E59: Usage-Based Comp, AI, and the End of the SDR (Tyler Will, VP of RevOps @ Intercom)

Tyler Will is the VP of RevOps at Intercom. Before that, he spent 6 years at LinkedIn and 5 years as a consultant at Bain. And now he's building for the new GTM challenges we're all facing in 2026.

We've rethought compensation, we've rethought pricing, and we've rethought our ICP. The post-sales model is actually quite different in an AI world.

In this episode, Tyler breaks down why too many CROs are obsessed with the wrong metrics.

Your CRO or whoever gets up at a sales kickoff and says, close more bigger deals this year, now go get them. And everyone's like, what am I supposed to do with that?

We discuss the reality of which GTM roles are already getting automated out of existence.

That SDR agent is going to be one of the first things that goes. You used to have however many people sitting there dealing with a bunch of junky leads, and now that's, you know, that's an AI job.

Plus, Tyler shares how RevOps leaders can protect themselves in this shifting landscape.

The days of sitting there waiting for people to ask you to complete something, I think, are gone. If you're showing up in the forecast meeting to take notes, you are going to lose your job to Gemini or whatever it is that is on that call very quickly.

And we cover a ton more: comp plans, consumption models, and the rise of forward-deployed engineer. Plus, Tyler shares the essential advice he's given to every new VP of RevOps.

At the end of the day, it sort of boils down to 3 words, which is—

Welcome to the Revenue Leadership Podcast. Here's a familiar scene: a deal's in commit and the buyer says, can I talk to somebody like me? Next thing you know, sales reps are asking CSMs after hours, 5 people are in Slack trying to remember who helped last time, and that one customer who always says yes is getting burned out after 3 requests in a week. That's why this episode's brought to you by User Evidence, the platform that turns customer proof from chaos into a repeatable GTM system. With advocates, references, and real customer evidence all in one place, your team can share credible proof instantly and keep deals moving forward instead of stalling. Companies like Gong, Ramp, and Workato already use User Evidence to capture customer proof, build buyer confidence, and shorten sales cycles. See how at userevidence.com. That's userevidence.com. Today's guest is Tyler Will, who is the VP of RevOps at Intercom, where he has been for 3 years in revenue and sales ops leadership, uh, following a 6-year stint at LinkedIn in RevOps and strategy and a 5-year career as a management consultant at Bain. And today we are going to talk a bunch about the AI transformation journey at Intercom, which has gained a lot of notoriety. Obviously on the product side, they've added Fin, which has done really well in the market. We're a Fin customer actually. And I think unpacking the journey and lessons and how other companies can be thinking about this transformation, it's top of mind for literally everybody right now. So I'm excited for this one and thanks for joining, Tyler.

Yeah, Kyle, appreciate you having me on. I'm excited to dig into this. Cool.

So what did I miss in your background apart from my two-line intro?

Not much. I don't think my stint as a consultant learning about price fixing in corrugated cardboard industry and stuff like that is what you need to talk about. So that's the key highlights for us today.

That is a conversation I want to have over a beer because I'm like, I love Malcolm Gladwell. So that screams Freakonomics and Malcolm Gladwell.

Uh, which is still a lot. There's a lot of, there's a lot of interesting tales of old school industries from, uh, from that time in my career, but not AI in any way, shape, or form. No. Cool.

So, uh, I want to start with a tangential warmup question. So, uh, before we recorded, you mentioned something which I loved is that, uh, revenue leaders are oftentimes too obsessed with success metrics. And not rigorous enough on input metrics. And I would say that is true. Many of us are guilty of being too myopic about the success metrics. And so just like unpack that. It's a cool, it's an interesting opinion. So tell me what you mean.

Yeah, I mean like the sort of joke version of that is, you know, your CRO or whoever gets up at sales kickoff and says, you know, everyone's fired up and goes, close more bigger deals this year. Now go get 'em. And everyone's like, What am I supposed to do with that? Right. And that's, that's sort of the silly extreme version of it. But you know, what I'm trying to get to is the notion of like understanding the entire machine that you want to build in your go-to-market. And so you can't just do that by looking at the outcome at the end of the day. We sold, you know, X many deals, this ACV and this much ARR. You need to actually understand the whole system and how it works and be able to instrument it, monitor it, talk about it. And importantly said, targets. So you can do that. And then, you know, that lets you as a leader actually go out and give meaningful instruction to your team. You can talk about, you know, how much pipeline we want to generate, what that pipeline should look like, how many accounts RevOps needs to assign to people, or what signals we're looking for in our accounts. And so you just want to be able to build a system, I think, that lets you have some sort of like X-ray into that entire process. Of course, the other extreme is like you work all the way backwards and you say, it turns out the key input metric is people coming to work. So like we want everybody to come to work every day. Like that, you know, that doesn't also breed success. It's the, the other extreme of this sort of silly way of phrasing it. Like it's, it's understanding the entire system, um, and being able to monitor it and talk about it and give people guidance about the things that really matter in it.

And I would imagine, well, My argument is in the AI world, this systems engineering mindset is going to be more important than ever. And this is what's going to differentiate, you know, good to great CROs over the next 5, 10+ years. Fair statement in your mind?

I do think so. And I think the, you hear about this in individuals and their jobs, right? Those people who embrace AI and embrace technology are those who are going to rise to the top. And those people who resisted are going to be the ones who ultimately get left behind. They might actually feel like they're ahead briefly in the future while the rest of it, or in the near future while we're all stumbling around figuring out what this is, but ultimately, and I think the same is true for a CRR. You're going to need to be able to think creatively about how to apply this, where AI creates real value in your customer interactions, where it does things as well or well enough compared to a human that you can really take some of the painful work out and then find the ways to get really creative and apply humans where humans do great human things, which is like connect with people and talk about interesting topics and get to understand a business in a way that maybe an AI sales agent or something can't do it.

Yeah. And so you listened to last episode with Jason Lemkin, who is at this end of the pendulum saying a CRO or CMO needs to actually go and implement one AI tool soup to nuts on their own to truly understand it. That would be like sort of one extreme. The other extreme being like, oh my, like my RevOps Tyler's going to figure this out for me. I don't need to be super AI native. Where do you land on that?

I think it probably depends a bit on the size of the company and like the practicality of that, which I think he also probably would have said in that, uh, in that, and you've talked about like scale mattering a lot, like You know, at Intercom we've got 250 sellers and then my team's going to be, you know, 45 people kind of early next year. Like that's probably not a great way for our CRO equivalent to be, you know, who's spending their time. But I think it would also be a mistake to be at the other extreme of that and sort of say, I'm going to put blinders on and assume this is going to happen in the right way. I think while maybe it's not implementing it probably in a way that that any CRO has not in the past, like understanding the use case, maybe being in on some of the vendor demos. Um, I think that education will be really valuable to anyone at any scale. And so if you sort of draw the pendulum and then, you know, use the extreme, like you got to do it yourself to the, you can be completely ignorant of what's going on. I fall probably two-thirds of the way toward, toward Jason's position on that one. Okay.

So what level of, I want to get one level more granular. So what level of like AI understanding do you think a, a CRO at a late stage growth company actually needs? Like I often, I often say like, if you don't understand the difference between pre-training and post-training, if you don't understand like recursive learning and like ingestion versus, and like tuning, like if, if some of those concepts aren't, maybe not second nature, but if you don't have like a, a rough understanding of, of what that is, it's you're probably going to struggle over the next little while, but like, is, is that too far? Like, what, where do you put the, where do you put the, the, the setting?

Yeah, I think there, there are almost two dimensions to that to, to think about, at least coming from an Intercom perspective. One of which is, you know, we're selling an AI product, right? That is our— Finn earlier, and sort of to put it as impolitely as possible, you're going to look pretty stupid if you go out to your customers and go, you got to get on board and use AI, and they go, How does it work? And you go, I, you know, I don't know.

It's AI. Don't worry about it.

Let's, you know, yeah, it's AI. Like, let's pull the sales engineering team in here to talk. Like, so, so there's sort of a necessity of if you want to sell this, you better be very, very fluent in your own product and how your own product works. And so anyone trying to, to skate by on, on, you know, this sort of what is RAG? Like, you better know what that looks like and The, we have this notion of like ATTD, which analyze, test, train, deploy, and like it's this flywheel and, and with Fin, but you gotta be able to talk about that. What does it actually mean to analyze? How are you using AI to analyze it? What does it mean to test and train and, and deploy? And so you're gonna have to be very fluent in your own product and whatever that language that is a necessity, uh, sort of a necessity to, um, be able to talk intelligently to your customers. I think it seems one vector that you're gonna have to be. Kind of as far out as your product is on that. If it's, you know, a lightweight wrapper around AI, like maybe it doesn't matter as much, but for anybody going deep on this, I think you've got to be very, very literate and conversant in that. I think the other angle is sort of like the internal use case. And I'd probably give a little bit more grace for flexibility on that front. Like, is it vital that our CRO knows exactly how Clayworks or how, you know, to use Claude Code and Snowflake to pull their own data and do, you know, this, that, or the other, like how we build agents in Glean. There's probably a little bit more forgiveness of like exactly what people to understand there. But again, if you're a CRO or another leader in the go-to-market space and you're not embracing these things, I think you're going to be quickly left behind and you're going to have people who have embraced it kind of running past you and you're going to be left. Left in trouble. So probably, probably a place better to overcorrect than, uh, than to undershoot, I would say.

