DevRel's biggest stage yet

Angie Jones
Angie Jones
Global VP of Developer Relations at Block
DevRelCon New York 2025
17th to 18th July 2025
Industry City, New York, USA

Amid industry-wide anxiety about AI, Angie Jones argues that developer advocates are already perfectly equipped to lead this shift by using their core skills of learning in public.

She shares how her team successfully pivoted to an AI product by being transparent with their community, admitting what they didn't know and positioning themselves as fellow learners.

This approach shows DevRel’s true value is in guiding people through massive change with authenticity, shaping the culture far beyond a single product.

Watch the talk

Key takeaways

  • πŸ™ Embrace vulnerability with your community Be honest about learning AI with them to build trust and encourage open collaboration.
  • πŸ› οΈ Focus on practical AI use cases Show developers how AI improves their daily tasks, as these real-world examples are more valuable than flashy demos.
  • πŸ”Ž Adapt to new content discovery methods Recognize that developers increasingly use AI chat tools to find answers, shifting focus from traditional SEO.
  • πŸ§‘β€πŸ€β€πŸ§‘ Broaden your educational reach Apply DevRel skills to teach non-technical teams, expanding your influence and becoming a central guide for company-wide adoption.

Transcript

Angie: Thank you so much. Just a couple of years ago, people were asking if DevRel was dead. Y'all remember that? Yeah. You heard it, right? Yeah. And yet today we're standing on our biggest stage yet developer relations is leading one of the most important and influential shifts of our time, and we're using our core DevRel skills to do so. I want to tell you a personal story. At my current company, I built a state-of-the-art, art DevRel team. We were everywhere doing all the things, living our best DevRel lives. And then for many tech companies, nowadays, priorities began to shift and some tough decisions were made. Unfortunately, the product that we've been working on for the last three years and pouring our blood, sweat and tears into had been sunset. And I'll admit that this was painful, right? This was a hard pill to swallow because we had given our all to make that product successful.

And now that it was gone, we couldn't help but wonder, what does this even mean for us? Fortunately, instead of letting us go, the DevRel team was asked to pivot to a brand new project. Goops, an AI agent. We were super grateful for this opportunity. We quickly pivoted and we were determined to hit the ground running. That ability though, to pivot without losing momentum, that is something that developer advocates already excel at. Yes, right? Adapting to new technologies, to new audiences, to new priorities on the fly, we do that. But there was one teeny tiny little problem. This time we didn't know much about ai, at least not this new AI where developers are using LLMs to basically build AI features into their products, or they're using AI agents to build whole applications. Higher role had just changed. How were we something that we didn't know ourselves?

Had we taken on more than we could chew? Of course we did, right? That's what we do, but that's never stopped us. So after that brief moment of panic, I realised that we weren't alone. Nobody knows this stuff, right? It is not just brand new to us, it's brand new to the entire developer ecosystem. Someone has to teach the masses. Why not us? And then suddenly I realised why they kept us around. They asked us to stay at the company because we have the exact skills that are needed for such a time as this. A time where no one knows everything, not even the experts, but someone has to be willing to figure it out and bring others along for the ride. So going first, figuring stuff out, guiding others, that's literally the job description of a developer advocate. We've always done that. And as AI is redefining everything at lightning speed software development as we know it is changing right before our very eyes, if you are not learning about how to develop software, then you'll soon be out of touch with the very community that you were hired to serve.

And I say this out of love. I know that there's a lot of hype around ai. There's a lot of noise. Sometimes it feels like a fad, but come on, you and I both know AI is not going anywhere, and I don't want my fellow developer advocates to be left behind, especially when developers look to us to lead the way. They'll never admit that, but they do. You know it. And fortunately, it's still early. And guess what? You already are well prepared for this challenge. So I'm going to share with you how my team used the skills that they already had to gain these new skills needed. The very first thing that we had to do was to be honest with our community. Five minutes ago we were talking about payment protocols, and now we're all in your face talking about AI that looks a little suspect.

