#234: How HubSpot Is Reimagining the Entire Customer Journey With AI Agents

0:00Go-to-market problems haven't really changed that much over the last century. It's basically like, how do I build demand? How do I win deals? And how do I create happy, loyal customers who retain or spend more with me? These are age-old problems. I think what's really interesting about AI is it creates new solutions. Welcome to AI Transformations, a special series from the Artificial Intelligence Show. I'm Mike Kaput, Chief Content Officer at SmarterX and Marketing AI Institute, and I'll be your host. Every business's AI journey looks different. In each of these episodes, I sit down with leaders who have lived through real AI transformation, including the actual stories, the pain points that push them to act, the moment things started to click, and the results they can point to today.

0:50So join us as we accelerate AI literacy for all as part of our AI transformations. series. Welcome everyone to episode 234 of the artificial intelligence show. My name is Mike Kaput. I am the co host of the artificial intelligence show and chief content officer here at smarter X. Now if you've been following along with this series, you might understand that today's episode is a little different than our regularly scheduled programming. This is a special episode in a limited series we're running called AI Transformations presented by Google Cloud. And in this series, we are spotlighting how real companies are driving real change using AI.

1:36So in each of these episodes, we aim to explore how leaders at some of the world's most innovative companies are actively using AI to transform how their teams, departments, or even in some cases, their entire organization's work. We'll do that by looking at their AI transformation journeys, figuring out what sparked those journeys, how those journeys unfolded, and even in some cases, talk about what results they got and what got messy along the way. And we do that by interviewing leaders firsthand right here on the show, including the leader I have the privilege of talking to today, John Dick from HubSpot.

2:12Much more on John in just a minute. One quick note here, if you are a regular podcast listener and this series is new to you, do not worry. Me and Paul will still be doing our regular weekly episode of the show. Paul and Kathy are still doing their AI Answers episodes. Nothing changes except you get even more episodes of the Artificial Intelligence show thanks to this series. So with that, let's get into today's AI transformation story. Now, I want to note before we dive in, today's episode is brought to us by Gemini Enterprise. Businesses of every shape and size are turning to AI. They're learning how to move faster, do more, and improve their performance.

2:55But you can't just hand your data over to any platform. You need a trusted partner with years of experience. That's where Gemini Enterprise comes in. Gemini Enterprise helps you build sophisticated AI agents that can connect with your business data while also keeping it protected by world-class security and governance. That means no more compromising between performance and protection. You can get started with Gemini Enterprise today at cloud.google.com forward slash Gemini dash enterprise. That's cloud.google.com forward slash Gemini dash enterprise. Okay, so today's guest I am super excited to talk to.

3:40Today's guest is John Dick, Chief Customer Officer at HubSpot. HubSpot is one of the defining companies in go-to-market software. it began as a major force in inbound marketing. It's grown into a customer platform used by hundreds of thousands of companies across marketing, sales, service, and more. John leads HubSpot's global sales and customer success organizations, which gives him a front row seat to that full customer journey, including how companies create demand, build pipeline, close deals, support customers, and grow relationships over time. Now, today's transformation story is all about how HubSpot is beginning to reimagine the entire customer journey around AI agents and AI-assisted workflows.

4:30We're going to talk about how AI shows up across the entire go-to-market motion at HubSpot, and we'll also get into the organizational side of the story, because John is coming at this as a business leader who helps transform how the company as a whole thinks about AI and how work gets done based on what AI enables. And HubSpot's organizational structures for some of its teams have changed in some big ways as a result of AI. John, welcome to the show. So great to have you. Hey, thanks for having me, Mike. How are you? I'm doing great. Doing great. We were just chatting. Yeah, it's a busy week here, but there's always something new to learn, always something new to dive into.

5:08So I want to just get right into it with you. Let's do it. Awesome. So let's start at the beginning. What were the kind of core problems that HubSpot was trying to solve across the customer journey? And why did you all start to turn to AI to solve them? Yeah. Listen, the way I generally think about AI and I talk to our customers about AI is that go-to-market problems haven't really changed that much over the last century. it's basically like uh how do i build demand right how do i win deals uh and how do i create happy loyal customers who retain or you know spend more with me like these are like age old problems i think what's really interesting about ai is it it creates new solutions um and the things that I started to get really excited about as the chat GPT moment happened.