Yeah, that's a good— that I think is a good takeaway.

Like, there's plenty of things in life where you could be like off by 20% either way and it kind of doesn't matter. This one is like, you're probably better at 95% than you are at 75% if, if 85% is the right answer. Yeah, very, very, very well said.

So, so going back to this like, uh, success metric output versus input metric. How do you— what is the mental model for finding the right input metrics in your business?

Yeah, I'm trying to think if there's like a really good framework. All my consulting years maybe didn't produce this one for me, but, um, I think there's an element like really understanding what the key things are that make a difference, right? And I sort of joked earlier, like, coming to work, like, nobody's monitoring whether you come to work, right? We're not doing like factory punch cards. We need to know that, you know, you're going to use certain tools, uh, you're going to contact your accounts in the right way, you're going to respond to signals that we can throw off from product usage or, you know, other sort of parts of the data-driven business. And so we're going to try to manage a few things that I think we find to be really key, uh, key like engagement points in the customer lifecycle. Right? So it might be if you're on the like account management side, hey, for certain types of accounts, we want to know that you're having a QBR with them to talk about the value and the ROI, because that's what ultimately leads to big growth and sort of the right profile of customers. So we, we, we monitor that. We might monitor something very different on the kind of hunter new business side where it could be, uh, you know, number of, you know, number of demos or, or whatever it is that we think is, is really important to. Generate good healthy pipe that leads to hitting your target. You could try to instrument everything. You've talked about this in other settings of like, you know, keep it as what was like the minimum dose for the maximum result. I always think of minimum effective dosage. Yeah. Like I was a tennis player and then my coach would talk about hitting the minimum shot for the maximum result. Like if you've pulled your opponent way off the court, you don't have to hit the ball 100 miles an hour to the opposite corner. Hit it in there and like, and win. And we want to sort of come up with the same equivalent in, in this instrumentation exercise. Like, what is it we need to do to really understand our business to not find ourselves in a spot where like, oh, if only everybody had done this thing, we, you know, we would have been fine. And it turned out nobody was doing it, but not, you know, how many emails did you send? How many calls did you say? Because people start managing themselves to that metric and the efficacy of the thing they're actually doing kind of goes away. So yeah. It's going to vary, I think, business by business, but the effort that you put into trying to figure those things out, I think can be really, really valuable in improving your own understanding of your business. And so I like the exercise itself to kind of go, it's this and not this, I think can be hugely, hugely useful.

Yeah. And you made an interesting subtle point there too about not being myopic about it. Like if, if you see this in BDR teams with dials, if you just like breathe down everybody's neck and you make it all about dials, then you invariably see dials go up and decision maker connect rate go down because people are just doing ghost dials to get their manager off their back. And so like every, like anything in life, every strength is a weakness. And so how do you, how do you dial into the right amount of pressure? Like, yeah, right amount of pressure on a metric so it's enough to drive the behavior, but not too much to start to give, to lead to a bunch of second-order effects where people are, you know, they're, they're not accepting ops in their pipeline because you've been hounding them about conversion rates.

And then, you know, it's like every, like every, or the opposite of that, which is you gotta have X much pipeline and they'll accept anything just to get you to leave 'em alone on, on the manager, or you get their manager off their back. Back and said, yeah, it's, it's, it's tough, right? You need, you need enough guidance and enough motivation that people have a sense of how to, uh, to do their job, to what's going to make them successful. Uh, but I think enough flexibility where, you know, people who— you've probably seen this, right? There, there are BDRs, there is a BDR in the world who'll make 10 calls a day, but those calls are such high quality, you know, that it's, that it generates the 6 ops that they need, right? Or whatever it is. Meanwhile, there's somebody making 50 and it's their way of getting, you know, getting the same outcome. And so you want to, I think, understand where you can sort of set some guardrail. Like if you're doing less than 10, you're not really working. And if you're doing more than 50, the quality is so low. But can you like bound things? I think it'd be helpful. Can you give flexibility? Yeah. The other one we've talked a lot about recently is where do you do the inspection? Like at the manager level where there's maybe a little bit less, right? Like, like if you tell every rep you need to generate $5 million in pipe this quarter, everybody's going to go generate $5 million of pipe. Yeah. Yeah. Conversion rate's going to be, it's going to be all over the place, but everyone will find a way to create opportunities and, you know, put in the sizing so it adds up to $5 million. Like you now accomplished absolutely nothing except tell people to write down $5 million on a sheet of paper. Right.

Yeah.

But if you do that at the manager level and kind of say, hey, look, we think about $5 million is right per person. We want you to stay on top of this. but we really need the quality to be there. So, you know, how do you inspect that in your, you know, your sort of team forecast meetings? Maybe that's a more effective way of, you know, this particular example of doing that. There are things we do at the exec level where we're going to look down and sort of say, like pacing is a classic example. You never tell every rep to like pace exactly this way, you know, through a quarter, but you need to look at that in aggregate and understand how you're doing and if that's creating risk for the business. And so, not only I think is it the metrics, but it's the level of your org that you're going to inspect them at that matters a lot where you can get the goodness without getting some of that kind of bad behavior, like people just behaving in a way that gets them left alone, but doesn't actually create value for the company. Yeah.

I think a lot of that is how do you tell the frontline manager how to enforce this metric or like guide to this metric? And if it's like, you know, a barrage of Slack messages, just like forcing something down somebody's throat, it's like you will generate second-order effects that you're not that happy with. But if you can teach nuance and you can teach people to like understand the— actually, these are the 3 metrics that matter. You know, back to your point on success metrics, everybody wants to just have like one metric. Oh, this is the one metric that defines success for whatever. But it's never that easy. And so if you can teach, like going back to BDR, because it's simple, if you can teach them, hey, DMCs matter, your decision maker connects matter, your conversion rates matter, but really like ops out the other side are the thing that we care about and like, you know, quality ops. And if you're triangulating those things in problem solving, you know, one of the, one of the Things that I like try to drumbeat in is, is like if a rep doesn't have a lot of calls, I don't send them a Slack message being like, hey, go make them more, make more calls. I'm like, hey, go figure out what's, what's broken. Like go figure out what the problem is and, and like, like help this person sort it out. 'Cause you know, people want to do a good job. Like every person in your sales team, you know, if you're a company like Intercom or Owner, like people want to do a good job. You get to hire good high quality folks. You just oftentimes there's things in the way.

All right.

I wanna shift the, the conversation to, to AI and a bunch of this thread I'm gonna pull through on success metric, but maybe just start with the internal transformation. Like where, where did you start? What are, what is the, what does the broad AI transformation look like today? And then, then we'll unpack a bunch of the components.

Yeah. Yeah. So, so Intercom, they're sort of two worlds. There's the R&D world and they were probably a bit ahead of us in doing it. Finn was out months after ChatGPT launched. What is that? 3-ish years ago at this point. And so they went through their own transformation. I can't really speak to all the details, but I think there's plenty of stuff on the web about Built for You and then other sort of product development kind of things that they talked about on this, but they were certainly at the forefront. So that's one big like element of Intercom's internal transformation. The go-to-market side has been almost everything, to be honest. You sort of talk about this, this, and what are we going to talk about? Like, well, pretty much, pretty much everything. So we, you know, we've thought, we've rethought compensation, we've rethought pricing and how then DealDesk interacts with pricing and kind of what levers we want to have for discounting policies and things like that. We rethought our ICP. Who, who do we actually want to go after? Who is good for a rules-based help desk model is not necessarily the exact same customer who's an ideal fit for, for Fin, uh, in a resolutions-based model. Um, we're forecasting differently. Uh, the data we're looking at on a regular basis is different and more, maybe, maybe more so than some of the hunting and kind of pre-sales part of the work, but the, the post-sales model is actually quite different in an AI world and love to spend some time talking about that. Cause I think it's quite interesting as we've seen, you know, an evolution of kind of renewing seats and what customer success motions looked like in that world to one of constant, you know, driving usage, pushing up more usage, expanding the use case of Fin and AI products is a very different motion for us. And it's really required us to rethink that a lot. I think we're, we're probably in the very early days of that evolution, but it's been the one that I think probably feels the most different internally from our GTM teams.

