Wade (00:01.898)
Alrighty, folks, welcome to Agents of Scale. We got a couple of great guests today. Our guests are Nick Mehta, the CEO at Gainsight, and Seth Wiley, who is their principal for Enterprise AI Transformation. For those of you don't know, Gainsight, widely regarded for popularizing customer success, has an amazing customer success management platform. more recently, under Nick's leadership, they've done a lot of work around AI, both inside the product to provide AI capabilities for customer success teams,

but also within side game site just to help them run the team and the company better. So really excited to have Nick and Seth with us. I want to kick it off with maybe a question for you, Nick, which was maybe share your aha moment like the first point in time where you're like, this AI thing, we're not doing enough. We got to figure something out here.

Nick Mehta (00:54.676)
every single minute. But what was the first one? was it? Yeah, right. So it's funny. I wish I'd like some eureka moment. Like all of us had the chat GPT moment, Like November 30th, And for me, like it wasn't right after using it, being totally honest. It was a couple months starting to play with it.

Wade (00:56.43)
What was the first minute? was the first minute?

Wade (01:06.083)
Sure.

Wade (01:09.678)
Mm-hmm.

Nick Mehta (01:20.936)
And then just seeing like the basics, you know, we built a feature in Gainsight that did like, I think you guys have done something similar in your own system, like a summary of a customer, like a briefing doc, right? Like an automated briefing. And you see that and you're like, my God, this is like as good as.

the average doc I would get before, which would be like a human created briefing doc, preparing me for a meeting with a customer. like, nobody liked making that human created briefing doc. They actually hated it. And it was never that good. So was like a bad on both sides. And then I saw our team built this feature, which is just a basic, you know, rag, like pull data from game site running through LM. And I was like, my God, this is amazing. And so then, yeah, then we went to...

building copilots and then of course, Seth can talk a lot about it, but then all the internal use cases. But I think it was that, like when we built that first feature and it was like, wow, this changes everything.

Wade (02:15.256)
Seth, did you have a similar, like, aha moment? Was it the same feature? Was there something else that caught your eye?

Seth Wylie (02:20.982)
So for me, I've always from the beginning of my usage of JetGPT at all, I've been attracted to its ability to be a thought partner. And so the turning point for me was when I first started to have it design characters for it to act as those characters for certain, you know, situations or challenges that I was facing. And it would like write a whole backstory for the character and come up with what it's what this character skills were and so on and

that just started to feel like this is a magical way of having new perspectives on things or having access to perspective you don't usually get access to. And that just combined with the ability for someone to say, I've listed six colors, what should the next seven be? And just for it to be able to just speed up the way that you're able to think powerfully, that's what really lit the fire under me. And I wanted to bring that to...

as many people as possible so that their minds could work at a level that they didn't even realize was possible.

Wade (03:24.226)
you have a character that was like especially good for you?

Nick Mehta (03:24.596)
One thing I'm...

Seth Wylie (03:28.502)
gosh. Well, one character is my tutor I wrote for learning Brazilian Portuguese. This is not work related. I can offer a different answer if it's work related. he is basically like, he likes to hang out basically at a beach bar in Florianópolis, Brazil. And he knows that I'm a fan of like Harry Potter and Star Wars, so he's just constantly peppering his comments with those. But also the fact that I did that means that he speaks in a really casual way.

Because I don't want to sound like a dictionary. I don't want to sound like a textbook. And I don't like hang out with a bunch of people who are teaching me the language in person. it's fun, but it's also effective and useful for me.

Nick Mehta (04:11.224)
One thing that brought up for me, because when Seth shared, I think, wasn't one of your characters like a philosopher or something like that? Yeah, Buddhist philosopher maybe? Yeah, I love that one. And it's funny because around that time, I had another aha moment. I think all of us are having so many aha moments.

Seth Wylie (04:20.052)
Yeah, Buddhist. Yep.

Nick Mehta (04:30.36)
right? So they just take you all the way up, you know, to the the moon, so to speak. And one of my ahas was like last year. So I write on the side, like I think that's why Seth and I get along pretty well. I'm like outside of work, I'm like kind of weird. Well, I'm weird at work and weird outside of work. So that's consistent. But for different reasons. Outside of work, I write a lot of poetry. And it's just like for myself or whatever. But I've written a ton, like 100 plus poems. And so one day I was like, let me just

just give them all the chat to see what happens. So I downloaded them from a Google doc into a PDF, uploaded them to chat. And of course, in five seconds, it gave me like incredible feedback. It told me how great of a poet I was, which I love how sycophantic it is. It's so great. It's like every poem is like heartbreaking and brilliant and reminds it of T.S. Eliot and things like that, which are totally not real, but like I'll take it. But like the fact that it like got to know me so well and frankly, that kind of like a really good like.

co-therapist to my human therapist. That was like an interesting thing where it kind of got me then being like, wow.