6:07The first thing was there have just been these historical constraints in go-to-market for a really long time. It was one of the first times that I felt like maybe those constraints could break. For instance, take sales reps. One of the most unmovable numbers in the last 20 years of B2B sales has been the percent of time that sales rep spends talking to customers. And every year, ambitious leaders say, we're going to move that number up. We are going to get to 36 % instead of 35%. We're doing it this year. And it just never moves because the amount of research, the amount of admin, the amount of coordination is so high when you are trying to execute on a complex sales process.

6:56And so that's a constraint, right? Is all this work. And for the first time with AI, I was like, wow, like maybe we could break that constraint. Like AI can do research. It can manage coordination. Uh, it, you know, it can do all of all sorts of things. So maybe we could actually start moving that number up. And it was kind of the same on the customer side. Um, you know, when the chat GPT moment happened, I was, um, you know, in our customer success org and like, we have a huge problem in like customer success, which is that we want to be able to offer really personalized, relevant digital help to our customers and say, Hey, you know, we can see that you're using HubSpot this way.

7:35We think you might get more value if you use it this way, but it is so hard to do that at scale in a deterministic marketing automation system. You of course can use personalization tokens. You can insert their name. You can insert their company. You know, you can build some like lists that are like a segment of people that you have a message for. But with AI, I was like, wow, maybe that constraint is broken. Maybe we could draft one of one emails for every single one of our hundreds of thousands of customers that offers them personalized advice on how to use HubSpot better. So those are the things that we got pretty excited about was just this idea of we could help more customers and we could help more reps.

8:16That's awesome. So it sounds like from pretty much the beginning of the kind of generative AI revolution, so to speak here, you know, ChatGPT coming out, HubSpot was on top of starting to think about this. So I wonder if you could maybe walk me through some of the story. How did HubSpot's AI transformation in this area get started? How did you progress? And how is AI now being used across HubSpot's customer journey today? Yeah. You know, over the last three years, we pretty much have like systematically rebuilt how we attract, engage, and delight customers. and have really built a pretty new agentic go-to-market model across that whole customer journey.

8:57First of all, I would say it actually started with a desire to get our employees to be AI fluent. And every company has to figure out how to do this. I think there's a lot of advice out there about how to do it. My advice on how to do it would be a couple of things. And these worked really well at HubSpot. The first was, it needs to come from the top. This needs to be something that the CEO cares deeply about, talks about, everybody in leadership cares about, talks about, and not just, you know, cares about and talks about, but actually does. Like, you know, one of the more impactful things to drive AI fluency at HubSpot was very early, you know, our CEO shared a video of how she essentially used AI to do certain tasks in her day.

9:40And people were like, oh, wow, like if Yomity is doing it, like I really should try it, you know? So I think that that's kind of the starting point is like, it has to be important. Um, you know, the second thing is you've got to carve out space. Um, you know, and I think that, uh, you know, having actual dedicated time to do it really matters. Um, and a third thing is, I think you have to create a culture of, uh, like no fear sharing. Um, one of the things I've observed with AI, and I think everybody feels this a little bit is like, everybody feels like they're, they're behind and like somebody else has figured out something that they haven't.

10:11And it moves so fast that like, it's exhausting, right? You're like, I like try and keep up with everything. And you're like, still, you're like, I'm behind on that. Oh my gosh, I'm like terrible at using it. And so you've got to create this culture where it's like totally okay for everybody to share. Like we did it through, you know, shared Slack channels and, you know, where people could post wins, examples, whatever, really inspired people. So I think it started with just a desire that the company was going to be AI fluent. And as the company moved to be AI fluent, it allowed us to actually not just think about how we, you know, operated as a team, but how we actually built the customer journey.

10:46Customer journey, of course, the first place we started was support. You know, I think that was the first place where generative AI had super clear product market fit. We were pretty early to that. We had a lot of lessons learned in doing it. I think, you know, everybody got really excited about resolution rates and companies still talk about resolution rates all the time. I care way more about CSAT, honestly. And I think that was like an early lesson is like, wow, look at the resolution rate. And meanwhile, CSAT was like, so, you know, we've really retilted that and it turns out, you know, if you build great CSAT, your resolution also goes up.

