Qualtrics on Closing the Experience Gap in the AI Era

For its first two decades, Qualtrics focused on closing the insight gap: helping organizations understand why customers and employees felt the way they did. CEO Jason Maynard says that problem is largely solved. The next challenge he's solving for is the distance between what a business intends to deliver and what people actually experience.

Maynard believes AI can close that gap by moving companies from manual analysis to simulation: model an outcome, predict the response, and act on a trusted result.

Daniel Newman welcomed Maynard to open the Enterprise AI Software and Agents Track at The Six Five Summit: AI Unleashed 2026, recorded at Qualtrics headquarters in Plano, Texas.

At the center of Maynard’s argument is a different view of enterprise intellectual property. A brand’s most valuable asset is not its model, but the accumulated history of its customer and employee relationships. He expects enterprises to protect that data rather than hand it to external LLM providers.

Qualtrics operationalizes that idea by combining probabilistic reasoning models with deterministic business rules in what Maynard calls an “experience ontology.” The goal is to ensure AI understands how a company actually operates, including its policies, relationships, and constraints, instead of producing the generic, unreliable output he describes as “AI slop.”

That ontology draws on more than 20 years of first-party Qualtrics experience data, now expanded through the acquisition of Press Ganey, which Maynard says gives the company access to the largest patient experience dataset available. He compares the resulting simulation layer to Formula 1 teams testing race strategies before committing them on the track: model the decision, evaluate it against a digital twin, and act with greater confidence in the outcome.

Maynard applies an equally practical standard to AI ROI. The goal is not to maximize token consumption, but to increase customer lifetime value. His clearest example is in-the-moment service recovery: identifying and resolving a problem before a customer calls for help or quietly leaves.

Key Insights:

🔹 The insight gap is giving way to the experience gap. Qualtrics has spent more than two decades helping organizations understand why something happened. Maynard says the larger opportunity now is aligning the experience a business intends to deliver with the one customers and employees actually receive.

🔹 A company’s relationships—not its model—are its real intellectual property. The accumulated history of customer and employee interactions is what differentiates a brand, giving enterprises a strong incentive to keep that data within their own environments.

🔹 Trusted AI requires both reasoning and rules. Qualtrics combines probabilistic models with deterministic business logic in an experience ontology designed to ground AI in how an organization actually operates and reduce unreliable output.

🔹 Simulation turns insight into action. Using more than 20 years of first-party experience data, expanded through Press Ganey, Qualtrics allows companies to test decisions against a digital twin before committing real budgets or affecting real customers.

🔹 Lifetime value matters more than token volume. Maynard rejects “token maxing” as a measure of AI success. He points instead to outcomes such as resolving a customer’s problem before it escalates into a support call or churn.

Six months into Qualtrics’ own AI transformation and its integration of Press Ganey, Maynard is defining the company’s next phase around a straightforward test: whether AI helps businesses deliver the experiences they intended and creates measurable value when it does.

Watch the full session at sixfivemedia.com, and explore the rest of our Six Five Summit: AI Unleashed 2026 Enterprise AI Software and Agents track coverage.

Disclaimer: Six Five Media is for information and entertainment purposes only. Over the course of this video, we may discuss companies that are publicly traded, and we may reference their equity share prices. Nothing discussed during this webcast should be considered investment advice or a recommendation to buy or sell any security. We are not investment advisors, and you should not rely on this content as financial advice. Six Five Media collaborates with technology companies and industry leaders to produce research-driven interviews and multimedia programming for enterprise technology audiences.

‍

Jason Maynard:
The biggest change is, for experience, you used to just generate insights, and you had to manually act. Today, we're going to enable you to simulate it, to predict it, and deliver a trusted outcome.

Daniel Newman: 

Hey everyone, welcome back to the Six Five Summit 2026. This year's theme is AI Unleashed. And we're going to be kicking off today's track. It is the agentic track, talking about agentic AI and software. And I couldn't be more excited than to have this track opening conversation with someone I consider a friend, a leader, and the CEO of Qualtrics, Jason Maynard. Jason, you've been on the Six Five before, but not since you took this gig. Welcome back. It's so great to have you here at the Summit.

Jason Maynard:

I am super excited to be here ready to unleash the AI as you said and it's great to be part of your community And and talk with you.

