Community in the Age of AI

SJ Morris: Alright. So again, thank you so much for that intro. It was really nice. I wrote it. So and if anyone is feeling also conflicted about the tech industry overall or AI, I am happy to talk about those things, but I am simultaneously so excited about AI and about the tech industry these days.

I've been around since the eras of, you know, when APIs as products like, you know, Twilio and SendGrid and a lot of other companies started to get funding and started to become crucial infrastructure for all kinds of, you know, amazing apps that we know and love today. And so I haven't felt this excited about tech since that era, and that's saying a lot. So I wanna talk to you a little bit about how I'm sort of interpreting the potential for AI when it comes to building out developer community. I do work at HubSpot, and my colleague Nikita will be moderating the panel at the end of the day too, so it's nice to have some HubSpot representation here. For those of you who don't know, HubSpot is a CRM and business platform that businesses can use to manage all aspects of their business, really driven by the core data around their contacts.

So, you know, if you want to run your marketing campaigns, run your sales campaigns, run your operations and services, we facilitate all of that. DevRel sits on the side of allowing folks to build integrations, customizations. We also have a CMS offering. So we have a lot of front end developer community, which I'll talk a little a little in a moment. But we're doing a lot of interesting things on DevRel, and I think we're very excited and also cautiously, you know, careful about all all of these changes that are happening right now.

Navigating AI's Role in Developer Community

So let's get into it, what I wanna talk about today. Go back slide. So, essentially, what I'm trying to do is, you know, talk to you today about how we've been trying to navigate that intersection of AI and dev community. I wanna tell you a story about how we're resisting the urge to automate everything even though that's been a big ethos in HubSpot for a long time. The more you can automate, the more time you have to do really meaningful work.

But we're trying to balance that automation that AI can really accelerate with, the ability to allow folks to connect even more deeply. So our philosophy on DevRel I apologize. This slide got a little wacky. It's my fault for sending last minute slide edits, like, an hour before this, so I get to see everything. But our our philosophy on dev community is really that AI needs to amplify but not replace human potential.

So I'm gonna talk about what we're working on, what's still a little bit messy, but mainly our vision. I've put together this year and a half to two year vision. We're not looking like five year visions these days because AI is changing so rapidly. But I'm making a lot of assumptions that I hope are gonna be true that will help us balance this, you know, AI functionality, also AI enablement enablement as well as that human connection that's gonna be, in my opinion, more important than ever. So it's not really a HubSpot story exclusively.

What I'm really hoping today is that I'm gonna share a framework of thinking that folks here can learn from, adapt, and use as their as needed for their own communities, developer communities. And let's get started. So I'm gonna take a step back. I think most of us are pretty comfortable and familiar with, I guess, what I would call, for lack of a better term, traditional developers that are working in the code day in and day out. But we're starting to see especially I think all of us are starting to see this across our companies and platforms, but in HubSpot, we're specifically starting to see more and more business users starting to leverage AI not even just for enhanced workflows and, you know, quicker, automations, but also for generating code that does stuff on HubSpot.

And so we're in this really interesting space right now where we have evolving traditional developers, you know, these folks that we're calling AI engineers that are somewhere in the middle of these vibe coders on the business side and our traditional developers that are looking more and more at things like infrastructure, scalability, cleaning up messy code generated by tools like Lovable. You know, shout out to Lovable. I did build my website on them, I'm a big fan. But, and there's gonna be all kinds of emerging roles coming up in the next few years that we can't possibly anticipate. I'm hoping there's gonna be more and more AI ethicists, hence the internal conflict.

But, we're gonna start to see our our world shift quite a bit. But I am confident and hopeful that developer the term developer will still be a word, but we'll just have a broadened definition and a different set of sort of table stakes requirements. With that, I wanna talk about our strategy on dev community. It's really about less noise, less broadcasting, and more connection between developers building on HubSpot. And just as a caveat, I will be using the word developer to sort of encapsulate all of those audiences I shared in the last slide.

And I did wanna also add something. I don't know for folks who were at Ricky's talk this morning. My sort of like community minded heart loves the concept of wonder as being a defining factor of a developer. So if you're really into those moments of wonder and inspiring them in yourself and others, then I I consider you to be a developer. So that's just my aside.

