, why? That's what everybody wants to know. When they know why, then you can go change it or allow an agent to go change it. Now let me fast forward to how you actually measure things. Those semantics can be very broad. Most processes are measured on throughput time, efficiency, volume, and so on, or they can be very narrow. If you're working in revenue marketing and you want to know how many leads are dropped, that's a very narrow semantic. You can do things that are broad and things that are narrow. By way of examples, in a typical process on the purchase order side, you can measure how many times they need to be handled by a human being. Once you benchmark and identify the variations where that happens, you can eliminate those. We have a really good example of a customer that did that and their sales order throughput grew by a factor of five by reducing unnecessary human touch. That's immediately value creative and absolutely economically measurable.
M
Moderator15:12
And Robin, do you want to add anything here?
R
Robin Daniels15:14
Yeah, sure. Taking the human angle again, when I think about my time at Salesforce or Box or LinkedIn, over 90% of my coaching or interactions with people have been on the soft skills, not so much the hard skills. It's usually those soft skills that get in the way of efficiency, great teamwork, great cultures. We think about how do we use AI in some ways to coach people to be more empathetic, more compassionate, use the right language, use the right goals, use the right motivation so you can get people to work well together towards a common outcome. What does that mean? It sounds very lofty. So what we think about is can we give people the nudges and insights on how their interactions are going to push them in a certain direction? Use these words instead of this word, learn this skill instead of this skill, all in the service of reaching business outcomes, reaching the goals you've set. If you have the goal of getting from 5 million to 10 million in revenue or entering a new market, all of this has to be in the service of better business outcomes. The AI we're putting into our platform and in our industry is about how do you push the humans and teams to break down that friction and work really well together. What we think about is that there's been a lack of visibility and insight into how do you measure that, how do you correlate that. If you think about what chief customer officers care about, NRR; sales leaders care about ARR; finance leaders care about cash or EBITDA. But when it comes to how you actually measure how people are doing, are they working well together, there's not really any way to measure it. So I think we as a company and as an industry are really focused on how we use AI to give us that visibility, that knowledge into how we're doing so we can progress towards the goals we've set and take the friction out of getting there in a very seamless way.
M
Moderator17:23
Got it. I wanted to switch gears a little bit and talk about the adoption curve of AI. Since ChatGPT kind of put generative AI front and center in 2022, are businesses on the whole operating much more efficiently than they were then, or how far into this adoption curve are we? What are you guys seeing? Whoever wants to start can start.
R
Robin Daniels17:47
Well, I saw some recent stats that were quite eye opening. 86% of companies and workers feel like it's making them more productive, yet over 70% of people also feel like it's giving them even more work because they have to check the outputs of all the stuff happening behind the scenes. So it's like one side yes, the other side yes. And then another stat that was mind blowing: over 70% of people say it's leading to increased anxiety and burnout because they feel they have to do so much more. The pressure is there. So my take is that there are a lot of conflicting numbers. We're still in the early days of figuring out how this is going to net out. I think we're all optimists up here. We think AI is going to be a superpower if used in the right way. But it's still early days, really figuring out how to harness this. It's pockets: marketing is doing this, finance, sales, customer support. But is it cohesive, centralized, do we have one full purview? We're getting there, but we're not quite there yet.
M
Mei Dent18:48
I don't think I agree. In terms of adoption, different industries, different use cases will see it differently depending on how threatened people feel. Some studies also say that executives feel more excited about it than the actual people working on it. For us, knowledge workers, there's a big divide between knowledge workers and the deskless jobs, the frontline workers. 80% of the world's positions are actually in deskless jobs. We're lucky to be knowledge workers, but we only represent about 20% of the workforce, and 60% of that happens in Asia-Pacific regions. When you think of the scale of potential impact from digital transformation and being able to take large quantities of data, working in large factory settings, it's early in its adoption. We still have to come through quite a bit of change management on processes, on skills, and on the technology to be able to get there.
D
Dan Brown19:50
I'll agree with you as well. Not only are we early, we are super early. We're talking about adoption. Maybe most employees are using generative AI in some form in their daily workflows, but what are the use cases? Use case discovery is so early. We don't know whether we're proliferating a million applications or they're going to consolidate, whether companies are going to make their own applications or buy them from someone else. It's exciting, and we are in the hype curve right now, probably still early in the hype curve. At some point, people are going to stop talking about AI, and at that point, we'll know it's actually starting to catch on.