11:22So it's kind of like, it's not like one or the other, but you know, we had, we had some lessons on that, but that's where we started, which is kind of in theory at the end of the customer journey, because it kind of had the best fit. Our marketing team got pretty amped about it for content creation, as you would kind of expect and rebuilt a bunch of workflows on that. And those are really the, and of course, coding on our engineering team. But that's not really core go-to-market. So those are the places we started. Okay, cool. So it sounds like HubSpot starts by leaning into this AI fluency, AI literacy.

11:55And then from there, there were some very clear use cases, workflows, what have you, where AI had great product market fit, including coding, support, and then some marketing workflows. Maybe take me through what the next step was there because I know we've talked about more recently, HubSpot has started to get into AI for demand creation, AEO qualification. Could you give me a sense of kind of how that's evolved beyond these individual use cases? Yeah, so we pretty much just looked at the entire customer journey and we said, what are the constraints and how can AI break them? And so it honestly, like, you know, the number one thing I talk to our customers about right now is actually how much demand generation has changed.

12:40And, you know, you noted that HubSpot kind of got our start by pioneering inbound marketing. It was really revolutionary for so many companies, including HubSpot. And inbound marketing has changed so much over the last couple of years. Not only has search traffic declined a ton, but there are new places that people are searching. They're searching and answer engines. And so there's a whole new type of marketing. That was kind of the first thing that we kind of did there was we built out an AEO strategy. We built agents to basically help us execute that AEO strategy so that we could show up better, get cited better for the things that really matter to our buyers.

13:23And that was kind of the first thing on the demand gen side that we did. Another big thing, though, that we did was we did a lot of focus on, like, how do we get the right TAM added to HubSpot for our customers? You know, and essentially we have like a demand agent that allows us to, you know, go out, set criteria, find companies that we think would be really good fit for HubSpot, get the right intent signals for them, et cetera. So that's kind of the second thing. And then, you know, the third thing, of course, is, you know, converting buyers who do show up on our website and who do show interest.

14:01You know, that's kind of our, you know, AISDR, our, you know, our sales bot. And, you know, unsurprisingly, I think our sales bot has done quite well. Over 80 % of our website chats are handled by the bot. The results are very, very good. our AEO agent has grown our AEO conversions dramatically, almost 2 ,000 % growth over the last couple months. And we've added, I think, the best version of our understanding of our TAM that we've ever been able to kind of into HubSpot to help our teams know who to focus on. I love that. So it sounds like you've actually, correct me if I'm wrong, been deploying agents to almost every phase of this cycle, from AEO to further demand qualification and to the actual strategy in TAM itself.

14:53Yeah, and this is what I actually advise most go-to-market leaders to do is just draw out your customer journey and the steps that your team do to show up in that customer journey and move people along it and then literally build a strategy for each of them. And of course, you have to prioritize. You can't do it all at once unless you have a huge amount of resources, but pick the ones that are the biggest issues for you. like if one of your top issues is like you um you know don't have enough sales reps to um you know get the pipeline generated that you need you know one one option of course is you can go out and you can hire a lot of bdrs and others you can deploy a prospecting agent to actually go and do the bdr job for you um you know and research and monitor the signals and all that um and so i i really do like i think the easiest thing to do is like write down on a piece of paper uh like or describe it to to Claude or whoever, here's what my customer journey is.

15:46Help me map AI solutions against each of those. Nice. I love that. That's a great way to look at it. So to that point, as much as you can kind of publicly share, what tools or agents or systems are actually powering this? Like how much of this is HubSpot native? How much involves other platforms or models? Could you just kind of give us a sense of that at a high level? Yeah. A lot of it's HubSpot. but a couple of things I would say about, you know, HubSpot's AI versus the HubSpot go-to-market team. You know, one really cool thing about working at HubSpot and probably the coolest thing about working at HubSpot is that everybody at HubSpot loves go-to-market, including the product leaders and the engineers.