So what we're doing is rebuilding our community space that we are running on a dev community team right now and really focusing it around connection. And that's something I've been wanting to do in various ways for years. And we've been kinda subject to HubSpot is a large company. We have, I think, almost 10,000 employees at this point. And so there's a lot of different inputs coming from different places into our community spaces.

The Dev Slack is something my team runs. It has existed for the last ten years, and it on paper, if you join it, you will see channels that have 25,000 plus members. Very deceiving kind of metric because the folks who are actually active in there are much much lower, just a percentage of that. And what we see is a ton of churn and turnover. Folks get referred to that space by our support team who are amazing but can't always handle the nuance complexities that our developers come in asking.

So they'll send them to our community spaces and well, what's happened is because we're a really small team, we have not had the ability to create programs and structures to really shape the the way that that Slack interaction is going. So you see walls and walls of unanswered questions, and it's a bad experience for everybody. We've been trying all these different approaches to fix that, but what we're doing now is really looking around how AI can help speed up and facilitate some of the some of the repeatable tasks, the ongoing support questions to make space for real contact connection. And we're looking at also helping our community grow in this AI minded future. So we're creating community programs.

We launched our ecosystem mentorship program this year, where we've paired a 100 folks with you know, folks who've been seeing a lot of success in the HubSpot ecosystem with folks who are trying to upscale and uplevel in their own journeys on HubSpot. We wanna lean into that even more and have more time to do that. And fundamentally, what we're trying to do is to change the definition of leadership in our community. I think historically, because of these sort of barriers to entry and the amount of noise, it's been kind of the squeaky wheels who have been the kind of most influential and loudest and the folks been able to get what they need from us. But we want to use AI to allow those sort of squeaky wheels to find and maybe interact with a bot a little bit more and less vocally in our community and make more space for the folks who might be historically a little quieter and really give them the chance to shine.

Before I get into the tactics specifically, I wanna talk about how we're gonna be, you know, measuring everything. And I would say the first set of metrics around community growth, I don't wanna be thrown off by the word growth. What I mean by growth here is really growth of the metrics that I think are most important in community. And the tried and true metric for the work I've been doing for the last twelve to fifteen years has been peer to peer connection. So as much as we can measure, non HubSpotters in our case because that's the way our community works.

In some cases, your company is gonna have your advocates in the spaces all day answering the questions and that's one model. And I think that works to a certain size of company, but at the point that we're at, for a two person dev community team, we don't have the time to answer those questions all day. And so we wanna make it so that our developer community feels inspired to help each other out. So peer to peer connection is a huge metric for us, and we wanna see more impact across groups. And when I talk about event attendance in our community, I'll also include, like, virtual events, in person events.

I don't know if anyone was at Erica Hansen's talk earlier today where she talked about building community with Google and and Shutterfly or Flutterfly. They're not Shutterfly. It's a different company. But she referenced the community commitment curve, and I wish I had that slide up here right now. But essentially, the premise of the community commitment curve is all of these different milestones that lead folks to in a community, if they're giving more time to a community, then they're more committed.

And that generally results in better business outcomes. And so to me, events like this are part of, you know, in our case is maybe an overall moment of catalyzing our ourselves around our community ident or our career identities and meeting our peers around this career. And so events for us would be things like, you know, let's get folks together around the mentorship program in person and talk about, you know, what it's meant to them, deepen those relationships, and it will just influence that kind of journey even more moving forward. I think I may have skipped a slide. Sorry.

Not used to these clickers. So let me talk a little bit about AI integration. This is a whole new set of metrics for us. What we're gonna be hoping for is this new tool that I'll talk about in a moment. His name is going to be Sprocky.

Our logo is a sprocket. And Sprocky's actually been around in HubSpot's history as a kind of clippy adjacent assistant in various spaces. But Sprocky is about to get a makeover and get, I don't know, AI steroids and become incredibly helpful. And so we are gonna be looking at how much Sprocky is being used, how many folks are getting unblocked more quickly, how many you know, like the accuracy of Sprocky. And fundamentally, two areas of measurement are gonna result in ecosystem impact.

So more retention both across developers and customers, more marketplace growth and quality, and higher developer trust, which is I think probably one of the most important metrics for development. And so I'm gonna quickly go through the phases and the timeline that we're looking at for this. Right now, we're in phase zero, and I'm here talking to you about it already, which is very ambitious. But really focusing on building out that Slack experience, improving it, getting that bot in there. And the bot itself is going to be let's see.