G
Greg Fallon20:39
I'll pick up on that last point and add something else. A lot of times, customers are already using AI and they don't know it. When you're writing something and your document fills in something for you, that's a great point. That's when we know we've arrived, when it just feels seamless. One thing we've noticed: we tend to focus on large enterprise companies. There is a very big appetite to explore use cases and adopt technology. Everybody's talking about Gen AI, agents. A lot of people are not really sure how to apply this. I completely agree, it's like what's the use case? What's the outcome we can get? I love that it doesn't matter whether you're a manufacturing company, agri, medtech, or public care. Everyone is interested in this technology and how it can improve what they're doing, improve outcomes, and fit the business for customers. I'll give an example of how we've seen this play out. We have a company that was interested in why blocks were happening on order processing and how they could expedite those. This seems basic, but if you can get rid of that problem, you can unplug a bottleneck. They understood the root cause by getting process intelligence information and then using an LLM-based assistant to summarize the problem and provide alternatives for action. This is a human-in-the-loop agent that shows up in a collab tool and helps that individual make a choice and make a change. At some point, that will move to something automated. The amount of human-in-the-loop will go down, freeing up that individual to do higher touch things. The evolution will be: we're eager, we want you to try. Every company's board is asking, 'What's your Gen AI story? How much are you going to save?' We all know this is happening. Everybody has the pressure. The curiosity is there. I think we're going to go from playing around to human-in-the-loop to fully automated.
R
Robin Daniels23:35
I have a confession to make. My name is Robin Daniels and I suffer from AI FOMO. It's real. When you look online and read articles, you just feel like you're constantly behind on AI. It's mind blowing, especially on LinkedIn and X. Last week I was at an AI go-to-market conference with about 150 go-to-market leaders from elite tech companies, and they're all feeling this way. Everyone feels behind. We went around the room, everyone lifted their hand and said where they are on a scale of 1 to 10, and everyone gave themselves a two, three, maybe a four. Everyone feels a little behind on the adoption curve. You see pockets of it, but there's not really any organizing strategy yet. So stay curious, stay hungry, keep looking at it. That's the takeaway. I'll make a quote here that I think is super relevant. Benioff said, 'People dramatically overestimate what you can do in a year and dramatically underestimate what you can do in 10.' I think that is absolutely the case here.
M
Moderator24:46
How do you deal with that sense of FOMO and that pressure from above? How do you make sure you're focused on things that actually matter?
R
Robin Daniels24:53
Just meditate, breathe, chill a little bit, because it's definitely real. You feel like you're always behind. You just have to stay curious, consistently lean into finding new tools. At our company, we've created an AI council and we have a go-to person for helping us across the company get much better at using AI, because we certainly have it in pockets. I also think you're going to get a lot of shadow AI popping up everywhere, just like Shadow IT many years ago. So having an organizing principle, somebody who has the full purview across all your different business processes, is what we're in the early stages of doing. I'm seeing more thinking about having an AI project manager or organizer who's looking across versus all these different groups doing their own thing.
M
Mei Dent25:48
The best way is just roll up your sleeves and do it. Maybe it's a very small step in the beginning, but the moment you feel like you are doing something, you feel more in control. Then the next incremental development won't be as hard. Creating a safe space for employees to try, safe for them to feel it's okay to fail, to do something and it doesn't quite work, to have that framework and that culture is important. Otherwise, I don't think this will ever work top-down. Even though top-down right now is asking a lot of questions, what you really need is for people to feel empowered, to not even think about whether they're using AI or not before it will be adopted.
D
Dan Brown26:36
I love that. The idea of empowering everybody. There is no single human being that is possibly able to stay on top of this. This is a perfect application for Gen AI to tell us what we should be doing at any given moment to stay on top of AI. But I'm seeing in more and more organizations there will be a centralized approach for governance, but also people in every group who have an inherent interest and are staying on top of it, influencing their colleagues. That's what gets me excited, because those little pods of human beings are coming up with new use cases and proliferating information. As an introvert, I have to spend more time talking to other people, even though I love the agent. There's a product for that. I have a team member here, Reagan. I created a digital clone of myself. I took all the content from my writing, podcasts, videos, put it into a digital clone. Now people can talk to me and ask questions, and I come back with 90% plus accurate answers of what I would say. You can scale yourself.