I'm just going to do everything. So now if you look at my left rail and chat GPT, which I think is a really interesting, like it's a good personality test to see that left rail. It's not just work stuff. It's everything. It's all the basic things like looking up restaurants and recipes and it's personal stuff and it's poetry and it's questions about quantum physics and math and it's stuff about SAS all together. yeah, it's funny how the PR to me, the personal

use cases like Seth was talking about and the work ones they just build on each other.

Wade (06:13.196)
Yeah, I definitely agree. I feel like, you know, so many folks I see do fall in love with it on the personal side first because those use cases are more accessible, more tractable. It's like just easier to get started where, you know, in a workplace setting, know, you can use it for yourself for work, but the real high impact use cases happen when you can start to do it within your team or within your organization. And those require coordination. That's trickier.

Nick Mehta (06:24.716)
Totally,

Nick Mehta (06:38.136)
Exactly.

Wade (06:42.796)
which maybe leads into my next question. Nick, you've talked a lot about this concept of like, the patient tortoise beats the overanxious hare. And I'd love to hear you talk about that with respect to AI, because things are moving just at lightning speed here. I've never seen the industry move this fast. And so how do you think that like analogy pairs well with the pace of AI?

Nick Mehta (06:49.549)
yeah.

Nick Mehta (07:03.768)
Yeah, it's really interesting because yeah, it's fine. Well, firstly, you quoted one of my favorite paradoxes, you know, Zeno's paradox. Well, the how does the tortoise, how does the hare ever catch up to the tortoise if the tortoise has a little bit of a head start? We can talk a lot about that philosophy later. But that idea that I think there is something that we have to be both the hare and the tortoise right now. It's very strange. You know, on one hand, yes, you want to move really, really fast. But on the other hand,

Wade (07:17.603)
Mm-hmm.

Nick Mehta (07:32.096)
we're going to be moving really, really fast for a long, long time. Right. This is not some, all of a sudden your organization is now, is now AI-ified and you're done. Right. It's like a constant evolution. I'm sure you, Wade, you feel the same thing Seth and I do, which is just like almost impossible to keep up with everything going on, you know? And so the truth is, like, you could get into paralysis if you're just waiting for perfection.

But at the same time, if you believe that like, you're just gonna go sprint and be done, you're also gonna be wrong. So you have to almost be both the tortoise and the hare at the same time. There's some analogy to quantum physics here, but like, you actually have to be both all at once right now, I think.

Wade (08:16.034)
Well, maybe this is a question for both of you. How do you actually operationalize that? Like, that is a tricky mindset to sort of balance for one person, much less for, you know, a multi-hundred person organization.

Nick Mehta (08:26.071)
Yeah.

Nick Mehta (08:30.968)
Yeah, I'll talk about it from a product strategy point of view. then I'd love to have Seth has done such amazing work on how we can all leverage this as employees and individuals and as humans. From a product perspective, the analogy I thought of recently was it's like building sandcastles where you build a sandcastle when you're a little kid or even as an adult. And you know that you're building it and the waves are going to eventually like.

take that St. Castle away and you have to rebuild it again. And I think that's the new world that we're in.

Is yet to be willing to have like build stuff that's kind of throw away. Like it's going to be throw away. There's no such thing as like five year architecture decisions, right? Cause none of us know five years from now. I was actually in a, went to a meetup last week in San Francisco and I went, there were these, was like on post-training and there were these three researchers, one from open AI, one entropic and one Amazon. And I was asking about enterprise use cases. They're talking about post-training and you know, verifiable rewards and the RRL and things like that.

time about how much better it's gonna get over time and I was like well what if you have to build an enterprise app now do they I was like do you recommend doing supervised fine-tuning or something else or just just going with the out-of-the-box model and they're like well if you can wait the models will keep getting better I'm like well what if we can't wait then like you probably have to build something and assume it's just gonna have to be thrown away when the next model comes out

And that's, think, the hard reality is you have to be comfortable that every new model rev means you kind of have to start over. If you have a mindset like a sandcastle, you can be like, that's kind of beautiful. You get to do it over again. If you have a mindset like somebody who like classic engineering where you're like, I want my software to last 10 years, good luck. No software is going to last 10 years, in my opinion.

Seth Wylie (10:19.973)
And the way that I think about this internally is, a way there's so many metaphors that we can draw on because the concept of innovation has been studied so extensively in practice, so extensively in different kinds of ways. But one way I think about it is the crossing the chasm curve. So you have the innovators and the sort of like,

Bleeding edge folks at the on the far left then you have this chasm and then on the right you have early adopters late adopters laggards the as I have heard said To be at the bleeding edge. You have to be willing to bleed a little so the folks the left of the chasm are Doing the things that require coding that require extensive experimentation and so on and then there's this motion of transitioning those things to the majority and then we could talk about like

the zone to win framework or other sort of frameworks that have existed to talk about that transition. But the thing is that in the case of AI, it's not like adopting like Zappi or Slack in your organization. It's not like we get the people over the chasm and then we're done. And then there are like some additional features. No, this isn't a curve, it's a wave. It's constantly moving. So how do you set up an organization where like the messy bleeding is happening at the forefront? But that's not the whole organization. It's just the little tip of the tail.