16:26Like you can have like the, when I joined HubSpot, like the greatest thing was like sitting down with like the, you know, the postmaster to like talk about email deliverability. I was like, this is awesome. Like I'm learning so much. And so, you know, one of the things that we often do at our, uh, at our company is that the go to market team actually, um, innovates out in front of the product team a little bit, um, and, uh, actually tries to like innovate solutions. And then, um, oftentimes we do them in partnership with the product team and the product team then kind of builds them into solutions for our customers.

16:57Um, and for some of our solutions that, that was definitely the case. Um, you know, at this point though, we are using a decent amount of the HubSpot stack for our core, um, you know, jobs to be done across that customer journey from AEO to finding our TAM support, et cetera, our prospecting agent, all that. So the HubSpot stack has gotten pretty good at delivering agents across those journeys. Beyond that though, we have experimented with almost everything and we'd like to keep an experimental culture. Our team members have access to lots of tools. If anything, there's probably like a little bit of tool sprawl.

17:36That's something I hear a lot from, from our customers is kind of like, yikes, like people are doing all sorts of stuff. I don't know. I am not measuring it. I'm not governing it. I'm nervous about it. I don't know what's going out. Big reason people actually want to build their agent to go to market in HubSpot is for that governance and control and compliance and all that. So yeah, we use a wide range of tools. The other thing I would say though, is that, you know, we you know, HubSpot product is this way, but also HubSpot go to market teams We look for the best model for the best job. And we do think that's something that HubSpot can really help our customers with is being sure that if you're using a HubSpot solution, it's able to orchestrate across different models today to help with the best results.

18:21Though in the future, probably also to help with costs and all sorts of other things that our buyers really deeply care about. Yeah, it's a huge issue, it sounds like, recently for many, many organizations, is which models to use, how to regulate usage. That's awesome to hear. So a big part of, John, where I wanted to focus this conversation with you is based on your specific expertise and background, you're coming at this as a business leader, trying to transform how an organization thinks about AI and how the work gets done. And I'm curious, as you have gone from these isolated use cases to actually deploying agentic AI across many phases, if not all the phases of the customer journey, what has had to change organizationally inside GoToMarket team, inside HubSpot as a whole, as this transformation has progressed?

19:13Yeah. Well, I mean, I think the way that we build AI solutions has changed for our GoToMarket team. That's probably the biggest place we've seen change. Okay. You know, when we started, it was Wild West. Everybody, here's some tools. Have a go. Share what you've built in a Slack channel. We hosted tons of hackathons. I mean, it was awesome. It was like really fun. It like unlocked so much creativity and ingenuity and whatnot. You know, but we pretty quickly realized that if we were going to make like, you know, something we think a lot about is the framework of moving from like individual productivity to institutional productivity.

19:50Okay. And really like obsessing about this idea that, you know, AI, of course, can help with individual productivity, but like the real unlock for businesses is if you can get institutional productivity, you can meaningfully move, you know, items in your P &L or growth results or whatever. And so we pretty quickly realized that if we wanted to get to institutional productivity, we needed a different way to build. And it's taken several evolutions. The first evolution was that we built a bunch of pods. So we took, you know, people from different teams, including subject matter experts and engineers and data people and whatever, and we cobbled them together into some pods.

20:26And we put all of them under one leader who is kind of an AI-pilled, you know, marketing growth leader, Kieran Flanagan. And that was awesome. We did that for about a year, and we got really great progress. We moved all the KPIs we were working on, and it was great. you know but we realized that we could probably go further and go faster if we went even further and the pod model was great but in the pod model they're still competing priorities and it was a lot of coordination cost and and whatnot and so we actually made a pretty big change to how we operate you know about you know three months ago and we had our all of our you know essentially engineers, systems people, data people, um, you know, uh, product people, uh, SMEs that work on go to market agentic work all actually, you know, um, direct line in now to, to Kieran Flanagan.

21:25That was a big, it was a big change for us. Um, and so, you know, he really now is able to, um, execute at a higher pace. Um, and you know, there's always, I always get questions about like, oh, like, you know, don't, aren't you nervous? And I'm like, well, no, because I'm like super aligned with him. And, uh, like we have the same KPIs. So, you know, as long as we're all running together, um, you know, but I think sometimes people get nervous about like losing quote control. I think what I would argue is that in AI, the pace that you can and should move, it is worth finding someone who can really lead that, um, and, and letting them run with it.