Daniel Newman:

Yeah, it's great to have you back I feel sometimes like I'm looking in the mirror, you know, it's a good-looking guy, isn't he? You guys see that you know how close we are to the summit by the length of the mustache I mean growing it out straight in the summit. So the SAS era was no facial hair. The AI era is all about the stache We have no time. Our agents are working, and we've just got all kinds of things happening at once, and you've got to let the facial hair grow.

Jason Maynard:

You used one of the image generators of what would I look like with a stache, and then you're just like, I'm going to make it in real life.

Daniel Newman: 

We do that all the time. We're here in Dallas at the Qualtrics office here. And things are happening in Dallas, so, you know, I could put a stash on at a cowboy hat, throw on some boots, and I could check it out because of, you know, these AI tools are just, they're so much fun.

Jason Maynard: 

We might have to do a second take or a cut and run down to the store and get you some gear.

Daniel Newman: 

Yeah, we could definitely spend some time talking about the South and talking about Texas, and I'm glad you did come down so we could do this in person. I know everybody's here at the summit. They want to talk about AI. They want to talk about what's going on. And you've been at the center of it. Let's start big. I mean, people might know you now as CEO of Qualtrics. They may remember you from your executive roles across tech, NetSuite, Oracle. But you also had a career before that. You were a Wall Streeter, too, which everyone knows I love that stuff. Working across tech, we are in a time, what I would call exponential change, probably the biggest technological revolution that you or I will experience in our lifetimes. I'm just curious, what do you see as the overall happenings right now across technology, software? What is your big viewpoint on all the change we're experiencing right now?

Jason Maynard: 

It's the most amazing time to be in technology. What a time to be alive. You're living through arguably the fastest and greatest transformation ever. It's kind of crazy, right? And it's funny, when you take a step back and you think about things from a historical perspective, we went from mainframe to client server, to web, to the mobile, cloud, SaaS era, and now AI. Each of these you know, call them platform shifts, I would argue are happening faster and faster, right? We're iterating faster and faster. So the pace of change, I think it makes it harder for everyone, whether you're a customer or a vendor or anyone, analysts following this, the pace of change is faster than ever. And I think that's probably one of the biggest things. The disruptive nature of this is obviously, you know, just starting to be felt. I don't think it's fully distributed yet in terms of the impact, but the pace is what I think is really starting to stand out.

Daniel Newman: 

It's fun to deal with this pace, but I think, especially in the businesses you're in, because you work with so many customers and enterprises that are delivering experiences, the pace creates excitement. It creates a lot of possibilities. It also creates, I would say, consternation. It can be a little bit sticky for people to figure out what to do next. It's like, oh, hey. That model that just came out is great, we should apply it. And then you wake up on the next day and you're like, oh, no, no, that's not the new model. There's another model now that's better at doing that particular thing. But wait a minute, our engineering resources just spent all kinds of energy to deploy this one. No, no, no, no, sorry, you're going to have to do, I mean, that is the kind, like when you were doing, mobile era, when you did even cloud era one, SAS one, you could deploy a tool and you could count on having what, two years, three years before really having to seriously look at a server upgrade or a application change. You could literally deploy something now and 24 hours later question whether or not you have to change it again.

Jason Maynard: 

Well, you look at, you look at the Chinese open source impact just in the, you know, the, the impact that's had in the last 90 days, right? 90 days, right? We go, you know, you think about the horse race on the frontier model front, it's, you know, it's back and forth all the time, right? And this orthogonal impact from, from the, the Chinese and the open source models, a huge, huge disruptive event for the business models of everyone to think about what does this mean in terms of inferencing costs, right? Where does that go? So I think you hit the nail on the head. You don't have two years to sort of get in, see what it's like. It goes 90, 180 day cycles. Like we haven't seen that before. That's all new.

Daniel Newman: 