I think this might be actually the next slide. I'll go into that after. But really getting that, you know, tightened journey solid for our community space, our developer Slack in this case. From then from there, all these phases kind of bleed into each other, but we're gonna lean even more into our mentorship, scale it, pull out some specific AI themed, I guess, tracks for the mentorship program, do more events, and facilitate more peer learning. From there, we're going to actually, this is a call out to this community for folks who are building anything that would, like, integrate with Hubs benefit from HubSpot's data.

We've already had some conversations at lunch with some folks who are building out some agentic experiences that would benefit from CRM data. We wanna build more partnerships with those folks. So, yes, historically, we've worked with individual developers and some agencies, but we also want to start working with actual agents. So I think agents are gonna become part of our audience as we grow. And finally, you know, all of that is gonna lead to what I'm hopeful will be much more intelligent expansion.

So this bot that we're creating, we'll be able to roll it out into different spaces. It is a part of an overall kind of approach that we're taking to our docs. A lot of stuff is changing, in the next few months for the HubSpot developer experience that I'm very excited about, so all this stuff ties together. And what I'm hoping is that we can use not only AI and the bot to improve the unblocking of the of devs as they're building and leave more space for sort of, like, deeper relationship based connections, but also use the the questions that are incoming to inform our product and to, you know, inform the road map and allow us to have more real time insights on those things. Timeline, it's probably a little redundant, but we are hoping that our Slack two point o will go live by early September.

From there, we're gonna roll out an AI powered mentorship pilot and start leaning more and more into developer gatherings, async learning spaces, more virtual and in person events, and then really lean into those partnerships I was discussing in 2026 and so on. I wanted to show that there is a little bit of momentum we need to hit here at this moment because I do think that between now and a year and a half or two years from now, we're gonna start to see the establishment of foundations and expectations of how of this messy middle between these tools that are, you know, vibe coding tools and actual, like, scalable production tools and outcomes are gonna meet in the middle, and we'll we'll understand what that looks like more. And so I wanna deepen do a bit of a deeper dive into phase one. So we're really we're I'm really, really proud of the mentorship program. It's probably one of the things I'm most proud of that I've built in as part of the HubSpot developer community work that I've done.

We've been able to really build some tight connections with people that are very intentional and really focused on allowing folks to build careers. And so we want to take what we've learned from it so far, lean into a tool that we're using called Mentor Loop, which does have some AI functionality itself, some algorithmic machine learning tooling to pair folks together as well as more and more AI features in the tool itself. So it's you know, when I say AI, it's this very big thing, and I think in some cases, this applied to the tooling itself, and it's also applied to us facilitating learning around AI. We're gonna lean more into gatherings, and we're gonna lean more and more into those feedback loops that I think are kinda missing today on the scale that we need them. And from there, here's the team that's pulling it together.

We have some folks on strategy. That would be me. I have an amazing colleague. So we're a two person team like I said before. Her name is Jennifer Nixon, and she is leading our execution and operations of all of these great things.

And we have our extended crew. We have the developer advocates that can help out with a lot of this, benefit from a lot of this, contractors that are helping us build out Sprocky's functionality, and some of these partners that we wanna build the deeper relationships with. And finally, I'm sort of seeing Sprocky, like this is where I hesitate a little bit because by no means do I ever want AI to replace a human, but Sprocky's kinda gonna be on our team. We're gonna be leaning into this tool to really help us accelerate all of these other things. And then this is my sort of closing slide.

What we really are aiming for with all of this is the ability to sort of facilitate more innovation. You know, when developers are able to build cool stuff with minimal friction, the flywheel just spins and spins much more quickly. So my hope is that if we get the balance right of where AI is automating, AI is unblocking folks with answers to quick support questions, then it's just gonna open up the time to to work on the stuff that matters, And our community will improve. We'll have more ability for, you know, maybe typically quieter voices to to be heard because they're gonna be funneled into the relevant conversations at the right moments based on their kind of AI bot interactions. And finally, the the bigger vision is to have all of this drive career growth.

Career growth, of course, in our case, based on, you know, the skills that it requires to help customers be successful in HubSpot. But, ultimately, that's, like, you know, something that you can apply to your ultimate business goals as well. I don't know if we have time for questions, but if anyone We do have has time

Juan Pa: for a few questions. Okay.

very curious about: I am very curious about kind of like if anyone's been playing with tools with their community that have like, you know, resulted in embarrassing hallucinations or incorrect answers. I think that's some of the messy stuff that I didn't talk about. So if anyone wants to chat with me about before or after, that's great. Think he's off.