Nick Mehta (11:27.03)
Yeah.

Seth Wylie (11:40.608)
then you have an ongoing process or people of some kind that are helping with the chasm crossing motion. And then the constant adoption that's happening amongst the majority. Because being part of the early or late majority, or even a laggard, is not a static process. It's an active engagement of continuing to learn the things that are being hucked over the chasm at you to be able to learn.

And so this kind of practice of how do you exist in an organization that's a constantly moving wave, that like, I haven't dug too deeply into this, but it's like learning organizations, I've heard them called blue organizations, like there's a understanding too of what it's like to have an innovation culture there. And I think it's about being able to have a foot in both of those camps and respect them each in the ways that they each deserve to be respected, which look very different from each other.

Nick Mehta (12:39.562)
One thing Wade, that think that Seth has done really well, he's very humble, but I'll just flatter him because he's not gonna do it himself, is there's so much of this is like human psychology. And Seth's thought a lot, I think you've thought a lot about the psychology of each of the people and the different phases of that wave.

and how you appeal to each of those, even some of the terminology you use around the different cohorts. I'll let you talk about later, but there's just a lot that you put into how you motivate people, which I think is critical in this process.

Wade (13:11.544)
Yeah, let's actually double-click into some of the specifics on this. So it sounds like the way in which you sort of have approached this AI transformation is that you do have these bleeding-edge teams that maybe experiment and then push stuff into the rest of the organization. But maybe share a little bit of some of the practical tips and tips on how that actually works and what actually really... What choices you made that...

actually drove change versus what choices you were like, we tried it, but, you know, that wasn't the winner for us.

Seth Wylie (13:44.79)
So there are a couple of teams that had leaders that just absolutely led the way. Early on, even before I was doing AI work at Gainsight when I was still in my previous role, they were figuring out how ChatGPT or other AI tools fit into their team. Our support team went ahead and bought a solution, like really invested leaders who led the way. In general, I'm finding that it's not like a few teams lead the organization that everyone else follows. It's more like...

Every team has some experimenters in it. So the question is, how do you foster that bleeding edge within each team and then create an environment where the conversations are happening for those insights and discoveries and techniques and so on to surface and spread from those individuals. So my approach has been, how do I find the people who are already experimenting and help them not be alone? Because AI is such a single player game, regardless of what you

essentially regardless of which tool you're using, that it's so easy for those insights and ideas to be completely isolated. So the very first thing that I did in the whole AI world, even before I was officially in this role, was I like saw some guy on Slack, Brady, who seemed to be experimenting as much as I was and like, hey, let's hang out. And then pretty quickly was like, we should invite other people to this.

That led to a weekly meetup that's 90 minutes every Friday. And there people who do stay the full 90 minutes. There are people who have joined while on PTO because they are thinking like, is just a, they've said that this is just a fun conversation to be a part of. And it is very much a experimentation space. like, everyone just shows up and they say, what have we been trying this week? What's been announced this week? There's always something that's been announced. And so we explore through all the different lenses of the different teams together.

And the thing that has been later coming, but I think that's appropriate is once you have this rich soil of experimentation, then you can start to build the infrastructure for management to say, Ooh, how, what is this person doing that the whole team should know? How do I get the ops team involved in making sure that my whole team is, is trained there? How do we talk about this with executives or with our board about what the opportunities are for those opportunities that actually require investment?

Seth Wylie (16:06.496)
purchasing big blocks of licenses or new tools, those kinds of things. So that's the kind of way that we're, that it is still in the process of playing out a gainsight.

Nick Mehta (16:17.016)
One thing I like really I love about the way Seth approached it is like, there are some things that can be top down, like certain business processes that are naturally corporate processes. Support I think is a good example because a lot of companies have a formal support process. So it's logical that you can slot in a chat bot or a co-pilot or whatever you want to call it. Right. But a lot of stuff in work is pretty like ambiguous.

Wade (16:17.506)
you

Nick Mehta (16:43.69)
ill-defined and very much on the front lines. So for example, actually the work, what a CSM does or a salesperson or marketing, it's actually pretty hard for senior management to just be like, this is what you should be doing with AI. Cause frankly, you don't even know what people are doing day to day. What I love, what Seth did is Seth and Brady and the rest of the team bottoms up like what are the cool use cases. So then when I got to meet some of the folks that were in Seth's program, they would show me amazing things that they were doing that I would have never thought.

because I don't know the job in depth enough to know what those things are. So I think what's interesting is there's certain jobs that are very process oriented where they can be top down. Most jobs are not very process oriented in software companies. And so I think bottoms up is really important.

Wade (17:29.302)
Yeah, Nick, you read my mind there on sort of where I wanted to take this next. Like, what do you think are the areas that you or, you know, insert sea level leader should take a tops down approach? Are there a couple of places where you're like, hey, these are just not going to happen organically and you do need to sort of put your thumb on the scale a little bit?