Yeah, but it is fun. I mean, it is fun. I mean, in my world as a as an analyst covering broadly tech, like I can count every week, there's gonna be a new thing that I'm going to have to learn. And then we had to learn the tooling incredibly quickly, because you had to figure out how to apply this stuff to your work to your business in real time. And, you know, this is exactly what all your customers are going through, by the way, you know, Qualtrics, You come in now, what are you? About six, eight months, beginning of the year. I still remember when I was in Europe and you texted me and you're like, I'm coming over here and you're like, we should talk. And I'm like, great, let's do it. But like underneath the hood, like you guys a little differently are working closely with the, I always say AI's successful adoption will be most notable when the industry customers, meaning manufacturers, hotels, restaurants, pharmaceutical companies, retail, sports franchises, go on their earnings calls and start to talk about how AI is driving their business. I said, right now, you see it in the infrastructure, picks and shovels, you see it with the cloud. I go, but when Eli Lilly comes on and says, we're shortening drug discovery by X amount of months, or when Dick's Sporting Goods says, we've increased retail foot traffic by Y% because of AI, like that, I said, and you guys are literally the company that's helping. So talk about like, in your viewpoint, because you talk about something called the experience gap. Is this the part that like, because we're so busy talking about the tech, the models, input cost, output cost, token generation, speed, but in the end, it's going to come back to experiences.

Jason Maynard: 

Well, we're, I think it's an interesting way to frame it, right? We're at the end point of our customers' customers. What do they experience? What gets diffused to them? When Qualtrics was started, It was the insight gap, which is, help me understand why something happened, right? It was the whole concept of experience data, not operational data. And that was a new thing. And we developed metrics, experience metrics, experience measurement around CSAT, customer satisfaction scores, NPS. And then we would give you these insights, and then you had to manually put it in place. What's changing today is it's not about closing the insight gap. I think that's largely closed. It's this experience gap, which is, what are you experiencing as a customer, as an employee, versus what is actually being delivered to you, right? How do you actually close that gap? And I'd argue, in a world where things move so fast, so many parts of people's business models are being commoditized, right? What differentiates you? What makes you stand out as a brand with a customer? I would argue a giant piece of this, the most important piece of this, is the experiences that are wrapped around the product or service. And so we're helping our customers solve that problem, which is how do you deliver these amazing experiences for the people they serve, whether it's a customer, employee, or a partner. AI can enable customers to do things that they could never do before. So instead of everything being a manual, you know, how do we hard code this business logic to a task? Right? We can activate it. We can make it real. We can scale it. You can deliver personalization at scale to everyone so we all get treated, you know, in a unique basis. That's a big change. You can actually do the things that we all dreamed of 25 years ago.

Daniel Newman: 

And it feels like experience management is an opportunity that right now needs to be prioritized by a lot of the brands you talk to. and work with. And I know a lot of the customers of yours probably are already focusing on it. But I'll give you a caveat. The debate right now is where is an experience going to take place? Yep. Okay, you know, whether it's in specifically your kind of customers have largely built app ecosystems and web services. Yep. And they really deeply understand how their customers consume, whether it's in the foot traffic in a retail environment, whether it's, you know, how they book a hotel online, you know, how they purchase an insurance policy, like all the kinds of clients you work with, and you work across many industries. But like, we are heading to a world where you could argue that the front lines of technology, you know, out in the Silicon Valley, Sam Altman and Dario Allende, they said, we're gonna, we're gonna, everything's just gonna happen on our platform. Now, again, there's a lot of alpha in that for them. There's a lot of value in them to get $2 trillion of IPOs. But at the same time, the markets clearly reacted to it. They created the SaaSpocalypse. They've crushed tons of services industries. It hasn't really directly hit so much the retails and others yet. But the idea that you could book your airline, talk a little bit about where you see that and maybe why. and I don't want to put words in your mouth, but why you don't necessarily see that being the case that suddenly all the experiences are just gonna happen on a chatbot.

Jason Maynard: 

This is really about what will our customers do versus even what we're doing. We're trying to enable our customers, so we're part of that trend. I would say this, AI without context means you can accelerate bad decisions really quick, or bad outcomes, or bad experiences, or hallucinate, all those things. AI slop, right? It's AI without context. And you know, it's interesting because you don't want to sound like an AI naysayer, because in many areas where you can generate verifiable truth, AI right now is amazing. And you look at coding, right? It's a very simple and verifiable truth. It's deterministic though, right? Code is deterministic. But not everything is deterministic on the public internet. If you're distilling things that are out in the public domain, you can find those deterministic places where you can have verifiable truth. But you and I, as a customer of an airline, or a hotel, or a restaurant, or in a B2B scenario, Those deterministic rule sets do not belong in the public internet. They belong to the companies. It's their data, right? So there's got to be a marriage of the probabilistic, amazing reasoning machines of the LLMs with the deterministic rule sets. that govern what's a personal experience for you versus me, right? Because we have different preferences, behavioral attributes, all of the things that determine why we care about certain things. So it's not just going to be the LLMs determining your experience when you get on the flight later today and fly to New York.