I'm going to ask one broader question and then I would love to just dive in right into that seats to usage transformation because so many people are in it. So what from a systems and rituals perspective is the most different now? Like what have you had to completely rebuild?

I'm not sure if anything is like completely rebuilt on a like RevOps systems side. Probably like compensation is probably the thing that's been the most overhauled in the last calling 18 months going through things. And so that's the place we continue to evolve and try to figure out exactly what the right, you know, the right levers are for us to pull. But the, you know, forecasting is a little different. There are things there, but still like similar base. I think it's probably that. Okay.

All right. Let's get into this post-sales motion because this is definitely relevant to so many. So maybe just talk about the monetization model of, of the AI product of Fin, and then we can get into what changes you've made in post-sales.

Yeah. So, so really simply, we have a resolutions-based pricing model. So $0.99 per resolution, where resolution is, at least in the customer support use case, um, is viewed or is defined as, uh, either the The customer's customer saying like thumbs up, that helped me. Thank you. Or what we call a soft resolution, which is basically the conversation ends where Finn has made a recommendation, provide an answer, and they basically don't come back and ask anything for, I think, 40 days. So those are the ways we assess that the conversation was resolved successfully. And that's what we monetize. Okay.

Interesting. And so now what does the contract look like they're signing? Is it platform plus resolution? Is it 100% resolution?

It's, it's, uh, it depends. So we have both versions of those things. So if you're a happy customer of, of whoever your help desk provider is, it's not Intercom, we can put Finn on top of that. So if you're a Zendesk user or Salesforce Service Cloud or whatever, you can integrate Finn into that and have it be your AI agent, which is pretty cool. That's been out for, yeah, coming up on a year now. And then we also have this classic, like, suite, use Intercom help desk and have Finn on top of that. And so one place it's purely resolution contract or agreement and the other it's a contract for seats and resolutions. Okay.

So then how does this impact the post-sales motion? What are the changes you've made?

I think there are a few. One is kind of that onboarding phase. And I don't want to suggest that a CSM's job or sort of onboarding team's job was ever simple, but You know, there's something almost quaint at this point in time about like selling seats and they provision it to everybody and they assign a bunch of users and they kind of run a tutorial and it's like, you know, have at it. And LinkedIn, when I worked there, had, you know, that was kind of the model. It was like, bought LinkedIn Learning, you bought recruiter seats, like, let's show you how to use them and like, call us if you need anything, right? And, and there's, so there's that today, but that process is much more now about helping the, our customer, Intercom's customer, uh, set up everything they need for AI to work really well. And we found this especially in the early days with, with Finn, like the documentation was terrible and, or there was conflicting documentation or, you know, sort of all the things you need to feed into the, the model and give it guidance weren't actually available. It was in people's heads. It was sort of, you know, in the, in the cultural waters, but it wasn't written down anywhere that you could get an AI agent. So we spent lots of time early on in those kind of first 3-month periods with a new customer, just try to like almost helping them write documentation so that they use the, you know, use the tool successfully and have it answer correctly. Um, and so there's, there's a, you know, that's a completely different world than like everybody has a seat and you kind of have already trained everybody and it's off training. And then beyond that, it becomes much more of an effort of getting customers to unlock Fin on, on more and more conversations. We talk about this notion of involvement rate. So if you have 10,000, you know, kind of incoming customer support needs in a month. What percentage of those do you let Fin interact with? And customers often start with maybe 1,000 of them or 15,000 of them where, where they're saying tier 1 simple stuff. They're like, yeah, what are your hours? What's your return policy? How, you know, how does this thing work? Like pretty basic stuff. And you get set up that way and you continue then to build. Well, you know, what's the next thing? Now we need procedures that can actually interact with the other systems. Now we, you know, now let's turn on your phone channel. Now let's turn on your email channel. And so there's much more of a kind of continuous learning and continuous engagement that is, that is needed. Um, in addition to always having the, the sort of informational part that, that feeds Finn and gets what I want. Complete opposite of a SaaS world where you didn't like, again, using the LinkedIn example, you didn't give people LinkedIn Learning seat, which is a product I worked on there. And then like only let 'em use 15% of the courses for the first, you know, the first 3 months. And then like, Once they got good at that, you let them use the next 15%. It was like, have at it, here, here you go. And now we're in this world of like the CSM team and, and the account management team needing to continue to engage, continue to, to get the customer to be willing to unlock that next tier of, of things or the next channel of things, which is then interesting because it becomes both a sort of success motion and, and engage motion, like getting people comfortable with that. But then it, then it becomes a commercial conversation. Again where you are, well, now we've used up our budget, right? We bought whatever, 10, you know, 5,000 Revolutions, but now we're, now we, we found use cases for 7,500. Well, that wasn't in our budget. We got to go back and do that. And, you know, maybe you're not yet ready to rip out the human seat, you know, the human support seats to save the money and kind of repurpose things. Um, and so it's just a very different kind of almost constant sales process and just really and prompt you about the next thing versus a, like I said, I'm using kind of my past examples of like, there were more sophisticated ways to use LinkedIn Learning or use a recruiter seat, but like not that much more sophisticated, right? You didn't go from what are, you know, what are your banking hours to like, I need to, you know, change the owner on my mortgage and I need a new credit card. And you know, and now you're in identity verification procedures and, you know, taking actions with, you know, financial transactions and a bunch of regulations and, that's a very, very different world than like, I now watch leadership courses, not just, you know, how to use Excel kind of a course. So it's been an interesting, and we are like just figuring out what this looks like.

And so how has the structure of the team changed? Does that mean you have account executives doing post-sales? Is the CSM more commercially oriented today? Like what has it meant structurally?

Yeah, it's, we've done a couple different things. One is, is I think continuing to evolve the nature of the CSM role is kind of the core part. We have AEs as hunters and then we call them relationship managers, RMs, plus a CSM on the post-sale side, which is kind of the core account team. So we're asking the CSMs to be more technical than they probably were before. A little bit more, we're not asking them to sign the deals, but they need to be more kind of commercially like savvy in that we've got to identify the opportunity and pursue it and bring bring the RM in at the right time. We've also introduced some new roles. Uh, we, we built out an entire R&D services team, which is kind of our— they'll both hate this and it'll be the easiest thing to say— is like the forward deployed engineer. I knew you were going to come.

And would you call this R&D services?

R&D services. Yeah. It's not, it's not Palantir model, but it's, you know, it's a little bit more hands-on. Like we can help people build stuff. Um, but ultimately we pride ourselves on the fact like our product is is sophisticated in its outcome, but not in its setup. And so we want people to be able to self-serve ultimately, but that's a totally new team. And that team is, you know, 8-ish months old. We did not have that before. We've actually split the sales engineering role into pre- and post-sales expertise to get people to— there's one thing to run a demo and sort of show people how to set it up. And it's another thing to be moving on to these more advanced use cases after kind of your initial commitment to Intercom. And so We're splitting roles, we're creating new roles, we're asking more from, from certain roles as well. And that's all a product of, uh, the, the nature of this AI product.

Yeah, why wouldn't you want to be called a forward-deployed engineer? It's like the cool new thing.

Yeah, like the, the, the RevOps equivalent, the, the go-to-market engineer that, you know, Clay invented, uh, 6 months ago.

Yeah, everybody wants to be GTME now. And so talk to me about how you have upskilled the CSMs. Has this been about bringing in a bunch of new, more commercially savvy talent?

Is it—

is everybody making the leap? I would imagine not, but like, talk to me about how you transformed the team there.

Yeah, the, the enablement team, uh, that I don't own has been a sort of the key, the key player in that. So they run something we call FinU, Fin University, come teach everybody, uh, and the, our customer-facing teams how to, uh, how to talk about Fin, how to think about Fin, how to use Fin. And so there's been a, you know, first and foremost, right? A concerted effort to upskill the team we have. That's certainly cheaper than turning everybody over and trying to find new people. So that's been one, but we're also growing so much that it's been, I think, a big advantage to bring in people from the outside who either have similar experience or are eager to pick up that experience. And so you get a nice kind of mixing of Intercom veterans who have learned and evolved and people from the outside who can kind of bring a different perspective in.

Yeah, makes sense. Okay, let's talk comp. I'm curious to unpack how compensation has changed both in sales and CS, if it has there. Um, yeah, like at what point did you change the comp plan?

I think people are all wondering, like, to everyone who tells us at every point we were changing the comp plan. Yeah. Um, We've had, we've had multiple evolutions of this and, and this is, uh, we can get, we can get really into the weeds if you want on this one, but you know, the, the first thing we found was that customers were often sort of committing to a little bit of fin and then using a lot, but going into what we call overages or like kind of a pay-as-you-go model. So classic, like usage-based mechanism. And none of our comp plans had any way of capturing that. So we were, whatever you, whatever you signed for a contract was the, was the payment you made and the quota of time where you got. And so we, we quickly found that our, our sellers were out, you know, as soon as somebody go into overages, they'd run them down and go, you know, put another 100 FIN resolutions a month on, you know, under a contract.