Nick Mehta (17:42.765)
Yeah.

Nick Mehta (17:50.072)
Yeah, I think it comes down to like where you believe in your P &L, you could have direct financial impact and therefore I'm going to actually drive this top down and where it's like, like you could be like, hey, if I do this, this number will get better, right? So that's why I think a lot of people do customer support because there's a certain percentage of revenue you spend on it and it's either you can reduce that cost or you can...

Wade (17:55.97)
Mm-hmm.

Nick Mehta (18:12.728)
make it scale to more revenue without hiring more people, whatever way you monitor, you leverage that. Like for us, it just meant we didn't have to hire as many incremental people and also could deliver better service to our clients. Right. And so I think you can say there's corporate business problems that we've got. Like, so what are those business problems? Another example of a corporate business problem could be sales to service handoff, right? That's a corporate problem. It's a business process. It doesn't really make sense to, for just one sales rep and one CSM to do it.

You either all want to do it or probably nobody, right? And so there's some examples like that. I have to admit, and I think in a information oriented company, I think those are few and far between. There's not an infinite number of those. I think there's a lot more bottoms up than there are top down. But for us, it's been like, where is there a real impact? Because I think a lot of companies have, mean, Seth and I have lots of dialogue on this. A lot of companies have top down investor board pressure. And I don't say pressure, encouragement, whatever you want to call it, to say, what do you

doing with AI, right? And that's like, I mean, I'm sure your investors wait, ask you about that. Our investors ask us about that, right? And so of course, we want to show what are those things. you know, software engineering, customer support, etc. Right? We can point to how we use illegal, but then there's so much more bottoms up stuff that like would never make the news and the headlines with the investors. But actually, I think is like really important, both for like employees to be effective, and also for them to like their job as well, which matters.

Wade (19:42.35)
But bottoms up side, I'm curious for maybe both of you if there's been like a use case or two that's bubbled up where you're like, whoa, I did not anticipate this in like the best way possible.

Nick Mehta (19:52.46)
Why you start, son?

Seth Wylie (19:52.886)
not sure we would have predicted that one of the first most vile use cases would have been writing in the style of our executives. I'm not sure that's a thing that gets a line item on the P &L that this is saving time or what have you. But to Nick's point about being able to an executive document for a meeting, the quality of work for it to be pretty darn good

when it's handed off, when this ghostwritten note is handed off to an executive to give a review before sending, that has a big impact, especially when it's written honestly in the style. I think a big part of what has helped that to be possible at Gainsight, by the way, is that they were shared as GPTs, custom GPTs. We don't have chat GPT licenses for all employees.

but we have a system that connects to the API so we can build GPT-like things in there. Anyways, it's really hard to share a use case with someone if you're saying, oh, well, you can sort of ask questions like this. And then when it follows up with you, you can sort of say things like that and take it in whatever direction you want. That's not a very effective way to share a use case. Oh, and by the way, here's a giant Google Doc with a huge prompt you can copy paste. But having something like a custom GPT that you can easily share a link with someone.

helps it to spread. So I think it was a combination of a useful thing, useful to many people. It's also inherently social. It's inherently part of like a document that you're preparing to send to someone else. And something that, you know, execs could understand why this is why this is valuable. That was a that was a early low hanging fruit.

Nick Mehta (21:40.418)
I'll give a couple on my end that most of these are just ones where I'll be like in a meeting and somebody just showed me something like, my God, every time I'm like, I'm so impressed. Like not like at a tech level, cause obviously so much of this is the LLM, but just like the...

Ingenuity of the teammates and you know, kind of fan the flames by Seth. So three examples, one was a Reese, like a customer that was a long time customer where we had to go through like what some people call a resell where there's like a new executive and they're like, why do we need to insight? You know, typical thing. If somebody's in SASE, you've been through that before. And the customer had a list of requirements, like a requirements document in some kind of PDF. saw it. was very long, like 20 pages. And then we had a demo to the customer.

And so this CSM took the requirements document and then like our online documentation about our product and said, make a demo script and within the language of the client. And it was awesome. And then our SC came on and did the demo and he did an amazing job. And it's the kind of thing where like without AI, yeah, it would be a lot more work to prepare, but also it wouldn't be as good. It wouldn't be in the client's language and all that. So that was one. A second one is, I mean, it's just,

Wade (22:50.158)
Mm-hmm.

Nick Mehta (22:55.88)
deep research all the time, right? Like clearly, like that's the home run, like killer app for a lot of people. And the way somebody used it recently, they use deep research to like understand this particular customer's business and then connect, like exported some stuff out of GainSight and then had it connect.

the themes that the customers talk about about their business priorities online with what they're doing in Gainsight today. And they built an executive summary slide using Gamma, which is a site creation tool that folks that don't know that basically showed what they could be doing with Gainsight. Third one, very recently, this is one of our ops people, Kendra and her team, who basically built this tool that would take call recorded transcripts from Wiesgong and do the normal thing, pull the transcript and all that, run it through

LLM look for competitors, push alerts into Gainsight so that every time a competitor is mentioned, a CSM would know about it. Again, all three of these aren't like rocket science, right? The magic is really the LLM. But it's really the creativity by the teams and the fact that they're doing this without any top-down drive. It's 100 % bottoms up.