So we had to make sure that people didn't confuse us with one of these grifters out here who's trying to promise you AI pipe dreams that they can't deliver on. So instead we were honest, we said, look, I don't know ai, but I know I need to learn it quickly. And so I've helped you before when you had to keep up with the never ending JavaScript frameworks. I was dead when you were told that everything had to move to the cloud overnight. I was there when containers suddenly became the only way that we could ship software. I was there and I'm here now to help you. Again, we didn't position ourselves as the AI experts. This is important. We position ourselves as someone who's learning with you and is going to share our discoveries. Don't get it twisted this vulnerability. This doesn't mean that we show up with a lack of confidence.

No. We show up confident that we can figure it out together and we'll bring others along like we've always have. And this invites others to admit that they're learning too. And this wasn't a new tactic that we picked up. And DevRel approachability is a superpower that we already have, and this serves us quite well in this new uncharted territory. That balance of staying approachable while still being credible, that's our lane, is it not? Yes. So my team gets questions all the time and discord on social media at events, and many times we have to say, I don't know either. Let's figure it out together. And this presents this wonderful learning opportunity for us as well. People are bringing us their real life use cases and they're saying, help me figure out how do I apply AI to this? And because this is done and public spaces, others feel safe bringing their questions as well, or even jumping in to help fellow community members, I learned so much from our community members who are also experimenting as well.

This is very, very powerful. This is going to disarm the scepticism. This is going to instead build a stronger, more open community. And because of this approach, the GOOSE community, I must say, is one of the most welcoming and helpful ones out there. That's not just my opinion. Community members tell us this all the time because in an industry where everyone is posing as an expert, it's refreshing to see folks who are vulnerable and honest. This stands out as being authentic and trustworthy. Now being vulnerable is great, but we also got to deliver some value. So here's how we quickly got up to speed so that we could do that. In addition to leveraging the community to get real use cases, we also heavily use AI ourselves. So you can't relate to your developer audience if you've never walked in their shoes.

So we tried to stay one step ahead. Oh, that was funny. And we do that by building. So we use AI to vibe code products ourselves, demo apps or real products. We use it to learn about code basis. So if a community member is coming in there, how do I do X, Y, and Z? We can use AI to quickly figure out what their code is even doing. We use it to review new features of our own products. And so you know how it is. No one tells you anything. You learn with the masses about a new release and you got to quickly figure it out. So AI helps us with that as well and even helps us write the docs. So I got this new feature. Let me talk to Goose about it. Yeah, I understand it now. Hey, can you give me a draught on the docs?

And you tweak it. Don't put it out just like that, right? You tweak it and make it human. We use it in CICD pipelines and containers. We use it to write tasks to set up environments. That's my favourite use case. We use it to manage GitHub projects like organising and prioritising the backlog. So we get a lot of contributions from a lot of different people, and so our backlog is a mess. Help me organise and triage these issues. Maybe take a first pass at some of the prs. We also use it to gather and analyse community feedback and sentiment, and then we can take that see patterns, and we share that with the product and the engineering teams to help influence the roadmap. So these are the fundamentals of software development and everyone is trying to wild datto birds with these flashy demos of things that they don't even need.

We decided to focus on the day-to-day developer tasks and to show them how AI can truly take care of some of the tedious work so that they can be more efficient and showing developers how to apply tools to the fundamentals of their daily work. Another dev re special, we post content about how we use AI for these everyday tasks. And let me tell you, this content does better than any of the extravagant, flashy demos we have what people are hungry for. They want those practical use cases that actually help them to be better developers. We also found that the developers hard to tell you, they don't trust your little shiny demos anyway. They don't believe that you one shot it a complete application in 30 seconds. They want to see the full, real messy process of interacting with ai. So we do live streams where we build stuff as the community.

We don't shy away from the failures. AI is non-deterministic. People need to know this, right? They need to truly know what to expect when they use these tools. They need to know how to iterate when something goes wrong, how to recover from that. When AI fails, they need to know how to protect their code from ai totally messing it up, and they'll never learn these lessons if we keep them in the background as we're trying to edit the video. We're robbing them of this. And this goes back to trust. Don't lie to your community. Don't try to pull a wool over their eyes. They can see right through this. You know that and you're going to lose them and you're going to ruin your reputation and your credibility in the process. But we already know that authenticity beats perfection, does it not? That's why we've always built in.