SJ Morris: Hi. Hi. That was a great talk. Thank you. I had a question about your community strategy around, you know, when customers are answering questions to other customers.

SJ Morris: How do you how do you

SJ Morris: deal with, like, moderation and, you know, someone giving, like, an answer that's not what you would want them to give to another customer and things like that? Like, what's your strategy?

SJ Morris: Yeah. That's a really good question. And I think it's one of the balances you need to strike and sort of make a little bit of peace with when you are driving customers to community spaces. And so that includes some caveats before they post. That includes some, you know, warnings that you're not getting the official answer, you're getting an expert answer.

We do have other spaces that facilitate community support at HubSpot. I specifically talked about the dev Slack today, but we do have community.hubspot.com, which is where the majority of support questions are funneled. And in those cases, we have pretty amped up champions programs that are there. But those spaces are largely for the the customer focused questions. Our sort of mandate on DevRel is to help support developers, and we've created these alternate spaces because there wasn't a lot of frontline support for or even community support at the scale we needed.

So we have a dedicated space for them there. But yeah, that's a really good point. Like one of the things we just did was extract an entire basically, what's the word I'm looking for, export of our developer Slack over the last ten years, did some anal analysis of it and we're able to identify like this no longer applies. This is true. We can use this to, know, support some content that is being developed now.

So it's an internal and an external thing. I think the caveats are really important when folks are being sent there. And we also need the ability, and I've been pushing for it internally, to be able to send folks to actual product experts when the community isn't coming through. Sometimes the community is not gonna know because they don't have access to every aspect of the internal roadmap. They don't know everything.

And sometimes questions start out I'm yapping now. But sometimes questions start out with one one sort of like the question is coming from one place, it actually is about a bug or the other way around. A bug is something else altogether. So it's a combination of like giving the caveats, being able to triage the questions correctly, and using it to inform our actual default answers that hopefully live in our docs and hopefully live in our knowledge base so that we're reducing the inconsistency. So not a perfect answer, very complex issue.

Juanpa: Hi. I'm Juanpa. And I think my question with integrating AI into communities is that it feels that sometimes that as AI continues to be developed, less people go to these community spaces to make questions that other people can benefit from. Yeah. And on the other side, as a developer relation, like, professional, you don't get those messages in in your radar as they're being asked to, like, AI.

So wonder how are you navigating like these two pieces?

SJ Morris: Sorry. Can you repeat the second part of the question?

Juan Pa: Second part is like if folks make questions like AI assistants Yeah. It's really hard for them to appear under your radar or you just don't see them appearing in your chats.

SJ Morris: Yeah.

Juan Pa: Which make it hard to know what are some of the barriers that people are finding. Yeah. So how do you navigate the balance between these Okay.

so sorry: These components? Why am I blanking on the first question? I had an answer for it two seconds ago. Can you repeat it again? I'm so sorry.

It's the first question.

Juan Pa: The first question is, as we integrate more AI

SJ Morris: Yeah.

Juan Pa: We see less questions and Yeah. People not going to communities to to kinda like find answers to to problems they might find.

really curious to: Yeah. So I think we might be a little unique in the sense that we aren't necessarily creating our community spaces for a lot a high volume of questions. What we wanna do is leverage AI to get folks unblocked on the simple stuff that they're coming in asking over and over again, and hopefully that will set the stage for more nuanced conversations that other experts will have opinions on and not be just frustrated like, oh, someone's asking the same question again. It is it is, I mean, an interesting balance. Like, I I'm really curious to see how it kinda comes together.

But the other thing is that when we're building out the Sprocky bot, we are gonna be building out ways to pull out frequent reports of themes, things coming in. That's gonna be a crucial part of building that out. So we may lose visibility over the specific question. Although I do anticipate a world where we can map those specific questions back to a user's profile somewhere. We do use a tool called common room for our community management.

I'm sure if lots of you know it. And so maybe there's some creative stuff we can do there. But I think it's gonna be about those themes and being able to more directly or indirectly rather address the needs of the community based on creating content, experiences, events around those things that keep coming up. So yeah, I mean, we don't there's so much we don't know. We're also very willing to take a hit on the volume metrics and really look at those quality metrics more than anything.

We good? Yeah. Alright. Thank you everyone.