Daniel Newman: 

But if I'm reading between the lines of what you're saying and you're not… directly saying this, but another one of your peers, you know, Alex Karp talks a lot about the, he says the alpha of your business is this thin layer that sits between the infrastructure and the model. I'm hearing a little bit of similar sort of line of thinking here, because you're talking about what I like to call the rules and rails of business, that either way a model cannot govern unless you give it that data. and then you give it that data, then you're really kind of giving your business process and proprietary secrets away. And people are doing this. It was the same thing that happened to our privacy. Like you'd said, like what happened to it? Well, we just, we traded it freely and openly for social media or for search. But enterprises now, you're kind of asking a different question, because enterprises have typically protected their data, typically kept it in their data centers. They understand the value of ERP and CRM data and HR data and customer data and transaction data and all the things that you use to sort of help Develop experience data off of you're saying they're not they're not going to just give that up And so I'm a hundred percent confident. They are not going to give that up and so in order to do that though Does that mean like we end up having a lot more? Surfaces do agents as cool as they are and as many things as they're doing for us not necessarily become the epicenter of the average consumer's day. I don't wake up in the morning and start talking to my phone and saying, hey, I want to eat lunch here, and I want to do this later. And it's not just going to book my flight and pre-order my meal. And you're still going to go to a Marriott app, and you're still going to go to an American app. You still kind of see that as how things will work.

Jason Maynard: 

I think it's an interesting question. So I'll give you a couple of things. So if I'm one of those type of brands that you're talking about, I don't look at the model and say that's my moat, right? My intellectual capital, my intellectual property, right, is the sum of all those relationships, all those customer experiences, my employee experience. Those are, that's my IP, right? That's what determines how I generate, I'll use your word, alpha in my own business. I do not see a world where they're going to surrender that proprietary, trusted information. So you've got to change the construct. And this is where I think from a purely, what I'll call holistic view of the world, LLMs do amazing things, right? But they can't deliver the same verifiable truth on your business's IP unless you grant them permission, right? Unless you have the trust that your IP is not going to end up distilled in some model in some foreign country, right? So I don't see those brands handing the keys to their castle. over to any LLM, right? I think they're going to protect that. And that's why when I think about where this market goes, we talked about a verifiable truth for coding. Well, what's a verifiable truth for your experience, right, when you book a car, when you get on a plane, when you check into a hotel, when you go to a restaurant, when you do anything that touches your personal life? Think about your business life. It's the same thing. You have governing rules. deterministic systems, if you want to call it, that have to be respected to generate the verifiable truth, the verifiable trusted outcome. Without that, you end up with AI slop at scale on your business. And I don't think that's what these brands are going to do in terms of turning the keys to the castle over to an external model provider.

Daniel Newman: 

Yeah, I think over time, the technological ecosystem will allow for something in between handing over nothing and handing over everything. Because we've seen it in the past, like with cloud, right? The early days of cloud was like, oh, we got to keep everything on prime. But then they build air gaps and security and photography. But that's the hybrid. And so I think we'll end up probably hybrid along the way. And by the way, when we do, that means that Dario and Sam don't get the alpha. They can be a tooling layer that sits on top. But a lot of software companies are already doing that. They're already building that tooling layer where you have a chat window inside of the application. I think where the opportunity probably lies, and you, I think, would agree with this, is that the era of SaaS, because you talked about the different eras, didn't necessarily build the connective tissue across the different business applications needed, that you would have more maybe a simplified single source of truth that businesses do want. And that's, I think, where the appeal maybe came from of something like a model to one model to rule all the apps is because people want, by the way, this makes sense, people want a better

Jason Maynard: 