We've always shown our missteps and we've earned trust by being real. That hasn't changed. In fact, AI makes that even more important. All right, let me get off my soapbox. I'm going to have to take a collection after this, Jon. All right, so we've earned their trust. That's great. How do you actually reach them? Because developers aren't discovering your content the same way they used to. The traditional sites like Google and Stack Overflow, they're being used a lot less by developers to ask those technical questions. They're increasingly using things like chat, GPT and similar tools to find their answers very quickly, right? It's tailored to their context. They get the answer just like that. And this means that the way they're finding your content is changing. SEO is all but dead. I'm so sorry. I can't remember the last time that I searched for something and I saw one of your blogs in the results. Can you remember? No. Right? Social media is an absolute mess, right? You post, you got a hundred thousand followers and you got two likes on it. They won't show it to anyone. So we're trying to trick the algorithms. We're putting the link in the second and the third post hoping that someone sees it.

So you may be wondering, is it even worse creating content anymore when it feels harder to reach your audiences? But I promise you that high quality content, that authentic content, that still matters very much. So there are new models that are being trained all the time, and y'all, they've already consumed the entire web as it exists today. So they're hungry for some new fresh content. And now with the real time search capabilities, they're even better. So even if LinkedIn refuses to show your posts to your followers, that hard post that you wrote, it was so thought provoking, you knew you were going to kill 'em, right? The will gladly point folks to that. Often, that's where I go now. I want to see blog posts. I love reading from humans. And so I go in chat, GPT, I ask it. I say that specifically. I don't want news articles.

I don't want all of that. I want blog posts from developers and it'll give me a list of those on a specific topic that I asked for. Same with videos. I'm often watching videos now, right in my chat GBT sessions. So this is how they're going to find your content now, but what that means is DevRel leaders that we are going to have to adjust our expectations about engagement metrics, right? The views and the clicks, they're going to face a decline they already have, honestly. But that doesn't mean that the content is not impactful. In fact, if you just have to track those clicks, look at the ones where the refer is chat GPT, and you'll see that it's actually bringing you your future community members. Now we already know and DevRel how to measure success beyond those vanity metrics like clicks and stars, yes, community health, trust, influence.

These have always been at the heart of our work and that part has not changed. And you know what else? Helping the community confidently navigate this change, that's extremely valuable too, and that's not going to show up on our dashboards. Now, one thing about AI is that it enables everyone to become builders. Suddenly, it wasn't just the developers that were asking for our help, it was everyone. See, all 12,000 employees at my company now have access to that Agent Goose. And while it was designed for developers, other disciplines, they began to use it too. So it turns out being able to speak in natural language and build a prototype or an internal tool, it's a pretty sticky use case that everybody wants in on. But just like the developers, this presented a learning curve for those folks as well. So they needed to learn, how do I use effective prompting techniques to get the results that I actually want?

They needed to learn about LLMs. What is this? How does it work? How they can be non-deterministic? That's always a shock to people. They had to learn about MCP and how they could use this protocol to connect AI to the tools that they use every day. So guess who our CEO tapped to lead this charge DevRel. This was the moment that I realised how big our stage really was. It wasn't just about the developers anymore. The entire company was looking to us to help them step into this new era because again, we already have the skills to do this. Yes, in fact, we are doing this externally for the global developer community. So overnight we started teaching everyone in the company on how to use ai, and this was legal marketing, design research, the executive assistants, everybody. We had to teach them how to become builders too.

We ran weekly enablement sessions. We would teach workshops at people's offsite, so we're travelling to team offsites teaching these folks. We had to adjust our messaging to reach this audience, right? We're used to talking to developers, but now we have other folks. It became less about the code and more about the outcomes we learned about their use cases. This was fascinating stuff. Lemme tell you, they're way more creative than any developer have to say that. And we had to provide this customised guidance to help them automate their workflows with ai. And we treated them just as we would any other community of builders. We nurtured them. We made safe spaces so that they could ask questions, they could seek help. We highlighted their wins. This makes people feel good. All of the same things that we do as developer advocates. This applied really well for this new community that we have and these early wins where people were building things that they didn't think they could. This created so much momentum for the company-wide adoption. And so we got a lot of kudos for leading the charge here. We became central to the company's AI journey, not just as teachers, but now as cultural guides and our work set the tone for how people feel about change. We help them to move from fear to curiosity.