M
Moderator28:02
All right, I love it. So May, you mentioned earlier that executives might be a little more excited about AI than some workers below them. I'm curious, you work a lot on upskilling, Robin. Is that something you find as well, and how do you deal with that?
R
Robin Daniels28:25
I think enthusiasm is high at the top and high at the very junior employees. I think it's the middle that's getting squeezed. If you're new into the workforce, you're kind of AI native in many ways. They're naturals with AI tools, very impressive to see. There's a lot of enthusiasm at the top because the board wants it, shareholders want it. Everyone wants to get more efficient, better at predicting outcomes. But it's the middle that's feeling the pressure: how do I use it? Middle managers are probably getting pressure to get more efficient, do more with less. They're not quite sure how to do that. So we have to empower and enable them, give them confidence and tools, let them know it's early days, that we trust them to figure it out, but also push them in a certain direction to get more efficient and learn the right skills.
D
Dan Brown29:22
One of the tried and true methods of helping a company out is to make the stakeholders successful. Once you make the stakeholders successful, and middle managers are often the ones who can talk to the people putting pressure on them, just to connect the two things. Getting started seems hackneyed, but we have an AI lab. It's a product and go-to-market joint effort. We go out and do what we used to call joint application development, working with customers to identify where AI can be useful and try it out. Ultimately get it out in the wild. You can put numbers on a spreadsheet, prepare a presentation, do a board deck, but when you have legitimate value-creating use cases that a customer is doing, that's the magic. I look around, and I don't know half these companies. What I do know is we're working very closely with customers with whatever technology we have and trying to create value every day. That is exciting. I think that's also what employees like. At the end of the day, employees want to see what they develop being adopted. A developer's best accomplishment is a million users of a feature. When you put that perspective, they sometimes feel the burden of needing to define the perfect thing to fulfill expectations, but everybody is really just experimenting, figuring out which one sticks. So start, don't be afraid of taking whatever that is, co-innovation, lab to the customer, get validation. I bet if that customer validates this idea, another idea will come next week. Employees will upskill themselves rather than spending three weeks thinking about getting the perfect thing. That's probably not going to happen.
M
Moderator31:47
Greg, you probably have a unique perspective here working with industrial giants that maybe haven't thought about AI as much. How do you think about upskilling there?
G
Greg Fallon31:59
Oddly enough, they're hyperfocused on it. What I'm seeing in customers is a huge push on AI, actually in industries I wouldn't expect. Some of the more connected, sophisticated industries I'm not seeing do this as much, but it could be because they're more mature and it's just not visible. A lot of what we would consider dirty tech, people with very old assets and machines, are really pushing AI in big levels. They're trying to figure it out and moving forward. It's very hard for these companies that tend to be extremely profit-focused to make a decision to do this. The better ones have carved out AI teams given permission to experiment and develop applications. We try to come in and show hard numbers. The margin is so thin in traditional manufacturing, and they are looking for that edge. They probably don't think of AI as AI; they think of that margin and the goals they have. We have a Swiss-based customer that ships their machine to Asia where most manufacturing happens. They use remote connectivity technology without traveling at all. 80% of the questions are about how to operate the machine or issues. Production downtime is huge for these manufacturers, so they have a strict SLA. This particular OEM vendor needs to stick with that SLA. How do I stick with the SLA without traveling? Traveling from Switzerland to Malaysia costs time. How do I reduce downtime? They're looking everywhere for AI, machine learning, AR, VR technology, whatever works. As vendors, we need to offer a very concrete demonstration of return on investment for them.
M
Moderator34:19
We've touched on agents a couple times, but it's something a lot of people are excited about and have different definitions of. I'm curious how each of you are thinking about agents and the future of agents. Dan, why don't we start with you?