They do. Well, I think you hit on an interesting point, which is we've been struggling forever in IT to present a unified view of customer, employee, supply chain, whatever it is, right? SAS era created massive fragmentation of systems. everybody in every organization got a point product for their specific thing, right? I think that era is over. You're seeing consolidation of the SaaS era. Some of it's just companies are struggling. They take on too much debt, right? They get recapped because they can't live to fight another day, right? That's happening out there right now. So there's a lot of things like that that I think are reflective of the trends you're talking about. But I think ultimately what you hit on is like the promise of AI, right, is that I can interface across everything. I can bridge, right, if I'm a customer, right, and I'm looking at your business, I can bridge across all your silos, right? I'm not stuck. You're not shipping the org chart to me. I used to make a joke, if you want a great experience at certain places, you had to have a black belt in how their organization was actually structured. That's kind of a crazy thing. That's not a mass personalized or customized experience for me. It's like my ability to navigate your phone tree should not be the determinant of my success, right? Those are crazy concepts. So with AI, you can fix that. But I think there's a hybrid view here where you have the neural network Bayesian probabilistic reasoning machines that have to be married to the symbolic structure The neuro-symbolic AI architecture, you've heard people talk about this for years, it's not going to be just one thing. And I think that's where you hear the term context. AI without context is slop. Well, what is context? It's the why it is and how your business operates. It's the ontology. For us, it's an experience ontology. How do you build an experience ontology that describes why across this longitudinal set of experiences, you like things a certain way and I like things a different way, and then how do we put that into action? I think that's the right architecture for where the world's going, and that will solve the problem of making sure that any business can compete. Use the models where it's appropriate, but not turn their alpha over, because I think that is going to be a huge issue. You do not want your keys to your castle distilled into somebody's model.

Daniel Newman: 

By the way, 100% this aligns with exactly what I've been saying. You can see my opinions on X. I constantly put that out there. the business will not give the alpha away.

Jason Maynard: 

No.

Daniel Newman: 

It should not. Now, some will make the mistake. You know, it was like the early days. I remember there was one semiconductor company that ran their whole strategy through, like, early GPT and was training the model on their whole major company. Like, it was like, holy crap, this is creating a whole new need for policy within IT. You know, and we've seen it now. I mean, security, privacy, compliance, governance, sovereignty, these are all really important things. Those that can solve it can really, you know, accelerate their business and add a ton of value for enterprises. I do want to hit you on something that you do that's very unique at Qualtrics. Because data is, you know, you can talk about the data is the new oil, data is gold. For all these systems, and you just said this in your last answer, but basically for these systems to be good. They need a lot of data. Yep. You hear about trillion plus parameter models, but even for it to be valuable in your ontology and your business, you need to have a lot of data. Data can be somewhat finite. Experienced data certainly can be somewhat finite. But one of the things Qualtrics does is it creates synthetic data. And this is something I just think the world probably should better understand. But if you have a really high value corpus of data, you can start to expedite solving your experience hypothesis. using synthetic data, how does that work? Because it feels like that's an unlock, especially in the AI era.

Jason Maynard: 

It is. The old world was a lot of listening and getting solicited feedback data, which we still do. There's still going to be need for that. And then you would understand and analyze that data, and you would take a manual action. The new world is you're actually going to simulate these outcomes, right? At Qualtrics, we have arguably the world's largest human experiential data set, right? It's 20-plus years of Qualtrics first-party data. We acquired Press Ganey-Forsta. Press Ganey has the largest patient experience data set. So we probably have more experience data for any economically active individual in the United States than anyone. That gives us a benchmark. that you can simulate different cohorts, different segments, right? Once you simulate that, then you can put it into action and predict, right? Dan and Jason both are in the handsome, bald man category, right? You heard it here first, right? But when you want to predict, right, we have two totally different things that we like to do, right? And so you can start to build these audience segments, these cohorts at the simulation layer, so you're not guessing, right? And you can run A-B testing, you can stand up a digital twin, but then you can put it into effect. so that we get the customized experience based on that set. And so I think what you're gonna see happen is a world where the simulation actually becomes the front end of all experiences. And I use an analogy, I know you're an F1 racing fan. I follow you on X and I see all your pictures. You're a McLaren guy. You're a McLaren guy. But if you think about F1 racing, I think it's interesting. They do the same thing, right? They run Monte Carlo simulations before every race, right? They run tons of simulations to say, this is what we think is going to happen at the Austin track versus the Monaco track, right? They take the cars out. Go to the practice rounds. What do they do? They get more information, right? They do a practice run on a test on Saturday. They get more information. The weather changes on Sunday. They load this information into the machine. They predict what's going to happen, and then they act, right? And they have a higher degree of probability on the track because they've simulated, they're now making predictions, and there's no guarantees in racing, as you know, but you have a much higher probability of a trusted outcome when you do it that way. I think that concept will apply to most businesses. The market research world's changing. Everybody's going to be in market research. Everybody's going to be able to test. Stand up a digital twin. Should we change price? Should we lower price? Should we add this feature? Should we take it back? Should we tweak our service? You'll be able to run all that on digital twin synthetic data, benchmark your results against our data sets, and then put it into action agentically.