My team receives kudos messages every week from people around the company saying, thank you. You changed the way that I work. This is bigger than anything that we could have been a part of. And even if you're not running internal AI enablement, your stage is still bigger now too. Whatever you're teaching developers about ai, those lessons extend beyond your product, right? We're in a brand new era. Developers, they don't just need help learning about your tools. They need help learning how to work with AI in general. So if you're teaching anything that's AI related, your influence is going to go beyond your product, beyond your company, beyond your immediate community. Every single blog post you write, every workshop you teach, every talk you give is helping developers adapt to the biggest change that we've seen in software development in decades. And of course with great power comes what great responsibility. That's right. So whether you realise that an IU or shaping how developers feel about ai, your teachings can change their perceptions from confusion to confidence. Developers are looking to us, the DevRel community to help them understand what's real and what's hype.

They want us to point out the areas where they should focus and how their existing skills can translate into this new era. So even when you're just writing docs or examples, you are writing the playbook for how people are going to interact with these tools. You're modelling what's good. You're modelling ethical use of ai, right? You're showing folks what this looks like. You are defining the culture of AI adoption and developers are going to carry this with them into other jobs and companies. The trust that you're building today, this is going to extend or beyond this moment. You are essentially helping an entire industry level up. That's powerful. Don't underestimate how much influence you have. Developers are listening to you and what you say will ripple through companies and communities that you'll never see, but they'll feel it. And this is only the beginning. AI is moving fast, steadily evolving, and so will the needs of our community.

So what's next? Right now the space is flooded with the demo apps. Developers are going to need our help. How do I translate this into real production work? Those are cute and all, but hey, I'm on a team with people. I can't just be yolo vibe coding. So when we make those little cute videos for YouTube, we're skipping all of that stuff. We're have to address it soon enough though. Okay? They're going to need guidance. How do I integrate this into my team projects, right? They're going to need your help on how to build out automated workflows that use ai.

Now at the AI does more of the work. Your developers are going to be fearful, okay? Anytime I do a talk about Goose and I do a demo or something, people come up to me and ask, is this going to take my job? It is a little scary how good it is. That fear is definitely real. What we can do here is to help them to realise that the things that they actually value, like creativity, like judgement , these things are actually more important. Now we got humanity at the centre of technology. That means don't just focus on teaching them how to build, but teaching them why they're building it, whom they're building it for. Help them to realise this.

When people ask me how we did it, how we learned ai, how we're leading the masses, both externally and internally, I tell them we already have. We need it. We're storytellers and storytelling is how we help people see what's possible, how to help them feel inspired enough to give it a try. Storytelling is how we make the abstract tangible. It's how we make the scary feel approachable. It helps people see themselves in the technology. We're also naturally empathetic, at least I hope everybody is. Empathy is how we make people feel safe enough to admit what they don't know to ask those questions. This is how we created a community where everyone can learn without judgement and then translating new complex topics into something that makes sense to people. They actually understand it. They can actually use skills they already have. This is also one of our superpowers, right?

Helps bridge the gap between the experts and the beginners so everyone can join on the front. This is how we make AI accessible to the masses. So these have always been our tools. The stage may be bigger now, but these skills are definitely the same, and that's why this moment was made for DevRel. What you're doing now, guiding how developers approach ai, how they feel about change, how they carry themselves into the future. Again, this is bigger than your product. What you're doing is shaping culture. I need you to stand in this moment, embrace the lasting impact that you're making on the developer community and be beyond. And guess what? AI is not the last big shift that you're going to help guide developers through. The credibility that you're building now is going to carry on to whatever comes next. What you do today is setting the stage for tomorrow. And when that stage is even bigger, you'll be ready because you already are. Thank you.

Jon (MC)::

So honestly, thank you so much for that incredible framing. I feel like that is the perfect foundation for the rest of this conference. We're going to ask a couple of questions here, fireside chat style, and then we're going to open it up to the audience for any questions that you have. For Angie. The thing I'd like to start with, you talked a lot about how the people that you're serving are not necessarily the profile of developers that we're all familiar with that in mind. What are some skills growth areas that you think the average derail person can have to serve that new audience?