D
Dan Brown34:34
I sort of mentioned this earlier where agents are going to be human-in-the-loop. That's what we're seeing today, where an agent will inform a human to take action, and the next logical step is for an agent to do that work. We're starting to see interest in agents that orchestrate multiple agents. That is a bidirectional thing. An agent can be orchestrating or told what to do by another agent. I think that's something we'll see. There is immense investment in this to make agents more intelligent. For example, if you want an agent to reroute a shipment because it's delayed, that agent probably needs to understand something like the weather. When you have domain-specific LLMs that can provide that information, you're empowering an agent to do much more sophisticated things. The other thing people are interested in is: great, this agent is going to do work for us, but is it effective? Did it actually create value? You have to feed that back and measure the outcome. People are a little scared of just handing over the wheels to an agent. How do you do that? How do you test it? What if there's an exception? What if it's wrong? Who's responsible? Those things are very evident in orchestration and agentic AI. That's the cutting edge of what people are trying out.
G
Greg Fallon36:31
The concept I'm most excited about is agents controlling agents. This is one of the trends I think is going to take off. Each agent needs to be specialized rather than having a monolithic model. It's more efficient to have them connect, but it becomes incredibly complex. You have this systems engineering problem of managing them and making sure they talk to each other appropriately. From an agent management perspective, it's much easier to disaggregate the problem and look at them individually: Are they doing what you asked? Are they getting better? Is there a fault? This idea of quality control and possibly retraining when needed is important.
M
Mei Dent37:23
I have a slightly different take on agents. We spend a lot of time on automation. For me, automation is not autonomous; agent is more autonomous. An agent not only can follow a prescribed set of complex things efficiently, it's also about discovering what problem you need to solve and applying whatever it takes, all in the context of human-machine interaction. I love the bot of bots, the super agents concept. Going forward, it's also about seamless interaction between human and machines. Instead of us versus them, it's more using the end to be thinking more like us, helping us take out some of the mundaneness. It's exciting.
R
Robin Daniels38:26
I think in the future, all of us will have an army of agents working on our behalf. It's not going to replace everything, but it's certainly going to get us great answers and new thinking. We're already seeing it. I have someone on my team who runs product marketing. For the longest time, we talked about hiring somebody to do competitive intelligence, which is very hard to do. Keeping on top of competitors, figuring out your vectors, it's super complex and requires 24/7 focus. Now with AI, you can do this really well. You can create an agent that goes out and says, 'They've changed their price list, new feature set, customers saying this on G2,' and gives you information to action. You don't need to hire that person anymore. There's still work involved in checking, verifying, but it gets you a big part of the way. Agents will do this all the time for ICP, customers, fundraising, all kinds of things.
M
Moderator39:32
Great. We just have a few minutes left. Maybe we can end on one big prediction or closing thought about an upcoming trend in the next year impacting each of your areas. Whoever wants to start.
M
Mei Dent39:48
I can go. We see every workplace becoming a digital workplace. With knowledge workers, it's already digitized the majority of the way. But 80% of the workforce is actually in deskless positions, and they are quickly becoming a digital workplace as well. So providing more mature solutions, thinking out of the box about how that can be applied to deskless frontline workers. That is the challenge we're working with. Hopefully many innovators at this conference are also thinking beyond knowledge workers, about how we can get the rest of the workforce there too.
D
Dan Brown40:45
One thing I think will become more prominent is sequential decision making. Every enterprise, every day, there are choices made. One problem in traditional machine learning is the world is not stable. What a model is trained on today is not what the world looks like tomorrow. How do agents or machine learning, with the semantics from process intelligence, facilitate decision after decision? Whether that dynamic behavior is injected into an agent orchestrating other agents, or human-in-the-loop, it doesn't matter. I don't know if that's in the next year, but it's something we think about. The second is, adopting AI in your company is a fundamental business process change. How do you instrument that? If you want your NPS score to go up five points, what are you doing differently? How do you get the lead indicators? If it doesn't happen, how do you answer why? It's process re-engineering. Are agents going to help with that? I don't know. But that's ultimately where intelligence should be focused.
G
Greg Fallon42:20
I'll play on something you just said. I think AI ops is going to become really important. We have MLOps, model updating. AI ops is a major thing. It means so many things. We were talking about human resources changing, having to oversee agents. This whole idea of AI ops and how it builds into ethics, testing, validation, deployment, I think is going to be massive.
R
Robin Daniels42:53
We're still probably 3 to 5 years out from seeing companies being fully AI from end to end. But in the next year, every single knowledge worker employee must have a killer use case where they're using AI, or they probably won't last long in their job because AI is coming. You have to embrace it.
M
Moderator43:14
All right. With that, I want to thank you all for this great conversation. Thank you.