Like, right.

That's, you know, that wasn't a great experience for anybody and kind of a waste of our reps' time to bring a very small amount. It was the only way they were going to get paid. And so we first started introducing kind of our first phase was, was to say, well, we will, we will continue to count the contracted part as, you know, fully retiring your, your quota, but we'll also give you the, basically the MRR of non-contracted usage. So let's say you had a, some contract and you signed, you know, whatever, $60,000 contract with somebody. So you get $60,000 of credit against your quota. And then over that quarter or that half, they also used an extra $5,000 each month, we'd give you that $5,000 toward your thing. We try not to, we try not to annualize it because then you start getting this like, they use a ton in one spiky season, they're clawed all back and next one it comes down. And so we didn't do anything like that. We just gave them monthly credit, which was sort of a, I think a happy middle ground. So if people are struggling with this, like there's a nice mix where, where I think the reps felt cool on being rewarded for driving success and engaging with this customer. I don't have to make them do something they don't really want to do or put a bunch of discount on it to convince them to do this. Hey, if they're really seasonal, like imagine a tax software company, like no one's on chat with their tax software company in September. They don't, you know, they don't want to, they don't want to prepare for that. But in February, March, and April, you know, it's, it's crazy, right? So they need to staff up and be ready for that. And you don't want to force them to buy that contract for the year. So it was a way for us to reward those kind of short-term spikes or sort of steady usage before a contract with, and they would get that in the first 12 months or something like that, or they'd get, they'd get it immediately. So let's say it's, it's January, you sign a $60,000 contract and then they're using $5,000 in overages in February and March. You're going to have $70,000 in, in kind of quota credit from that example, $60K from the contract plus $5K plus another $5K. but then hey, if it went to zero in April, 'cause for whatever reason that's not their business model, like we didn't have to claw something back because we're needing something else. So, so it was a, it was a little bit of like first step into this process. Okay.

But would they get that for the first 12 months of that customer coming online? Is it into perpetuity?

Uh, yeah, that's, yeah, good question. So the, the, the, we set, ended up setting up so the account exec who landed the new logo would get 12 months of that credit. Hmm. As well as then the account management team would, would get it as long as they own the account. Okay.

Makes a lot of sense.

So that was like phase 1 and that was our first like toe in the water. We've got this problem. Uh, we want to do it and it felt like that worked pretty, pretty quickly. Um, and so we ended up, and here's where it gets, it gets real messy. We, we ended up moving our entire account management team, the, the RM role to a pure MRR added up over the course of the, the comp period. So they have a semiannual plan, and we say, you're going to take all your accounts that you own over that, that period, and every month each one generates some amount of MRR, or zero if it, if it churns. Uh, we're just going to add that up towards some target we've set for you, and that lets you basically be a purely like usage-based thing, even in a world where a lot of their, their revenue is under contract because it's contracts or seats contracts, but if it's a solid contract, it just flows for 6 months. You get that MRR, like you're set, but you then, uh, you then get the usage ups and downs and it's, it's been, it's been interesting. It's complicated. I don't not recommend it, but I also am not like, we've, we found the answer. Everyone should go, should go do this thing. It's, it's a little messy at times, but it has certainly brought focus back for us onto both expansion and, uh, sort of better churn contraction management. And it has let us weather some of those ups and downs of a usage-based product where we don't have reps running around chasing every, you know, every overage and trying to convince people to bring it under contract. It's been really nice, I think, both for our people, but also for our customers who can experiment and turn things on and off and they're not getting sort of harassed by commercial or about commercial terms for, from all of our reps. Yeah.

Um, is there a set of underpinning variable comp philosophies that you come back to?

Yeah, I mean, the, the, I think the classic, what I call like sales comp rules still, still apply, right? You don't want so many metrics that people can't understand it or don't know what to do. So we've kept every plan to, to 2 metrics. I think maybe there's one plan that has 3 metrics. Um, we've tried to make those meaningful so there's not little piddly like 5% of your comp is that, you know, this thing or that thing out. Um, we want to make it something we can instrument so people can always see how they're performing. It's, you know, I don't need to tell you, like maybe you're for your audience, like don't surprise your reps at the end of a period with how much money they're going to make. Like that should be a thing they can know and see all along. And so we've stuck with those and we always come back to those when we're, we're doing things. Is this, you know, is this sort of simple enough? You can explain it in 5 minutes on a whiteboard. Is this, uh, is it going to feel fair? Is it going to reward work and the behavior we want? Those, those I think are always true. And no matter how far into the AI era we go, I, I think always remain true and you need to come back to those. Um, but it's let us then experiment a little bit with like how we want to create the right incentives and do some other, other things like as the nature of the product and the use of the product changes.

Yeah. So simple. Fair, tied to the right behaviors, something that you can instrument and the rep can watch through the period.

Yeah, I think those are, those are kind of the four. If, if we feel like one of those breaks, we're gonna go back to the drawing board. Okay.

And how often do you feel like you can change a comp plan?

We did the one major overhaul now a year ago on the, the account management team being on, on MRR. We had a, a fairly unfortunate reputation of both quarterly planning and changing the comp plan quarterly for the first part of my tenure, which none of us wanted, but was, was a bit necessary as we were going through all these changes. We, we make, I would say we make tweaks every 6 months when we kind of do a, like a refresh on, on where things are. Uh, but we try to keep the core elements the same. We don't want people to have generated large amounts of pipe that looks one way thinking that's going to be really advantageous in the comp plan and then go. Just kidding. We, you know, we want everybody to go over here instead. So, you know, as we think about, we're starting, we're sort of deep into the planning process right now. One of our, one of our sort of added principles at this time has been how do we, how do we tweak the plan without really changing things? So we have the same core metrics, you know, FinARR, FinLogos, all, you know, all in ARR is kind of the core things. Maybe change the weight a little bit. Maybe we're, we're talking about putting some accelerators on certain aspects of things we want people to do. New products are coming out, Finvoice. We've got our customer agent that was announced at our, our big event, uh, back in September. Um, maybe it was October. And so, you know, you start to layer in some of those things, but it's, it's at its core, it's essentially the same plan, but we try to be responsive to the interruption. Yeah.

How often Do you think you can do like a real overhaul? Is annually too much?

I mean, I'd sort of say annually would be about the most you could do, but you've got to be responsive to the business, right? I mean, you're like dying on a hill of a principle. If we don't change our comp plan more than once every, you know, every 24 months and running the wrong business for 12 months is a, you know, as a terrible idea. And so. I think if you're, and we had this problem particularly on the CSM side where every quarter they were getting a new plan and they were just like, what are we doing here? That was crazy and pretty unfair, I think. But yeah, like minor tweaks at the 6-month mark, if you're kind of doing a semi-annual refresh is okay. And if the business really demands it, you can do every year. But I think you need to give people ample notice, make sure your enablement is really good. You're probably going to want to roadshow it. Cascaded down the org from your, you know, second line leaders down to your frontline leaders. And ultimately my team runs office hours for, you know, the month after new comp plans are just like, hey, coming out, we know you're going to ask the questions. So we're just going to be here and try to make this as accessible as possible for people because it's, it's probably the most sensitive topic I've, I've come across in my the cost you're generating, which makes sense, right? Is how people get paid and put food on the table. And so when you change the rules or people feel like the rules, maybe not fair, which is often just maybe not fair is code for, I don't actually understand how this, you know, this works or what I'm supposed to do with this. I think is, it's just essential that you kind of go out of your way to explain that to people, to walk people through that change, help them understand why. How they're going to see it, how it's going to show up. Often that's like, hey, that deal you sold last, you know, last time, like tell me how I would get paid in, you know, in this version of things. And you want to go through that as much as you can to do that. And I think, you know, one of the knocks on RevOps teams is we're all, you know, a bunch of cold-hearted, like Excel jockeys. And so how do we, how do we get out of that a little bit and get out there and talk to people and first take their feedback into the design? And even if it's Hey, we heard you, you want a spiff for, you know, this, that, and the other thing. Like, we're not going to do those and here's why we're not going to do them. I think helping people feel heard, especially about a topic as sensitive as compensation and in a world where we're changing it pretty, pretty regularly is, uh, is really critical.

Yeah. And like being honest about it, you know, like I find the, the worst comp plan rollouts I've seen is when you try to stick people with a change that just like straight up not good. It's like, hey, we're just paying people way too much on this type of deal because the accelerator was bad or the accelerator was too generous. You're like, we have to move back this accelerator and you go out and try to convince people why that accelerator going away is a good thing for them. It's like, no, it's like, you know, we're, we put this accelerator.