Wade (24:09.868)
All right, so this leads to the next thing I was curious to ask. So you just described a bunch of use cases kind of in and around the customer experience. And you all obviously are doing a lot within Gainsight to improve how folks...

Nick Mehta (24:18.838)
Yeah.

Wade (24:29.418)
operationalize their customer experience. So you all are fresh off your big conference. Talk a little bit about some of the areas and gaps that you think AI is going to be able to drive, improve customer experience and really change and uplevel the way companies can provide this level of proactive service.

Nick Mehta (24:48.664)
Absolutely. By the way, wait, do you mind if I pause for one second? There's just a knock at the door. just want to... Is that okay for the... We can just pause the recording for one second. Okay, back with me.

Wade (24:56.684)
Yeah, we'll pick it back up.

Wade (25:04.654)
Gotta get that Amazon delivery.

Maeve McGeorge (25:25.589)
You guys need to take a water break at all. Take advantage of the time.

Wade (25:28.372)
I'm good.

Seth Wylie (25:31.286)
Myself hydrated. I don't know if it's good for the video, but it is an open AI thing. So at least it's thematic

Maeve McGeorge (25:33.449)
Nice.

nice. Maybe we could get a few shots of you drinking that during that episode.

Seth Wylie (25:49.014)
swag I got as part of being an OpenAI champion. So I don't think they would mind the plug.

Maeve McGeorge (25:52.619)
Nice.

No, I'm sure they wouldn't. All right, feel free to take it over.

Nick Mehta (25:58.92)
I'm back, sorry about that. Yeah, I can go back. Okay, so I could just take it unless you want to rephrase the question. Yeah, I can just take it. Yeah, I'm good, yeah, totally.

Wade (26:01.335)
All good.

Wade (26:07.791)
you could take it if you remember what it was. Alright, go.

Nick Mehta (26:11.488)
Yeah, so one of the biggest opportunities out there is to reinvent customer experience with an AI first mindset. And that means a lot of different things. Gainsight's working on parts of it. I think there's just a big world of problems to be solved. Some of them are what I'll call classic kind of chat bot type problems, and some are more agentic. We had our conference, which we do once a year in the US, once a year in Europe, and we had it in May of this year, 2025. And we all got everyone together and said, OK, what does that roadmap look

like for using AI in the customer experience. The theme of the conference was the movie Wicked, or musical slash movie Wicked if you've seen it. So we made the yellow brick roadmap to the agentic world. So there's a little play on words there. And we talked about four phases that we think companies are going through. Phase one is what I think almost, hopefully almost everyone's at today, which is using LLMs as a great assistant. So whether it's just straight up using Chetchupd like Seth was talking about,

other LLMs, quad, et cetera, to summarize information, write emails, and all the normal, what I'll call commodity use cases, right? Everyone should be doing them. They're no brainers, super high ROI.

Or it's building those into your product. So like we built a lot of those features into our Gainsight products so that it'll write emails for you, summarize stuff, take notes, et cetera, et right? So that's phase one. I'll just call that like basic commodity LOM use cases, no brainer. That'll allow you to be more efficient, more effective, better experience. Phase two then is...

Rather than waiting for the employee to say, hey, I need help with X or can you help me summarize Y? It's taking more of an agentic approach where it's pushing stuff to you. Right? So we all know that that's a huge opportunity. In our case, we bought a company called Staircase.ai that basically analyzes communication with their customers and with no human interventional flag risks and opportunities and sentiment issues and renewal problems without any human intervention whatsoever.

Nick Mehta (28:15.778)
And so think there's a category of agents that will basically flag things for your team, know, anomalies, issues, things before that, you you would have to go looking for and now agents can help find them for you. And so I think there's a whole class of problems there. We're solving some of it. The next category, phase three, is then giving agents to your customers. And so this is a world that's very hot. This is like the self-service world where it's a agent chat bot for support. It's an agent in my knowledge base.

navigate the documentation. We build a community software product agents in the community to auto moderate posts and things like that. We have a learning product that helps you train your customers agents around creating more personalized learning. So a lot of agents that will show up in the way your customers interact with your service. But then the final, I think the Nirvana, at least what our customers want is the ability for an agent to truly autonomously achieve a business goal. Right. So agents that actually can

start with an open-ended goal. in the case of

our world, how can I renew a customer? How can I drive better adoption? How can I onboard them? And so we announce what we call Atlas, which is a family of autonomous agents where you give them open-ended goals. So for example, I've got a long, you guys have a huge long tail of customers at Zapier. So you probably have the really small ones where you could never put a human on them. How do I automatically onboard them, drive adoption, renew them and giving them open-ended goals. And then it would go query a bunch of information, design the right intervention.

whether it's a personalized email, a video, or even a phone call and actually reach out to the customer and guide them through the steps of that journey. So what's awesome is all of this stuff is being built as we go, but the first couple phases are much more mature. The autonomous stuff is bleeding edge as Seth alluded to.