Daniel Newman: 

It is sort of interesting because it's become very accepted in a lot of physical industries, meaning we simulate automotive development. We test thousands of crashes. We've been doing this for a long time. Before, we'll actually put it like, you see the dummies. They do that like once. The first thousand times they've simulated it. And by the way, that race data is actually Very good. I mean, if you've actually heard a lot of times, the team principals will come on the race day and they'll be like, yeah, we probably, best case, they'll finish sixth today. Like, they know that. Like, they've watched their cars. They've watched everyone else's cars. They've loaded up the data. And unless something outside of the norms, deviations from the norm happens meaningfully, they're going to get it within like one or two standards of deviation, you know, which is crazy because you have things going 200 miles an hour on the straight.

Jason Maynard: 

You're literally talking about seconds. Yeah.

Daniel Newman: 

And I mean, so that's an invite. I mean, it's such a cool industry. Can't wait for Qualtrics to sponsor a car and to come and sit.

Jason Maynard: 

Right. I'm going to get hit up by like by everyone. Yeah, they will.

Daniel Newman: 

I will tweet it. But the but back to the agent. So where does that inflect? Because right now, everything you're suggesting. It's still fairly engineering-led. You can do it, but you have data scientists engineering-led. You're turning this over to an era where this can be made available and delivered at scale for more companies. Because part of the AI promise isn't, I think it's more productivity, more jobs, but I also think it's more efficiency, more scale. Meaning, you don't need a thousand data scientists working on this problem anymore. Companies of all different sizes can partner with a company like Qualtrics now. and build synthetic data, assuming they have the right corpus to start with and start to make these kinds of simulations. Agents have to play a role here. So where does the agent step in and start to deliver scale?

Jason Maynard: 

This is the pace, right? We've compressed the legacy market research business of doing human panels into synthetic data, right? So you can run more and faster sims, right? The prediction layer is where we can build customizable experience models, right? That then tailor that experience for you as you interface with the brands that you're doing business with. Now the question is, and this is one of those terms I hate. I think it's like the worst term ever. I think it's the dumbest term ever. Everybody's like, the world's going headless. And I'm like, well, if you go headless, where's the brain? So we don't go headless. We're embedded, right? We embed insight. We inference that insight, that brain that says Dan's experience is different than Jason's experience, right? And so I think what you'll see from my standpoint is Qualtrics will embed that inferenced insight into whatever surface area Our customer wants it embedded. It could be their mobile app. It could be their website. Maybe they're using an LLM and we will embed that in a trusted, governable way so that that insight doesn't get distilled into the model. We're delivering a trusted outcome. a privacy-based outcome. So I think that's what you're going to see happen, which is there's a proliferation of services. You and I probably consume our information services in all different shapes and forms, your watch, your phone, your iPad. a kiosk, a check-in at a hotel. There's so many surface areas. You want to keep that context across the entire journey, right? And I actually don't even think it's a journey. I think it's a flywheel. I think if you're a business, you want to create these continuous loops of flywheels. Yeah, virtuous cycles. Virtuous cycles, right? So you want to have the context across multiple flywheels. You're a new customer, that's a flywheel. How do we get you to spend more time with us at the hotel? That's a flywheel, right? So you've got a lot of different flywheels. All along, and I think this is one of the big changes, it's happening quickly, I think there's blowback, but there's this notion of token maxing, which is like the most ludicrous idea I've ever heard, which is like, we're going to try and spend more money. I mean, it's great if you're selling tokens, but our customers are not, they don't wake up in the morning and go, how can I token max? They don't do that. You know what they wake up and they're like, how can I deliver the greatest experience to my customers or employees or whoever I serve? So I would argue the future is lifetime value maxing. How do you lifetime value max? Our job is to make sure we're using the most efficient model to help them LTB max. so they can drive the most economic value in their business. Margin maxing. Well, maybe you want growth. Maybe you want margin, but it's lifetime value. You're trying to generate that economic value. This is the alpha that you talked about. These businesses are going to maintain their alpha. They're not going to cede control. And the notion that it's a toking maxing strategy, I think, is enormously short-sighted. And I think the idea is, how do we enable customers to LTV max? Because that's what determines the most economic value in their business.