Yeah.

We put this accelerator out there. Now we feel like it's, it's compressing our unit economics at the end of the day. We are all agents of capitalism here. And like, so we're going to make this change, but, but not bullshitting people into, into like telling them why, hey, why your quota going up is a good thing for you.

It's like, no quotas go up because that's how startups work.

And there's, it's a, you know, a law of startup physics. So what's the bar for like a real overhaul? Like when, when, because you don't want to change it too often, but like, when do you know it's like, hey, we got to make a change?

Yeah. You know, I, I think for us at least, and then maybe we can extrapolate this into a, a generalizable principle, like significant change in pricing. Hey, we moved from a seats-based to a, a resolutions-based model. The comp plan, the way we do, the way we do commerce here has changed. Like our comp plans have to, to reflect that. It's a big pricing change or, or I think big product change is gonna be the other, like, time in which you're going to have to really step back and say, hey, are we, you know, do we need something new here? One of the things for us was early on, like nobody's making big commitments to Fin. And so we felt like the reps were, were sort of not, I'm not going to, I'm not going to drag this out to pitch AI to somebody who's maybe reluctant to get, you know, $5,000 more on this deal. I'd rather just move on to the next deal. And so, you know, that was a place for us, like we got to do something. To get the attention on this. So we, we gave a REIT 2x credit on, you know, on every FIN deal that they sold. Cool. And that got everybody's attention, you know, out right now. We've gotten rid of that because everybody now leads with, with FIN and, and talks about it. And so I, I think you need to really assess like, hey, is some fundamental part of, of not salespeople or, or whatever changed, but like, has the business changed in such a way that our comp model is now missing some key thing that we need everybody to go do. And that's certainly the point we got to when we made our big changes. It was not before that, it was not too far after that, and it was really tied to that. And it's been interesting for us because one of the things that I think as R&D and go-to-market are getting closer and closer on some of this stuff at Intercom is, you know, there are times where like, we want to launch this new thing right now. And you're like, ooh, our comp plan really isn't going to help with that. And we've got 3 weeks, you know, we got 3 weeks to go in the quarter until we can make a change or whatever it is. Like, can we coordinate that a little better? Can we do them simultaneously and enable the product and how we're working? So there's some opportunities I think people can take advantage of to like tie it up into one nice big package. That said, you know, you don't want to move too slow. You don't want to delay, you know, big product releases just because the comp plan, and then you use SPFs and you use other things to modify that. But Those are the two big things I've seen where we're like, old world doesn't work anymore. We need something new. Let's go. Let's go.

Yeah, that makes sense.

Yeah.

The thing we're trying to get better at is aligning product roadmap with go-to-market roadmap. It's really tough because you see these things coming, they're on a roadmap, and then they just sort of sometimes show up and we're like, we're ready to drop it. And unless go-to-market is ready, and sometimes it's comp change, most most of the time not. But, you know, like, we need to go take this, like, big message to market, or we want to turn everything on. You're like, well, actually, we've got 3 major initiatives all lined up, and we can't put a 4th in front of the— for the— in front of the reps without massively diluting it. It's, it's a— it's such a tricky thing because, like, you want to move with maximum speed, but you need to introduce some of these, some of these processes, which naturally do slow you down a touch. But, like, what is What is that minimum effective dosage?

Yeah. And, and it's, it's the slowdown to go fast, right? And everybody's probably had the example of like product launch something like nobody in the field knew about it. You know, that stuff gets missed. And then everybody realizes very quickly that waiting 2 weeks to launch it and running an enablement session would have saved everybody, you know, a bunch of trouble and a bunch of awkward conversations. And so great. Like you kind of learn from some of those painful moments and I think can improve that process. But it's, uh, yeah, it's a constant trade-off of, of speed versus like having a fully baked plan for how everything's gonna work together.

Yeah. So I'm wondering if you could share more about the comp planning process. Like what was the process you used to say, all right, so business model, product pricing, some, some like big change. We think we need to change pricing. Like, where did you go? Like, how did you start that? And just walk me through it.

Yeah, so a lot of it was, was led out of product and pricing of here's kind of, here's the direction we want to travel. We've done our research. We think, um, you know, sort of outcome-based is, is going to be the way to go. And, you know, Intercom, I think rightly, is very proud of itself for kind of being the leaders on that, you know, resolution-based pricing in the market. And so, you know, we, we in that case followed, you know, followed their lead. And then it's And then it's sort of weaving that into what we have as our pretty standard planning process. Finance has a, you know, sort of a top-down target. We're going to split that out into, uh, you know, products if there's product differences between teams, obviously region segment kind of things. And then we're, um, we're just going to break that down into kind of here's how much we need to generate. How do you want to set a quota? Do we have the right components of that quota? Like I said, you know, is it Just Fin, yeah, it's part of your quota. Just Fin is part of your quota. You know, the whole package, how do we want to think about those trade-offs? But it actually, it actually worked pretty well to like weave pricing some of the new product stuff into our, into our existing product or into our existing process. Sorry. You know, I think that the place where we've seen some interesting things has been on the, like maybe the periphery of planning a little bit, which is Uh, when we changed pricing and Intercom had a long history of pretty painful pricing, hopefully you weren't, we weren't a customer at that time and have only experienced the new, the new fairly good version, which is, which has been great to like clear that up. But you know, their customers on old plans and they migrate over and that changes their pricing a lot. Like we don't want reps telling customers to migrate or not migrate based on the quota and what we can move. So we set up an exceptions process where people get relief if the migration to a new pricing plan was You know, sort of not to their advantage.

Level up in abstraction. Like, how did you, like, what was the process by which you found the edge cases? Or like, what was the comp build strategy? Do you start with like, okay, these are business priorities, so these are the metrics. So this is like, like, I threw that strategy.

Yeah, sorry. I think that's, I think that's basically it, right? We have, you know, a strategy, which is. Introduce an AI, you know, an AI product. And at about the same time, we need to move people onto our new pricing plan. Those things are kind of happening in parallel. And so that, you know, those become the business objectives, right? Like get people onto the good pricing. Everybody's using Fin is going to have resolution pricing and everybody needs to go sell Fin. Those are like, we set company level objectives every, every year and often related to, to Fin or, or things like that. And so those are going to then drive everything in that process below that. So we're really trying to align to company strategy and, you know, people setting comp will often get 40 requests. We want the comp plan to do, you know, there's a litany of things. And I think for us to be able to anchor on the couple that we think really matter at the company level was kind of the key thing, sort of like those, whatever, we, 4, you know, 4 kind of core principles that should never be violated in the comp plan, period. We kind of had the same, like, what are the 2 or 3 strategic things that this plan really needs to reinforce? So that for us was, was the starting point of strong alignment to company strategy. And from there you can figure out all the, the gory details, what we're going to do, what's, what's going to, you know, what can you actually instrument? What can you actually set a target for? Um, but that's where we try to align ourselves most. Okay.

When do you bring in reps?

Yeah, it's, um, when do you bring in the feedback cycle?

Maybe more generally managers, reps, regionals.

Yeah, managers and sort of the director, the second line leaders. We brought in, particularly when we did the big overhaul, we tried to bring them in very early. So we used, uh, we used some consultants to help us with some of that, like think through things. And so we brought them to sessions very early in the process. So, you know, months, months in advance of making the change, we wanted them to understand it. There's no way if that change had just been like a RevOps announcement and everybody's staring at each other that it would have gone well. And so we needed to get the directors and the managers to go, we've seen it, it makes sense. Here's why we're doing it. You know, here's the kind of the reinforcement of the good and acknowledge the bad or the trade-offs we're making. And so we brought them in very early on, on that process. Now that that plan has been live for close to 12 months, we've done much more of the like, let's go talk to reps. You know, when somebody kind of raises their hand and says, I don't think this one's fair. I think the team has, you know, not every time, but at times and with people who've, you know, expressed constructive opinions are like, very happy to engage and say, what is it? Or did this, did you do something differently because of the comp plan and try to learn? And then yeah, if we can make improvements, we'll make improvements. And if, you know, like in life, not everything is perfectible. If we have to live with it, like at least we can, we can go talk to people about the pain points they're having and say, you know, the alternative to this was this thing and that thing breaks all this other stuff. So we're going to stick with this one even though it has these, you know, these kind of warts on it. Yeah.

If you did it all again, what would you do differently?