Wade (30:09.326)
What do you see when you look at Ural's customer base? When you look at the customers that are on the bleeding edge, the ones that are doing the most, that are impressing you the most, what do you think separates them from the rest of the pack?

Nick Mehta (30:25.14)
Yeah, I think there's a couple of things. So one thing is that it's easier for some companies to experiment than others. Like being very empathetic, right? If they are selling into a non-regulated industry, if they have smaller customers versus bigger ones, right? If you only have 100 big pharmaceutical customers, it's a little harder to be on the bleeding edge. If you've got...

10,000 small collaboration customers or for example, one of our.

pilot customers for this agentic stuff is DocuSign. They've got all these small customers using DocuSign every day, a little bit easier to experiment, try new things out. So that's, think, one ingredient, which is the nature of your business. Second one, I think, is the culture. Have they created a culture of experimentation, kind of like what Seth is doing for Gainsight, right? Have they empowered employees to try new things? Or is it a culture where it's like, no, no, no, everything has to be top down and we can't try anything out, right? Is there that? And then third is,

I think there's an element of like how much initiative people are taking.

versus waiting for the technical people to figure it out. hope one of the things I love about AI is I hope we break this idea that certain people are born anointed as technical or non-technical. I can't tell you the number of times people said, well, I can't do that. I'm not technical. I'm like, that's not a thing. It's not like a D, it's not a gene, right? That you're technical or non-technical. I was on a call with a round table we did with some of our customers and there was a chief customer officer on there. have a company, probably about a hundred million of revenue, something like that.

Nick Mehta (32:02.049)
And I was like, oh, what are you doing with AI? And he's like, oh, we built a whole tool that uses LLMs to take information from our customers and then build an onboarding plan and automate some of that plan. I was like, that's awesome. Did you involve your engineering team? the CCO, chief customer officer, was like, no, I built it myself on the weekend. And I was like, really? And he's like, yeah. And I'm like, have you ever coded before? He's like, no. I was like, that is awesome.

That's why this stuff is going to change the world, right? We are, we are like opening up this black box of like being technical and making it available to everyone in the company. And by the way, I'm not trying to try to make, I'm not trying to make commercial for Zapier, but you guys are always friend centered in that because of course people then want to integrate that with everything else that they do. And inevitably they organically bring up your.

Seth Wylie (32:38.594)
And if I could

Wade (32:38.775)
Yeah.

Seth Wylie (32:43.2)
You

Seth Wylie (32:52.436)
And I would love to build on that because I don't think it's just about the sense of like, am I technical or not? I think it's also a sense of, do I have the smarts and agency and permission to reinvent how I do my role? I think that, you know, the C-suite folks of the world generally don't feel like anyone's going to tell them exactly how to do their work, that it's up to them to figure that out. And they have chiefs of staff and things to help them figure that out.

Whereas I think especially individual contributors feel like their job is to do the job that's been prescribed to them. And they might not feel a sense of permission and certainly don't feel like a sense of they have gifted time to work on this thing, know, a la Google 20 % time or something, to figure out on their own how their work should be done. And then when they bring those ideas to management to expect that they'll be celebrated instead of greeted with some kind of skepticism.

And the degree of psychological safety, the degree of permission that's needed in order for people to become those reinventers of their role, I see that, you know, I don't think that just separates companies from each other. I think it separates teams from each other. I think it separates individuals on the same team.

Wade (34:02.284)
Yeah, I love this idea that like being technical or not technical is not a genetic trait. It's not like we were woke up, like woke up being born left handed or right handed. Do you like bringing it back to the maybe more like some of the internal transformation work? Have you have you found any techniques that instill that like agency in folks in the organization versus do you feel like that's just more innate?

Nick Mehta (34:09.073)
No. Right.

Seth Wylie (34:28.79)
I think they need to be invited into some kind of virtual or in-person space where that permission exists. I think a lot of us have had that experience when you go to the offsite. The rules of the room are different than if you're on a Zoom meeting, regardless of what level you're at. It's big picture thinking time. There's a whiteboard. There's post-it notes. Like it's a whole experience. Maybe there's a professional facilitator in the room. I think that it's possible to create those experiences for folks in ways that are not quite so grand.

And I think it helps to have, like in my case, I think it's an important thing that I wasn't, you know, a manager's manager within their team. I wasn't their peer. I was this other guy who now lives in the IT department, who is creating this optional space where if they're interested, they can come. And when they do come, like their boss isn't there. This is like, this is, it feels like that kind of offsite. And actually this is a way I describe using AI in general that,

good use case for anything, good use case for chat GPT is anything where you want to have like an offsite with yourself, because it's basically your conversational partner to figure out some big idea thing. So that's the main thing that I think about is how do I facilitate conversations where people feel like this is them stepping back from their day to day, which is the same kind of, like I learned that skill from being an operations leader, from being someone who helps a team leader to whiteboard their problems.