Angie: Yeah, it's interesting. Communication is something that we pride ourselves in. Oh, I'm a great communicator. This was quickly challenged as you talk to other folks. So we would find ourselves saying things like, oh yeah, just like somebody said, oh yeah, just pseudo with people are like, what the hell is pseudo mean? What is that? They're not familiar with things like terminals and things like that. And so we had to take a big step back. We take for granted how much of our lingo and the tools that we use are just a part of our everyday vocabulary and muscle. This is not the same for everyone else. Some folks didn't really even know how to unzip a foul, let alone open the terminal and start running commands and stuff like that. So we had to really kind of step back and think about, okay, how do I talk to humans and not just developers, but your regular everyday people. And that was a great skill in communication as well. I think that communication in general is even more important now when you're talking with the bots and stuff. If you are not precise what you're saying, they might do something that you don't want them to do. And specifically I'm talking about AI agents that can take actions. So you have to, we're basically building that communication muscle, not just for the devs but for the regular humans. And too,

Jon (MC)::

One of the things that you started with in your talk was kind of touching on the turmoil in DevRel and how this is a new opportunity for all of us. How do you think we build sustainability from now on, right? There is hype and maybe that will normalise at some point. So what do you see going forward?

Angie: Yeah, I think getting to the basics is the key. We're always going to have to test software. Don't matter who wrote it, the bots or the humans, right? It's funny, the humans judge the box code like, oh, AI slop. And I'm just thinking in my head, your code not that much better. They kind of learn from you if we're being honest. But so testing, documenting, all of these things are still important and that's why this is what I want to cycles on it, showing how do we use AI in these things? I don't know. Building software is going to likely change what that looks like, but these things stay the same or relatively close to the point. Well, okay, how do I use it effectively here where I'm not just losing all control and turning everything to the box? Like, oh, you write the code and the test and the docs. Well, where do I fit in and how do we trust that all of that is actually accurate?

Jon (MC)::

Absolutely. To transition to a more technical subject, and then I think we'll open it up for some q and a from the audience. One of the notable things about Goose is that it's open source and that's kind of at odds with a lot of the other tooling out there that may not be quite as open. What do you think that brings to the table for your team and for the company?

Angie: What it brings a lot to the table and that we're giving people access to something. So we get a lot of folks who maybe their companies won't buy these other tools. We just want to kind of play with it on the side on the weekends. And it being open source gives them access to do so, but also don't, we're not an LLM provider or anything like that. And so that gives us the space to really be free on Gemini, you want open ai, you want anthropic, bring whatever you want and have a dog in the fight.

Jon (MC)::

Yeah, absolutely. Sorry, I have one more question I wanted to ask before we give it to the audience, but you talked a bit about how the content landscape is changing. Do you have any hacks that you've discovered for effectively SEO for ai?

Angie: Yeah, so a great thing to do is maybe go in incognito window with chat GPT and then ask about your product or even you. And you can say, Hey, or the domain, tell me what's the top tools in X, Y, and Z. And if your product is not on that list, then you know got a problem. And so you can then ask questions, how do I strengthen my brand so that I make a list like this in the future? And it'll give you some tips. You need to make some more content. You need to be in these channels, you need to maybe work with these influencers. And

Jon (MC)::

So we're going to open this up to audience questions. Raise your hand and we'll run to you right in the front. And please introduce yourself too. We're all friends here.

Audience member 1:

Hi, my name's JE Timothy A. Little bit with a shot of order by trade software developer without can move into product design. So my biggest customer was when you were referencing having that, especially AP advocates for developers, other people. Then actual, so when you would have this a surgeon, so developer and designer relationships that come into play, how do you think Dan interjects itself that way?

Angie: Yeah, we help a lot of designers as well to empower them even more. So for example, the designers, they make their designs or their components don't really have the skills to turn that into an application. So you got to kind of wait for the developers and then there's this whole back and forth about, no, I didn't want it like that, stuff like that. Now with these tools, we are showing them how you can take these designs and turn them into an actual application. And so they feel so much more empowered. Now it doesn't mean the developers aren't needed. Yeah, that's great, but then it's still all of the other stuff. We got to make sure it's secure, got to make sure it scales and all of these different things, but they feel a lot more empowered with these tools and they were left out of the conversation for a long time. We kind of focus. I think overall the focus right now is on developers across the board and there's so many other use cases where AI can be helpful to other people and to kind of break the bottleneck, even thinking about internal tools. So now we have marketing whipping up internal tools for themselves. The developers were never going to make it right? When would they have time to make an internal dashboard for you? But now they can do these sorts of things for them.