22:03because I think you'll just get better results in a meaningful way. And Kieran's doing an awesome job. He's great. Yeah, I'd love to explore that a little further. It sounds like that this centralization or unification of these teams under one leader in one function, the benefit there is just the ability to move and react faster. How has that played out? And how is that materially different than the pods model or even just kind of the Wild Wild West model? yeah i mean i think it's about the ability to commit to bigger uh bolder goals more than anything uh and you know when you're running a pod model you're like i i think i can do my best here but like i don't actually like these resources aren't on my team and like you know uh in some cases it's like competing priority it just cleans everything up um and it also just allows for a lot more fluency like i think we're just learning like the skills that are needed continue to evolve and change, you know, just like I was like prompt engineers to, you know, go to market engineers to like engineer engineers to, you know, and so like to data scientists, you know, MLEs, like that whole world.

23:12And I think having all of those people on the same team allows for a lot more fluency in that stuff. We also think it's great for talent acquisition and retention, which is that, you know, those people feel like they are part of an organization where they're working with some of the best people in the world on these problems. I think this is how the best teams built marketing for a while. But we used to have this big debate about how we organized our SEO teams at HubSpot. SEO was huge to HubSpot. We work in many different languages. You've got regional marketing teams, of course. And should the regional marketing teams own the SEO strategy for their region or should the central team?

23:53And how do you make that decision? And one of the core ways that we made that decision was that you looked at whether the domain expertise in the topic area really, really mattered. And for something like SEO, it just does, right? If you were world-class SEO, you got a legitimate edge over the rest of the world. And number two, whether or not having that team be centralized would result in hiring people who cared more about the craft than others. And I think those are the benefits you get out of that. And model is, we feel like we can hire and retain really, really great, awesome people that want to be the best in the world with this stuff.

24:39That's awesome. I'm curious, as you're going through this evolution, it sounds like pretty recent last three months, kind of centralizing these folks under a single team or function. What has been in your experience a little messier or more confusing or harder than usual as we go through these kind of growing pains? I mean, the hardest thing about it is always trade-offs. And it's about, you know, local priorities versus global priorities or, you know, and, you know, I think one thing that the centralized team really ensures is that the global priorities are the priorities, which I think is actually a very good thing, because focus matters.

25:23And, you know, we know that in AI in particular, if you do a lot of stuff, mediocre, you get pretty mediocre results. And if you do get really focused and execute really well. But there's all sorts of pockets then where you're like, I wish I could kind of like play around over here. And you know, and it makes it a little hard to do that stuff. So but I think that the net on that is, you know, much more positive than negative. That's awesome. So it sounds like at every stage of this journey, both from kind of the individual exploration to the pods model to the more centralized model, it sounds like you've been getting increasingly effective results across kind of go to market across the customer journey.

26:01I'm curious, you know, you shared some numbers, some impacts, any other results that are coming out of or that you're starting to see in the early phases of this more unified model? Yeah. Um, well, I, I shared the, the data on the marketing side. I didn't really go into the engage or the, um, the, the, you know, the delight side, um, you know, but, uh, on the engage side, I mean, the biggest thing we've seen is we've seen win rate go up with, uh, AI, you know, basically we have a, um, an AI assistant called guided success, um, which is built in HubSpot's breeze assistant. Um, so we have an assistant in, um, in HubSpot and it's, there's not a box assistant, but there's also custom assistants.

26:42It's a custom assistant that, um, we built and tuned really for the job of the sales rep. Um, it relies on a whole bunch of things, um, all the context on how to win deals at HubSpot. And, uh, when reps use that win rate is up a bunch, um, similar with prospecting, uh, we've seen, you know, a lot of reps want to build their own like prospecting agent and they'll, they'll build a gem for it, the, you know, a glean agent or whatever. Um, and because, you know, they kind of like, number one, it's fun, uh, it's fun to build. Everybody wants to build. Number two, they're like, I think I've got a good way to do it.

