How to Lead DevRel in the AI Era: Stop Playing It Safe

A developer advocate describes leading Developer Relations teams through the AI boom by taking unconventional, hands-on approaches — including building a Copilot talk that became one of O'Reilly Media's most popular live workshops, launching a public livestream series at Block's TBD unit that brought competitors onto the company's own platform, and creating a "vibe coding" competition called The Great Goose Off around Block's open source AI agent, goose. She argues that DevRel leaders must keep building alongside their teams to set an execution standard and give meaningful strategic direction.

I’m awake typing this at 5 AM on a Sunday morning, plagued by two issues: the modern operating model for Developer Relations is broken, and I cannot find a comfortable position to sleep in at 36 weeks and 5 days. The latter is slightly less relevant to this post, but equally present. Right now, DevRel feels stuck in an existential struggle for impact. Every company has pivoted towards AI, and most companies are following a similar marketing playbook. It’s hard to stand out in a saturated market. Developers our customers are overwhelmed by the many choices. But I’ve always operated DevRel by going against the grain. For example, after joining GitHub as a Developer Advocate in 2021, GitHub Copilot launched. Our team leaned toward taking a patient approach since there was so much misunderstanding around the product. Users had strong opinions like: “GitHub is stealing our code” and “AI is unreliable”. However, I eagerly and possibly too hastily built a talk titled The Truth About Copilot to address developer skepticism head-on. I took feedback from my Director that the title was a little controversial, so I renamed it to Level Up with Copilot . I was pleasantly surprised that it became my most in-demand talk to this day, one of O'Reilly Media’s most popular live workshops, and I even wrote a blog version of the talk, which Thomas Dohmke, the CEO of GitHub at the time, distributed to enterprise customers. Then, I moved to TBD a business unit within Block as Staff Developer Advocate. I was tasked with building community and adoption around complex, misunderstood tech: decentralized identities and cross-border payments. I noticed a lot of the developer education happened strictly within our internal channels. For instance, our live sessions lived in Discord. I decided we needed a version where the broader public could also discover our tooling, so I started a livestream series on our YouTube channel. In that series, I brought on "competitors" who were solving similar problems. Bringing competitors onto our platform was seen as risky, but I used it to prove that what we were building wasn't weird or shady. A lot of developers kept associating our work with cryptocurrency. My stream helped prove that respected companies like Auth0 and Microsoft were exploring similar topics, and by collaborating, we were able to tap into their communities. A year later, I was promoted to lead Open Source Developer Relations at Block just before I went on maternity leave. I was sad to step away at a time when I felt I needed to immediately support my team, but that opportunity gave me an edge. I spent every nursing session with my newborn learning the agentic space, contributing to goose our open source AI agent , and evaluating how we showed up in the market. I recognized that our team was doing good work, but we could push further. For example, we had a livestream series showing how to use the product, but because early models in late 2024 and early 2025 were slower and less reliable, live demos carried a lot of friction for advocates and created dead air for viewers. So I thought: why don't we add a series that is short, snappy, and silly, but still demonstrates real engineering approaches? I introduced a vibe coding competition called The Great Goose Off . I challenged contestants to build the most chaotic tools. That way if the agent failed to deliver the correct output, it was still enjoyable to watch. It also lowered the pressure for contestants. Sidenote: The livestream did really well for a period of time. I will admit that once models improved, it felt like an outdated trick to watch agents generate a tool. Here are more tactics I used to lead my team through the AI boom: 1. Stay hands-on with the product Even though I was a leader, I continued to build. I recognized that I set the execution standard for my team, so I can’t give meaningful strategic direction beyond "we need more content" if I’m not experiencing the product friction myself. I did this by: 2. Meet developers where they are in real time Your foundation content installation, feature walkthroughs is necessary, but it won’t earn attention. To cut through, your team must inject your product directly into active ecosystem conversations. 3. Move from perfectionism to high-velocity quality The AI ecosystem moves too fast for a DevRel team to spend a full quarter polishing a single piece of content. The companies that can afford immaculate quarterly releases are frontier labs with natural monopolies like OpenAI or Anthropic . Emerging tools must be inside the shift while it is happening. 4. Keep technical meetups hands-on and actionable DevRel event budgets often get wasted on forty-thousand-dollar networking breakfasts that yield zero long-term developer adoption. Happy hours are great, but hands-on technical environments build real adoption. 5. Automate your documentation pipeline Developers aren't the only ones reading your documentation anymore; AI agents read them too. If your docs are out of date, agents fail to use your tooling. 6. Invest in community, but not just as sales activation Companies don't build developer communities out of pure altruism; they want adoption, feedback, and open-source contributors. But true community relies on authenticity. Coming from the West Indies, Black American neighborhoods, and church environments, community has always meant showing up, supporting one another, and building real relationships. In DevRel, that means: I’m confident that I took the right steps for two key reasons: “I've worked with a lot of talented people, but every so often you come across someone who raises the bar for everyone around them. Rizèl is one of those people. She is a brilliant engineer who consistently delivers high impact work with both speed and thoughtfulness. She is also the most creative person I've ever worked with. But what I love most is that she doesn't just toss good ideas on the table, she executes with precision. In fact, most times, the results are way better than you could have imagined. Rizèl is a natural leader. She sees gaps, steps in, takes ownership, and drives initiatives forward. She's someone others look up to, both inside and outside of her immediate workplace. Her work earns respect, and her approach sets a standard. All of this makes Rizèl an exceptional Developer Relations Engineer, in fact she's one of the best in the industry Managing Rizèl has been one of the highlights of my career, and I hope to get a chance to work with her again in the future.” Across all of these roles, the environment keeps evolving, but the core challenge stays consistent: figuring out how DevRel can move with developers instead of behind them. I realize my best developer advocacy moments came when I stopped trying to fit into a template and just trusted my instincts. My call to action to DevRel leaders and specialists is: stop trying to play it safe. Leading in this era requires building in the open, moving fast, taking creative risks, and treating your community as people you appreciate.