Daniel Newman: 

By the way, there are a few companies that love token maxing. And then there's everyone that actually had people token maxing. It's probably littered throughout this event because in different conversations I've had, I talk about this idea of we went from token maxing to token efficiency in like three months. Because every company was like, what really they were trying to do is change maxing. They were trying to get people to behave differently. use this model, go all in, build stuff. And then all of a sudden they're like, wait, we got a $40 million token bill like last month. They're like, slow down there, tiger. We've burned some tokens ourselves internally, don't get me wrong.

Jason Maynard:

I mean, everyone has.

Daniel Newman: 

The joke was everyone has an app and they all have one user, right? Everybody built something.

Jason Maynard: 

But that's the early part of it. That's why this market, go back to your first question, what's this cycle like? It's early, it's fast, there's experimentation with the technology, we're trying things. Like, we are so early in this game, it's kind of crazy. Like, you feel like every day you can wake up and be like, oh my God, I'm behind. It's 8 a.m. and I'm behind. But then you take a bigger picture. We're like in year, what, three, four? Right, of sort of.

Daniel Newman: 

And the models really only became useful this year. Like really before Opus 4.6, I would argue, nothing was enterprise deployable. It created some interesting slop and you could order, you could plan a vacation, but like, you weren't gonna run anything through a tool set with an API connection and try to actually send something that it created to a customer and be like, that's good. So we really only got started, I would argue, early this year. The fun thing though is like, literally, even the models from the beginning of this year are like antiques now. I mean, you wouldn't use Opus 4.6 for anything serious now. No. Like, you're going to use Fable 5 or Opus 5, I mean, or Sol 5. Like, all I'm saying is how quickly, though, what was state-of-the-art became antiquated. It's just absolutely fascinating. I want to wrap this up. Jason, it's been a ton of fun, by the way. We should do this more often. Absolutely. How does this kind of evolve? You know, how does experience XM, your industry, what comes next? Okay, and maybe if you had one more thing, because I know all we have many enterprises out there, probably a lot of your customers listening to this. How are you thinking about how they should measure ROI on sort of their experience, say, gentic strategy, the thing all of you spent the time talking about?

Jason Maynard: 

So the way I communicate to folks internally, because we're going through this transformation. We bought Press Ganey, which puts us out. Customer zero. Yeah, we're $3 billion-ish in revenue today on a run rate basis. We've got thousands, 40,000 plus customers. So we're going through this transformation ourselves internally, and also guiding and coaching our customers through this transformation. I would say the biggest change is for experience, you used to just generate insights and you had to manually act, right? Today, we're going to enable you to simulate it, to predict it, and deliver a trusted outcome. That's a big change, right? Now, I'll be really candid. We're so early in this process. Because we have yet, I think, to really sort out, and this is to your question, where along that experience do you add the most economic value, right? We've got half a dozen use cases where we deploy them all on that journey. And the metric, it's all in support of lifetime value. I joke internally, we're a lifetime value-creating machine. We are LTV maxing for our customers. Sometimes it's helping them acquire new customers. Sometimes it's to prevent churn and downsell. Sometimes it's to prevent cross-sell. My favorite one is prevent the angry customer from calling the call center. You don't want them to call the call center. You want to resolve their problem and make it right in the moment. My favorite use cases that drive value are in-the-moment service recovery. Solve the problem now so you don't leave and call the call center, or worst case, ghost the brand. NPS maxing. But I think it's LTV because NPS is a factor. It's a measurement on that whole experience flywheel that you're creating. But I think that's where you see companies right now picking the spots where they can get the most return and experimenting. And that's what makes this industry right now so much fun is we're reinventing everything. And it's a privilege and honor to do that and work with some great customers because we're in it together trying to figure out what comes next.

Daniel Newman: 

Absolutely. Well, congratulations on the appointment. Six months, lots of changes. I'm following it closely across LinkedIn. I'm building a heck of a team. Came right in and made a massive acquisition. Press Ganey, I mean, you wasted no time. And you have a really interesting opportunity to bring a whole new layer of context and value to this AI transformation. So appreciate you sitting down. Let's do this again soon. Absolutely. Appreciate you and thanks for all the time and questions. Yeah, absolutely. Thank you everybody for joining us for this opening track here at the Six Five Summit 2026 AI Unleashed. Hope you had a lot of fun there. Stick with us. So much more to come. See you soon.