Um, I think I still would have spent more time with the, with the management team, getting them educated on it. I think we probably like started really strong with engagement and then kind of went into heads down mode and got out of that engagement model. And so even though they'd, they'd signed off, uh, and kind of said like, yeah, we get it. This makes sense. I think that when it actually was revealed, it was kind of like, whoa. I'd forgotten what we agreed, you know, what we'd agreed to 2 months ago instead of being brought along on that journey more. And so I think that was one kind of change management mistake that was made. Um, there are probably some places where I don't know if we got like too cute for our own good. Uh, you know, sort of like overcomplicated. Oh, we've layered in all these different things and here's how you're, you know, the MRR is going to add up or here's how we're going to set CMS targets around, you know, around resolutions and it's going to be sliced and diced this way and that way. I think there are probably some places like the simplicity was, was lost a little bit in trying to address some of these like, oh, but won't this happen if we, you know, if we don't capture it, like let's, let's try to squeeze that in. But then I sit here today and I'm like, well, which one, you know, which exact ones are those? Like it's probably like the accumulation of too many being slightly too, too creative for your own good that creates challenges rather than any one particular one. But I, if I could only change one thing, I would, I I would land the change management a bit better.

Um, all right, I want to change topics, but I am really glad we got to talk comp because I've been trying to get somebody to talk comp on the pod for a while. So I think this was real— this was really useful. I always wonder if people listen to this and be like, Kyle, you're out of your mind, and this was really boring. But like, I, I'm like, this is so valuable and interesting. And, uh, so hopefully people have the same love for compensation planning than I do. But I'm curious, going back to the, the broader topic of AI transformation, so how does RevOps today look different? Like what, what roles are different? How have you constructed the team? Like maybe talk to me about RevOps structure.

Yeah, the, um, the, the, the one, the one very concrete thing, or maybe there are a couple of very concrete things we've done. One is, is we have hired our first go-to-market engineer. So that Uh, that is like the classic, uh, plant your flag on being AI enabled, uh, for a RevOps team. We've done that one. Uh, and then as part of this broader—

when was that? That was recently?

Yeah. Like a month ago.

Interesting.

Yeah. Okay. So, but the search took a while, I think, which is the recurring theme of folks trying to hire that person as well. Cause hundreds of them out there who, you know, who've been doing it for 5 years, they're like, people you think might be able to figure it out. Um, so yeah, really excited to have her on board. Um, and then what we're doing actually is we had both the sales team and the solutions team were kind of covered by the same ops function. And so a little bit of like, you guys be good at everything. And as these complexities of the post-sales motion that we've talked about have come out, we've actually decided to split that and we're creating a separate solutions ops team to really lean in and support that post-sales motion more effectively. So those are, those are kind of the, so far, the, the roles and org structure changes. I think where we're seeing the need for more and more is a sort of change is, is almost more like knowledge and skills. And I think we, we talked about this before as well, like kind of do— my hypothesis is that like generalists in RevOps will, will kind come back and, and sort of in the long run. And like the great example of this is, is we've been set up with Claude Code. And so I don't now have to go to my analytics team to ask it a question and have it, oh, this is really complicated. We got to do all this work. It, you know, come back to you. Like I can just do it. And I, you know, was doing it this morning and I was doing it yesterday. And it's like, it hasn't eliminated the need for an analytics team, but it's changing the nature of that work. They actually, we work closer to the business. There's no longer like a file a ticket, we'll build a dashboard, you know, here you go. It's them getting embedded and coming to the forecast call and hearing, here's a thing that's going on that we're concerned about. I can, you know, I can go dig into that and write a great write-up about it. It's much more, I think it's evolving much more toward a, like, people hate this, like ad hoc analysis all the time versus BI team. And we agreed on 10 dashboards we need and we went and built them. And then you introduced a new product and we, you know, we tucked that one in. Uh, it's more of like a business advisor and sort of investigator kind of role. Because if I need to find out how much ARR we have by customer segment and region, like I don't— I can do that myself now. And about You know, in about 3 minutes.

And, and so is Claude code going into your BI tool then, or directly into Snowflake?

Into Snowflake. Yeah. Okay.

And so you're just querying your local Claude code or, or the, like the computer app version of Claude? Yeah.

I just have the terminal up and like just ask it stuff and it, yeah, just like reads the, reads out of Snowflake. But we also have Glean and Glean will search documents and, you know, so again, it's the, ability to like self-serve, which is like steadily decreased in my career, right? As you get more senior, suddenly can't do anything except attend meetings. Like that's actually coming, like that's actually coming back up, right? I can now think about a problem and do more of the work myself. I don't need everything extracted into, you know, a pivot table that I can play around with. And so, and again, this is like, you know, I think you're a baseball guy, like This is like the first pitch to the first batter of the first inning in this evolution, kind of, you know, early days. But it is, it is, I think, going to take away some of these like highly specialized, I do this thing when people ask me to do this thing kind of jobs. And so what I'm trying to do in addition to like, which is funny because we sort of specialize that one team, but it's breaking down how analytics is going to work with kind of the ops people. It's requiring folks to understand a much bigger part of the business. And so we're actually spending a lot of time on my team right now, um, sort of teaching what I call like general business knowledge, both like intercom, but also just sort of like general context, you know, stuff in a business. So no longer is the comp team the only team that needs to understand the comp plan. Everybody's got to understand it because It's going to affect, you know, these other parts of what we're trying to do and there's going to be new product launch and I need, I need more people than just me to be able to go, hey guys, we got to connect these dots because a lot of the, the just like turn the crank, grind it out work, I think is we're going to see steadily go away and get automated away.

Yeah. Yeah. And I don't know if the framing is like, it'll be less specialization, but I think the second piece, like less grinded out. Like less, I'm a person that creates dashboards, certainly, certainly goes away.

You, the tickets in Salesforce, like, yeah, it's funny you mentioned Cloud Code.

So my VP of RevOps and I were like texting all weekend because we were both cooking in Cloud Code on, on different things. And he was like, I think I can just tell Cloud Code what to build in Salesforce and it will do it. And he was, he was building into our sandbox Salesforce Salesforce from the Claude code terminal, like at a very high level of proficiency. And so you're like, okay, well now we don't need to do the grinded out, like Salesforce development. If you can just take, you know, the business requirements, have the business requirements get transformed by Claude into the technical spec and then implement the technical spec. And it's, I do think that. I, I very much agree. I think those, the nature of the roles change a lot. And so what people are going to do in these teams, I think should be much more strategic and important. Like I was just having a conversation before this with one of the RevOps team members who's been in our, she was a, one of our very first BDRs, then RevOps, and she's learned the whole RevOps craft here. And so she's going to move into like a more like cross-functional coordination type role. And I was like, hey, just let me be clear. If your role ends up being writing documents and just telling one thing from one person to another, we have failed. So like, the, the AI tooling should be able to do all that. You need to like pull the strings and build that AI system to find where all this information is, pull it together, put it in the formats people want. Like, you should be building the system that does that work, and, and the people like doing the work should be, those roles hopefully go away. People don't want to do those roles.

Like the ability to ask the right question suddenly becomes way more valuable. I mean, it's always been valuable, but now like that's the thing that's going to keep you employed and busy versus I could follow a process and do it with high accuracy.

With AI, is your team, because you mentioned 250 reps, 40-ish RevOps people or next, next year. Do you think that ratio of RevOps to quota-carrying reps is higher now? You have more RevOps people or less RevOps people than 18, 24 months ago?

It's, uh, we have more right now, and I think that's a product of the complexity and the change. There's a, there's a little bit of like, I, I've been thinking and talking about AI for us and 3 levels. There's like our product that we sell out into the world. Then there's AI-enabled tools and things that we're providing our sellers or that, you know, our, the engineering teams are using. And then there's like AI for RevOps and we like never have enough time to really do that well. So we're about the first 2 that, that the last one has, has not quite taken hold as much as I, as I hope it Does, and, and we're gonna push hard on next year. But I think that all that complexity and all these new things we've been, we've been talking about this afternoon have required us to like add more people, right? First, we, you know, we gotta go support this new solutions, you know, post-sales solutions team that we're, we're building first. That's more people to figure out one thing you need and, and ask questions and partner with them and, and do those things. And so the team is getting bigger. It's getting bigger to support all of that. We were talking about our, our, you know, deal desk team is adding people because we're pumping more, you know, more deals through the system.

Yeah.

The leader of that team and I, you know, we've been talking for a while now, like, what is, what is the discount? Yeah, I deal ask look like, you know, well, like, what, you know, what is that? So every deal gets scrutinized that way. Like, do you guys need to read the red lines, or can, you know, an agent read the red lines and tell us what looks problematic or not? And so I think we're gonna have this kind increase and then we'll start to see that dwindle as we automate these systems, like take advantage of more of the automation ourselves. Yeah.

I still think the— no, go ahead.

Like personally, my, one of my big like top 3 goals for next year is to build a much more scalable RevOps function. I sort of challenged them with things like, how do we grow our business at, you know, 40% and, you know, we'll never probably never get there, but like, What if we ask for no new RevOps people at the end of next year?