27:16And I want to, you know, try my way. And I think what we have pretty consistently seen is that the, um, the stuff that has been globally built and really tuned with the right context and the right, uh, you know, evals and everything just really outperforms the, you know, the stuff that individual reps build. And so, and we see that, so we see that with our prospecting agent and everything, you know, and our prospecting agent, basically like, I think the number we had from last quarter was like book 10 ,000 meetings. So yeah, real scale. Yeah, no kidding. Yeah. And on the support side, I said, you know, we're industry standard resolution rates, you know, 60%, CSAT is high now on that experience after a couple months of chipping away at it.

28:04And we see that our CS reps also have an assistant that is able to trigger a bunch of agents for the work that they do. The challenge, of course, with any CSM team is how do you spend as much of your time with customers because it's just, you know, I'm supposed to get a lot of customers. And so we see when our CSMs use the CS Assistant, save rates go up, got a seven point increase in save rates, better usage, you know, all the metrics that you'd want to see. That's awesome. And you mentioned something so interesting there, which I've heard others echo, which is HubSpot's context at that global level is what seems to be able to help get much better results out of just the technology versus someone building on their own.

28:51So it really, I mean, HubSpot is such an advantage there, it sounds like, I mean, just because you've had so much time in the market, you have learned what works and what doesn't the hard way. And now you're able to train agents to then help new people learn that quite quickly, it sounds like. Yeah, I mean, we think so. And we've seen the data on it. And so we've seen our internal data, but also, you know, data from our customers using our agents. You know, HubSpot has a ton of context. And that context comes in a whole bunch of forms. There's just like, you know, the context that every customer has because their own data and workflows are already built out, which, of course, like, you know, I've never seen so many customers kick off data cleanup projects as, you know, 2026.

29:37Everyone is like, okay, now's the time. But that's obviously a big part of it. But to your point, we also just, you know, we're able to, you know, not just eval our agents that we build for our customers on, you know, standard metrics, but on go to market outcomes specifically. And those agents have access to the context on go to market. Like, yeah, we like to think that, like, the reason you would want to talk to somebody at HubSpot is because we know a lot about go to market and we spend our lives thinking about and developing solutions for go to market. And we really hope that our agents do the same way, that they have an advantage because they know a lot about go-to-market and have seen a lot of go-to-market in practice.

30:20So, John, honestly, I could talk about this for hours here. But as we wrap up this conversation, hopefully the first of many, it's been really cool to hear these real details of how this transformation journey has gone at HubSpot. So I'm curious for someone in your type of role listening who is a little earlier in their AI transformation, what is the best advice you could give them based on some of the things you've learned the hard way? Okay, my top advice right now would be the following. Number one is you are not behind. Everybody thinks they're behind. You're not behind. Everyone is trying to figure this out still, including companies like HubSpot.

31:03We are all out testing, learning. We learn every single hour of every single day, what is working, what's not. And so hop on in, uh, the water's warm. The, uh, second thing that I would advise people, um, you know, who are newer to this, um, is, you know, pick something to focus on, pick a real problem for your, your business and focus on trying to solve that problem versus of doing everything at once. And, you know, the third thing that I would probably say is just set a really high bar for quality. You know, I think, like, I hate slop. It's like, it's so terrible. Like, you see it everywhere, you know, and it's like, we really deeply believe at HubSpot that like the power of AI is not just AI doing everything on its own.

31:51It's like, you know, Human authenticity plus AI efficiency is like the secret sauce. And so set a high bar, have your people be part of it, not separate from it. And if you want to get started and go to market AI and you're a HubSpot customer, get started with HubSpot AI. It's got context that will help you get out to a faster start and higher quality outcomes, I think, sooner than you would if you kind of built everything from scratch on your own. Well, John, thank you so much for sharing today. This is an incredible transformation story. We're going to absolutely be following along closely. It sounds like HubSpot's doing incredible things.

32:30And this is just one of the many AI transformation stories we're hoping to share with you all through this series. So we appreciate everyone listening in. John, thank you so much for being here. Really appreciate it. Yeah, thanks, Mike. Good to talk to you all. Thanks for listening to our AI transformation series from the Artificial Intelligence Show. To keep learning, visit smarterx.ai, where you'll find on-demand courses, upcoming classes, and practical resources to guide your AI journey. Until next time, stay curious and explore AI.