And that's what feels like it unblocks people for a while. And then that builds the habit of that feeling like it's okay.

Nick Mehta (36:03.343)
One other thought that I'd add in, in addition to, think one of the biggest thing a leader can do is have people like Seth that can help inspire this bottoms up. think you can also, as a leader, if you want to make this kind of transformation, you have to be using it yourself. My opinion, like, don't, I don't, I think it's going to be really hard for a company to make an AI transformation if the CEO and CEO, founder, executive team, if they're not just using it all the time. like when I'm in a meeting now, basically deep research is running like all the time and it's, it's, it's.

very humbling because like somebody will be talking about something and I'll put in the question and then it'll come up the exact same thing that that person was just about to say, right? And so you to kind of accept that that's just the reality. But it's it's really important, my opinion, that leaders are using it like an example that I did recently, I did somewhat in partially intentionally, partially just serendipitously. I was on a call with a prospect and they said they were looking at some other competitor and it was a small

company I didn't know them well and they said how do you compare this other company now of course we have a competitive research team they're great but I was like you know what let me just go into deep research and say compare us against this build a 10 row grid with Harvey Ball comparison the vendors or whatever so I did it I sent it to the customer and by the way it's very authentic because it was like not me it was ChadGPD like actually doing the analysis right and then I posted it online internally and I said hey I just did this and I think that kind of thing encourages people to be like okay well if the CEO is doing it if the exec team's

doing it, I have the permission to be doing it, the encouragement to be doing it too. So now, mean, Seth, you've done all this research and I think a very high percentage of our employees are using AI in their jobs every day. It's thanks to all the great work you're doing and I think some of the top-down encouragement as well.

Wade (37:52.376)
shift it back to customer success for a minute. know, customer success, customer support, this is an area where there's a lot of stuff changing. And, you know, probably one of the few roles that is seeing the changes the earliest. I think all of us are going to see this in all of our roles eventually. But that is definitely one that is happening the fastest. Based on what you've seen, what, you know, what are the...

Nick Mehta (38:06.856)
Right.

Wade (38:17.784)
What are the best customer success reps doing? What are the folks that like really are like proving themselves to be irreplaceable doing today?

Nick Mehta (38:27.613)
Yeah, I'll throw out what I'm seeing and then I'd love to see what you're seeing as well. I think that you can divide the world into like maybe it's just oversimplification, but two segments of customers. So imagine you're a company and you've got the customers that you're able to spend a lot of time with and the ones where you're not only able to spend a lot of time with, right? So if you're someone like DocuSign or Zoom or somebody else, have millions of customers, right? So clearly there's a ton that you can spend a lot of time with mathematically, financially, and there's something you spend a lot of time with.

There's some kind of two different use cases in each of those cords. For the ones you're not spending a lot of time with, well, it's all upside. So how can you design digital programs, agents, automation? And so you're seeing a lot of people do really incredible things around everything from.

dynamically generated communications to like I was mentioning voice calls, video avatars, you in app, like, you know, you probably you've seen like traditionally people do like in app guides that like you would start a product like Zapier for the first time you get like a walk through having those be programmatically generated, right? Like having learning classes be programmatically generated and automatically updated when there's new content. So a lot of stuff for that long tail of your customers on the high touch side.

I've seen people just doing a much better job challenging customers in a very respectful way based on the real business goals. I'll give you an example. have an awesome CSN has been with us a long time named Elliot, and he did this presentation for a customer where we were both on together. He basically used AI.

to build a SWOT analysis, Strengths, Weaknesses, Opportunities, Threats, of how this customer was using GameSight. And it was like a one slide executive summary. And it was like the best slide I'd ever seen us present to an executive. Because it challenged them. It wasn't just like, here's the four things you're doing with GameSight, great. It was like, here's what you could be doing. Here's the opportunities. Here's the threats you roll out.

Nick Mehta (40:29.471)
and he used AI to prepare it, put his own color commentary into it. So I think that's an example where Elliot is already a great CSM, and he can be even better because of AI, and he can challenge a customer even more, and then the customer is going to get even more value. So I think it depends on the long tail, where it's more about do stuff for customers that you could never do before. And now for the high touches, how do we do it even better?

Wade (40:52.79)
What guidance do you give to leaders who are trying to navigate these transitions, but they might have teams that are resistant because they're fearful about, is this going to take my job? Is this going to replace the thing that I love? I like the creative side. Da-da-da. How do you help? What advice do you to leaders to help them get folks to be more excited about the opportunity here versus fearful about the change?

Nick Mehta (41:17.439)
Yeah, mean, it's fine. The fear is legitimate. It's understandable, right? Like it's very, first of all, don't like deny the fear. mean, it's legit. Like every human's wondering like, what's my role in this universe in 20 years, right? Like, and none of us really know the answer. That's like the truth. Don't pretend like you know the answer. Don't say some cliche like technology always creates more jobs than it replaces. You'll all be fine. Cause people see it through that from my way. But at the same time,

Wade (41:23.757)
Yeah.