Audience member 2:

Hi there, I let the boss specifically around prioritising relationships over content, these kind of things over the resonated with it really speaks to the core job be done for Debra. However, my Dki is very data oriented and he wants to see ethics that represent DE's impact to the business. And so as you said, SEO is search as a channel for distribution. It's declining, which needs way press to judge efficacy of our work and carry through to specific channels is count. And so I'm curious if you have any solution or lessons learned through Adapt Peak for overall wish. Bert B traditionally, they're more challenged to do that because they have better channels to get access to information and which we,

Angie: Yeah, we have a talk on that, don't we? Isn't somebody doing a talk? I want to go to that talk. Actually it is circled on my list. I need more insight on how to track that stuff too. I don't want to track blog clicks or social media likes. It just not, it was never that great, but it's definitely not given the signal I need now. So again, what I do now is very unofficial. It is basically talking to the boss and asking it where we are in the grand scheme of things. And I get a lot of insight on that and it also helps me to formalise what we should be focused on. So it's like, yeah, y'all are not coming up in these specific channels or these specific sites. And so I know, okay, let's focus there next quarter or something like that. But I don't have the answers. Nobody ever has the answers to metrics. Hopefully the person giving the talk on metrics. Are you here? Hey bestie. So let's all go to the metrics talk.

Audience member 3:

Hi, I'm Kevin. Thank you so much for the presentation, Angie. It was great.

Angie: Thanks.

Audience member 3:

My question is, I agree with you that developer advocates are the right people to drive this new message, but the question that I have is more about the audience because most of us worked at companies who had an existing product with an existing persona who had a different problem and our expanding into a different audience, which might be a different persona to the one that we have. So the question that I have for you is how do you do expand to this new market or to this new persona without hurting the one that you have? It's a challenge for companies who had an existing product outside of AI and now they're expanding to space, everyone's expanding to ai, I guess. So how have you not heard that existing person who might not relate to some of the problems you're trying to solve?

Angie: That's a great question. How many of you are actually working on an AI tool or ai? AI features? Yeah, everybody. Okay, so alright, you're right. So you have your legacy products and things like that. So for example, for me, we support everything under blocks. So that's our square SDKs and APIs, cash apps, same thing. But then also there's this shiny new open source AI agent that we also need to support. And so I have basically folks allocated to all of that, even the legacy stuff. But the AI stuff is our top priority just because this is what people are hungry for and this is what they need to learn. But that doesn't mean we neglect the other things, right? Like, Hey, I'm still running your SDK, I still need these APIs. So we can't neglect that. So we do have some folks still focused there as well.

Audience member 4:

Hello everyone. Thank you so much for this opportunity. I'm a student here pursuing a master's in it. I have no idea about developer relations. I would love to get your advice on how can I break into this industry. What I understand is basically we will be a bridge between technical as well as business side of things, which I think I would love to be part of this industry, but I have no idea where to get started, what to do so that potential founders can hire me. So please let me know. Thank you.

Angie: He is like, yeah, great talk. Didn't understand nothing you say. But yeah, I want to do that. Love it. I think the best thing, best advice I can give you right now is to just start doing the job now. So this is how I broke into developer relationship. I didn't even actually know developer relationship thing. So it really was authentic that I just wanted to share what I was learning with my community. So I started blogging, I started giving talks. It wasn't until I realised, oh, this has taken up a substantial amount of my time. I still have my software engineering role too. I need to figure out how to make money doing this. I actually like it. And so put it on my resume. I didn't have a company name to put it the top, I left it blank. But I called myself a developer advocate and I said all the things, the bullet points of things that I have done that fit, that showed the content, the portfolio, and got the job. So my advice would be to just start doing it now. You don't have to wait for someone to hire you. You need to build those skills anyway of how do you communicate with developers, how do you explain concepts? How do you share what you're learning? And in these various written and video and things like that spoken, I say spoken word.

Jon (MC)::

Awesome. Well thank you again to Angie. Let's give her a huge round of applause for that great kickoff for the conference.

Angie: Thank you so much.