Yeah.

Can we build, like, can we build that engine?

And I think it's possible.

Yeah.

You know, it's just all the people, all the people doing the junior number crunching, grinded out work all have to move up one level of abstraction and AI has to do all that.

Yeah, that's right.

Which like, if you're in cloud code as much as we are, like, yeah, what, what can't it do? I spent a bunch of my time on the weekend trying to get Claude MCP'd into all of the places, into more places than I have it today to yank a bunch of stuff in and build like even simple stuff. Like go, I set up, I use Fireflies for meeting recording. And so I set up Fireflies to push everything into one big Notion page. So all the links of every call are in there. And so now Claude Code, I built a skill for basically just taking meeting notes from an entire week and giving me like all of the critical outputs, like the things. And so there was, you know, I kept being like, okay, like look for this and look for this. And like, what else should I look for? Like, okay, I said I would do that thing, but it's not in my task list. Like tell me those. Or we agreed to do this thing, but it's not in that other team's task list. And And, uh, you know, you just start to like think about those are people, those were people's jobs and now it can't be, well, it shouldn't be in my opinion.

Right. And then you think about the benefit that creates for you. Now you're spending more time in front of customers or with your, you know, with your team helping them get right on our call. That's just a, that's good for everybody. For me, it's now my team can spend more time on these big transformative projects for our, you know, for our go-to-market engine instead of. You got to move this account from this owner to that owner. Mark the ticket correct. We got to, oops, comp exception, like go run through that process. So there's just so much potential and it's really fun to be on like the very, very beginnings of what this can look like. Yeah.

From that Karpathy interview with Dwarkesh, he was like, computers allowed you to automate anything that is specifiable And now AI can let you automate anything that's verifiable. I might be butchering that, but like, you know, it's, it's moving away from the if this, then that sort of deterministic logic and being able to be like, hey, like, this is my, these are my rules for exceptions. These are my DealDesk parameters. This is what a good contract looks like. And it's just, just like dump it all in and give me an output.

Well, and you know, you, I don't think we should trust our lives quite yet to this, but like. It's way faster to review a list of recommendations on comp exceptions or whatever it is than go through everyone. It's, you know, you can spot check that, you can do a sample of it and say, hey, this thing's working right. And so, yeah, it just, yeah, these things up so much. As our, our SVP of sales here is like, hey, I use Claude Code and found $9 million of pipeline, like crying emoji, you know. Yeah, yeah, yeah. Just think, right, like that, we didn't miss that one in Salesforce. So it still makes mistakes and I think it's incumbent upon us to keep pressure testing and learning how to use it most effectively because it's only getting better. I love that. Yeah. I'm all like the, like, remember, this is the worst AI you'll ever use. Like, that's always true, right?

Yeah. Last week I was on a call talking through what I want our sales tech stack to look like. And so I was trying to send a PRD to a vendor to be like, hey, here's what I want. And so I dumped in the transcript and I told Claude. I was like, hey, like, you know, shape it in this way. I want like a one-page format, yada yada. And then I was like, wait a second, I didn't say that on the call. And I was like, I don't think that's right. And so Claude's like, this is concerning. Here's my detailed review of fabrications and inaccuracies. And there were 13 of them of things that it just completely made up. And Claude, I mean, I rarely have this. I really rarely have this experience now, but like, like, you know, when it goes off the rails, if that was completely running autonomously.

Yeah.

And I don't have the like observer agent set up for that type of stuff. This is just me ripping in Claude. But okay. One or two last questions before we get into the quickfire. How has the broader shape of the revenue org changed? Does this mean reps can carry way more quota on a per-person basis? Is it more or less BDRs? Like, what are, what are the impacts of AI, not in your product, but more like internally facing, uh, on the broader org?

Yeah, the, uh, I think we're, you know, maybe start at the top of the funnel, but to give it some structure, like, you know, SDR is going to be one of the first things that goes handling inbound leads. We have, we have it set up. We have, I think we have some You know, we've announced it at our big customer event. Um, so like that SDR agent is going to be one of the first things that goes where, you know, you used to have however many people sitting there dealing with a bunch of junky leads. And now that's, you know, that's an AI job. You see that go away. I think, you know, the, the more sophisticated outbound prospecting, like that job just gets easier. I think you guys were talking about this in your last episode of like, now we can send 70,000 really great emails. With one tool instead of 7,000 maybe, you know, passable emails. And I mean, I'm sure you get tons of these outbound emails. You're like, these people learn how to write, like, terrible, right? And you go like, wow, that, that becomes so much more effective, so much more interesting. The enrichment on that, you know, stops becoming like, hey Tyler, like, I knew somebody from Wisconsin just like you're from, Yeah, great. Now I want to buy your software. Like, you know, that's not particularly impressive, uh, you know, personalization. Like, but that's going to get better and better. And so you'll see those, you know, almost that entire part of prospecting, at least the initial stages of prospecting, I think gets automated away. We haven't seen that yet, but we're working there and there's certainly a bunch of companies out exploring that space. We're seeing a bit on, you know, productivity on the kind of classic, like Hunter AE, farmer RM in our business where they can, they could do a bit more. And then I think the big one for us is things I mentioned before, which is that post-sales like success motion. I mean, we've just had to significantly expand that team, create new roles for, you know, the most sophisticated, largest customers to properly service them and get them set up correctly. Um, and so that, so that's a narrative violation.

Everybody's like, oh, success is going to go away. But I, I hear this a lot. Success is actually more important than ever.

Actually, more— it's, it's, it's, it's probably both, right? It's, um, right. You think about like AI in the product to help you do things, right? It's like, uh, I don't know if you're old enough for that Clippy guy in, uh, yeah, off suite, right? Like, like actually a functioning version of that to help you be really successful in the product and, and understand your business context and everything you're trying to do, like that's a hugely powerful thing that oftentimes is now like a success webinar kind of thing. Like they don't need to do anymore. And so I think that does go away, but the intersection of the commercial and the product usage and sort of continued expanding application of this is, I think is going to be a bit of a human endeavor for quite some time.

Yeah.

Makes sense.

So last question, because I generally have like, steered away from just asking people about their use cases, but everybody wants to know. So like, what are the handful of like specific AI use cases that have been most impactful for you internally with go-to-market that are not Intercom? Uh, and like, feel free to mention the vendor, talk about like some of the, some of the things.

Um, so the, the one we're, we're building now that has, has kind of gone to fits and starts, which is, is every RevOps professional's the bane of their existence is the constant influx of, of what feel like basic and well-documented questions. ROE, how does money work? Uh, so we're, we're, we've set up Fin to try to answer those questions so they can leave us alone. So Fin integrated into a Slack channel or, or a Glean agent, just fire it away. Don't stop asking me, you know, what your base rate is. It can just tell you. Um, so that's, that's one place. And then we need The problem there is actually adoption on the sales team side. You need them to be willing to stop talking to us and willing to use that. So that's one, I think, use case that will continue to grow significantly. We have built, not my team, but another team has built what we call the cockpit for our post-sales teams, which is starting, it's funny, 8 months ago it was a Tableau dashboard that showed a bunch of product telemetry and, and kind of other things. And now it is an entirely AI-driven tool where you chat and interact with it and it tells you all about your customer and it can recommend, I don't know, for product usage, like specific use cases of Fin that they should explore. Like you've done this and this, therefore you should, you know, try this one next given the nature of this, this company.

So where have you built that? That's cool.

It's, I think it sits mostly on Snowflake. I don't know, actually, I did, it's not my team. It's a data science team that has built that. I'm like all credit to them for pulling this thing together, but it's amazing what it can do and it can, it just makes the life of a rep so much easier. And so that's been a really powerful one.

Are you mostly building your own AI tools?

We have a mix. Like we just switched to Gong forecasting to take better advantage of the integration the sort of call intelligence and forecasting. So we're really excited about the potential of that. We can stop building our own top-down forecasting model in a spreadsheet, you know, to pressure cut things. Like it's happening right in there. That lets us do deep dives into sort of, you know, deal sentiment. And, you know, when a leader comes in and says, we're going to beat our number because we're going to land this, you know, 6-figure, 7-figure, big, you know, big deal. We can just click right into it and go, you know, no, you're not. Have you listened to this last call? Or like, you sure you wanna, you know, you sure you wanna pin your whole forecast on that particular deal? Or like, you guys are sandbagging, you know, you got 3 of these, you're in great shape, whatever it is. And so that, that ability for us to quickly inspect, uh, sort of pipeline and then forecasting health has, has been a big one. And then we tried to simplify reps. Reps Lives, we have, uh, win.ai up and running. So that kind of helps populate MedPic fields automatically. And so we've seen, you know, completion rates on basic like Salesforce hygiene stuff has, has gone up. Uh, so we're using it in a bunch of different places that are sort of, you know, augmenting things that existed before. And I think we'll, we'll continue to push on that front as well as try to do some of these more, you know, not quite rip out a whole function, but like the SDR agent is going to be, you know, it's going to be a big one. That will change how we interact with customers right at the top of the funnel.