Nick Mehta (41:43.069)
We all know that we don't know it's all gonna be bad and it's probably gonna be bumpy. And so what I like to tell people is it's gonna be really bumpy. Be honest with them about that. But you'd rather be on the side of the folks that know how to use this stuff, they're growing, they're learning, and we know that it's vital to any future job.

Right? Like learning is not just about being a game site. It's about all your future careers. Right. And like, this is about honing your personal craft and your professional craft all at once. So I think try to, my own perspective is try to avoid the platitudes to placate them because they're going to see right through it. But at the same time.

Encourage them to lean into the ambiguity. You know, when people say, wow, things are changing a lot, I'm like, yeah, get ready for it. It's going to change even more. And this is just the new world. like, it's best thing you can do is embrace it. Well, how about you, Seth? You think a lot about this.

Seth Wylie (42:36.931)
Yeah, I have a different but complimentary perspective, which is I feel like when I've heard the resistance, it's from people who generally their conceptualization of AI is this is going to automate the tasks that I do, whether it's coding or interacting with customers. And I have felt the temperature in the room change in a tour in the direction of relief.

Nick Mehta (42:49.809)
Mm-hmm.

Seth Wylie (43:03.142)
after they've had an experience where AI has helped them do something that they think is really cool or they care about. So if we're talking about software engineers, example, if they are afraid that AI is just gonna code everything that they ever wanna code, if instead they're able to have an experience where they wanted to build something like this before, but they didn't know the framework, it requires a really complicated API that they've never worked with before, what have you.

If they can use AI to do that in five to 10 minutes, then suddenly that doesn't feel to them like a threat. It feels to them like an unblock. And then they feel like whatever it is that AI, I think, what they think is whatever it is that AI takes up, I will be able to use that entire stack of technology to reach to the next step further that AI can't do yet. And that feels to me less like it is...

less like it's holding them back and more like it's giving them a sense of empowerment. And that's really the antidote, I think, to the sense of threat.

Wade (44:07.086)
You know, we've seen something. Go ahead, Nick.

Nick Mehta (44:07.421)
I love that one, I was gonna add one final thing, but that might be helpful. I think that, just get, got me thinking.

In life, often we can get typecast at work and even outside of work, right? You think of an actor who's like, oh, that's the actor who's only in comedies or only in horror movies or whatever, right? And the same thing happened at work. Oh, that person's not technical. That person's technical. That's just a marketing person. That's just an engineer. But what's awesome about AI is it allows you to do so much more than people think is possible. And so yes, you can be technical now.

You can actually be a designer, even if you're not a designer. You can be a marketer, even if you're not a marketer. And I think that's pretty inspiring.

Wade (44:49.454)
I think that's awesome. The thing we've seen folks have, our customers have a lot of success with is running hackathons. You at this point, like you give people the chance to put their hands on the keyboard and start playing around with this stuff, find that aha moment, and they'll quickly realize where this stuff is, you know, jaw-dropping, exceptionally impressive, but they'll also see the areas where it's not so good. And like it is kind of...

Nick Mehta (44:57.608)
Yes.

Wade (45:17.794)
funky. I've heard this term jagged intelligence talked about recently where it's like, man, there's these incredible spikes where it's just far outpaces human capabilities. And then there's other tasks that humans do with ease that the AI just simply cannot figure out how to do some of these things yet. And it starts to...

you start to see a path forward, think, as a human where you're like, oh, wow, there are certain things now that I get a sidekick that is way smarter sidekick than I've ever had in my entire life. And that is incredible. And I'm excited about collaborating with it on these areas. And then here's the place where I get to have some agency and push and pull and all that sort of stuff. And it becomes less of a...

I don't know, the fear part is still there, but there's a more curious, explorative part too where you're like, I wonder where this is going to go. And so it gets pretty exciting.

Seth Wylie (46:04.308)
Well, and I think that's why, you said so much earlier in this call, that that's why the outside of work use cases are the things that I think drive the most learning. There's something about, I mean, this sounded smarter in my head, but there's a strong relationship between curiosity and learning. And there's only so much curiosity that you can bring to making a SWOT analysis. Like that is an outcome. That's an outcome you get once you learn how to do the thing. But I can guarantee you that Elliot didn't

Wade (46:22.147)
Mm-hmm.

Nick Mehta (46:22.227)
Yes.

Nick Mehta (46:26.584)
Yeah

Seth Wylie (46:33.888)
build the skills for that by saying, I'm going to watch the YouTube video about how to make spot analyses.

Wade (46:40.846)
I love that. That sounds like an amazing place to sign off. So Seth, Nick, thanks for the conversation here, guys, and sharing a little bit more about how Gainsight's done the internal transformation and the things you're doing with customer success. And for those listening, stick around. We're going to have more Agents of Scale. We'll be back with more conversations around how other leaders, companies are navigating growth, complexity, and this crazy world of AI. See you next time, folks.