Very much so. What is the most important way that you're getting more done as an executive using AI? Like, what are your favorite personal productivity use cases?

Yeah, I mean, I think it's, it's, this is not sophisticated. Uh, I mentioned Cloud Code already, you know, just my ability to not have to sit down with my team to ask for something. And I'm just like kind of just getting started with that. but that's, you know, that saves me a bunch of time. That saves other people time. And so I'm really excited about that. The ability to self-serve on, you know, sort of data analytics questions where I'm just, you know, if I'm sitting around thinking about something, I wish I knew this, how quickly I can pull it up without hunting down a dashboard or hoping I, you know, hoping I get something right in Salesforce. Um, so that's been quite, quite nice. I think the, um, this sort of ability, like quickly generate documents, right? In my, my role, there's lots of like, write a memo, review a memo, you know, do like, do that kind of, kind of work, uh, with just ChatGPT is hugely helpful. Like the ability to just, I think I wrote this right, you know, pretend you're this person and read it and tell, you know, tell me what makes sense and what doesn't make sense. Just, it speeds things up just so much.

I'll send you my Claude skill. MD file for this to, to, I basically like put 150 pages of my writing into, originally it was ChatGPT. I'm like, tell me my writing style and excruciating detail, blah, blah, blah, blah, blah. And then it marries that with a bunch of context about owner. And that's now in like a Claude skill. Uh, so any, anytime I'm like, write this like me, I don't have to, 'cause I used to have a text expander snippet where I would hit Yeah, I would hit Z style and then it would be like, and my whole like how to write like Kyle thing, but just, you know, it's a lot of tokens and space, but it's pretty frigging great because Claude's so good at writing.

Right. Yeah. And I mean, you just like said that it's like you can send a draft to, you know, a totally trusted advisor who's going to tell you if it's garbage or not. Yeah. Or as you know, here's 3 bullet points, make it making a, you know, 4-page document in my style is hugely, it's like time-saving, productivity-enhancing. And so the, you know, so far it's those kind of things. I think as we get some of the, you know, like get our arms around some of these other use cases, we're just going to find more and more ways to like, do I really need to make the team meeting deck anymore? Like, can we do Eric? Can I just once a month, can I just ask it, you know, what are the most important things the team needs to know right now that they may not know? Like, make 10 slides kind of stuff. So I'm excited for those, those cases. And it's sort of fun to experiment and see what we can come up with. Totally.

Okay, let's move to quickfire so I can get you out of here on time. In your experience, what separates a good CRO from a truly great one?

Ah, um, it's a good question. I knew it was coming. I'm still not sure I have the perfect, the perfect answer. So I, I think a— I think there's sort of two things that I came out with. One, I don't know if you want to call it like balance, or when I was a consultant, we had this notion that people would talk about like being a porpoise, which is like you're kind of up on the surface and you're looking around and you can see everything, but when you need to, you can dive really deep. And so a CRO who has the ability to operate at the surface, and it's so sort of derisive to say, oh, the 30,000-foot view, but like you're kind of, you're looking around, oh, there's a demand gen problem. I can go over there. There's an SDR top of funnel problem. I can go there, I can swim to it, but then I have the ability to like get in there and inspect it myself. I, it's hugely valuable. You either, I think sometimes, and I've been fortunate that I've not had this problem, but like you hear the CRO who only kind of knows one thing. They know, I know, I really know how to do really big deals and I'm outstanding at that, but like demand gen and velocity, like I'm, you know, I'm, like kind of not my thing.

Toby from Shopify would call it being T-shaped.

Yeah.

Having, having a lot of like breadth over the top and then being able to go really deep on like one or two things that you're an expert at.

And so I think that ability, um, and it's, you know, it's a role where I think most people have kind of come up through the sales channel, right? Like that's kind of the, the typical background of a CRO, which makes tons of sense, right? Customer facing, you're carrying a number, Like it's, it's the right profile, but those people who can speak the language of everybody they need to talk to both within their org, right? So you got to be able to come to talk to me and have useful things to say and ask and an understanding, maybe not how to do the RevOps job, but of what the RevOps job should be and what good looks like and kind of what they want out of that partnership. I think it can be helpful if it's just me telling you stuff, like who knows if we're going to go in the right direction, hopefully yes, but helpful, but also like they gotta go talk to product and they've gotta be able to talk to, you know, product marketing and, you know, the finance team. And so, so there's a little bit of a, uh, like business generalist DNA that I think is, is really valuable. Uh, and so if you, if you bring those things together where like you're skilled enough, you're experienced enough in the, the part of the, the world you own that you can go deep and like, I can go rescue a deal, I can go, you know, have customer calls, I can tell people or think about prospecting what that looks like when it's great at scale. But also I can go talk to the product team and the marketing team and everybody else. I think is what separates the really great ones from somebody who's doing a good job.

Yeah. What's the most common advice you give to new VPs of RevOps or new senior RevOps leaders?

Yeah, I think at the end of the day, it sort of boils down to 3 words, which is have an opinion. Hmm. There is, you know, if you follow any of the RevOps or sales ops discussion, everyone, every third article is like how to stop being tactical and, you know, be strategic.

Yeah.

And it's a problem, right? The day-to-day firefighting sort of pushes you back into that. And so I really encourage people to find a way and this, you know, I do this with my own team and other people I talk to of like, you need to come with a point of view. If you're showing up in the forecast meeting to take notes, you're going to lose your job to Granola or whatever, you know, Gemini or whatever it is that is on that call very quickly. And so how are you equipping yourself to show up in that setting with a point of view? Maybe it just, you know, corroborates what everybody else has to say. Maybe it's a, you know, unique perspective, but the days of sort of sitting there waiting for people to ask you to complete something, I think are work on. And so if you want to really succeed, you gotta have an opinion.

Great advice. What's the hardest lesson you've had to learn?

Yeah, this, this is, uh, uh, probably a recurring one for, for me as well as many other people, which is you've got to give tough feedback to people early, particularly if you're inheriting a team, um, which I, I've done now repeatedly. Temporarily feels good to, you know, sort of explain away things you don't like or behaviors or performance that maybe isn't meeting it. It's, it's easy to rationalize and it feels like you're being maybe kind, but ultimately it's not kind. It ultimately sets you up to have to have a really tough conversation often too late. And I think I fall victim to this of like, oh, these people are new. They don't know me. If, you know, if I come in here and they quit, if they quit tomorrow because they don't like me, oh, I'm really in trouble. Like, Oh yeah, we're really in trouble. So I better be, you know, I better be extra nice. And that sort of develops a pattern of like, oh, now we're just nice. Yeah. And nice always doesn't, doesn't conclude with being nice. It concludes with, ooh, we got a big problem here and I haven't really told you about it. And now, you know, now we got to— yeah. So that's the, that's the one I've got to, I've got to avoid. And I think is, is a very easy trap for people to fall into.

It very much is.

Okay.

Last question. What's the best thing you've read in the last 12 months?

Yeah, the non-work answer to that, I read this book, Edward Abbey. He's like a naturalist writer. I think in the '50s he wrote this book, Desert Solitaire. If you spend any time in Southern Utah, he like spends time in Arches National Park before it was a national park and just like writes about his adventures and his experience. And he's pretty funny because he complains about all the tourists coming there and ruining everything. So that was a fun one. My family's big into hiking. That was a good one. And then the business one is, is I really enjoyed Ethan Mollick's Co-Intelligence book.

Uh, oh, I haven't read this.

Very, very quick one. Um, probably, probably horribly out of date now because it's, you know, a year old, but, uh, but it's a good one. He has 4 principles that I'm going to butcher if I try to repeat them. Um, but it basically encouraging people to embrace AI and, and bring it into their work and kind of how to treat it and how to think about it. And I think for me, that was a really good framework for how to approach a lot of the problems we've talked about today.

Nice. I recognize the COVID of this book now, now that you're saying it. Well, that's awesome. This was a great episode, Tyler. I'm really appreciative that we could bait people in with AI and actually talk about compensation because I have been excited to unpack it and I think a lot of good frameworks and lessons for people to learn. The transformation journey is really interesting and so this is going to be helpful. So thanks again for joining.

My pleasure. Glad to be here. Appreciate the time, Kyle.

Thank you for listening to the Revenue Leadership Podcast. If you enjoyed it, don't forget to subscribe and you can find a link in the show notes and be sure to leave a 5-star review, share it with your network, and please join me next Wednesday for another great conversation.