DevRel and AI: a department lead’s self-evaluation one year in

Reflecting on the past year's journey of integrating AI into our Developer Relations initiatives at Nylas, I'm taking a moment for self-assessment.

Last year I wrote a blog post titled “Experimenting with AI in DevRel”. It lays out four areas of approach for adopting AI in DevRel work. That post later led to my hosting a DevRelX Summit panel with a number of Developer Relations experts on the topic. Needless to say, this has been on my mind.

Reflecting on the past year's journey of integrating AI into our Developer Relations initiatives at Nylas, I'm taking a moment for self-assessment.

House-keeping note: This evaluation focuses solely on my leadership and strategic choices, and is not a judgement on the incredible efforts and accomplishments of my team or the teams we collaborate with.

Using AI in our day-to-day work: A-

Incorporating AI to enhance our workflows has been a transformative experience. Our achievements in content creation, sample code building, and workflow utility scripting with AI have added extra punching power to our team without increasing its size. Faster? To some extent. More capable? For sure. Greater confidence in our quality and depth? Most definitely.

As just one of many, many examples, I’ve been excited to see Laura, our Staff Documentarian, using AI to build scripts and troubleshoot issues for our Docs app stack. Some of these enhancements and fixes would have required Engineering intervention in the past. Now I feel like the Docs team is solving 80% or more of the app stack issues they encounter on their own.

I give myself an A- here mostly because I could be doing more to encourage a uniform approach to using AI for getting our work done. It's been very up to individual interpretation so far, but maybe that was what we needed for year 1.

Using AI in communicating API use cases: B

Our journey in blending Nylas APIs with AI APIs at the application level has yielded open-source examples and, wildly, sparked a global roadshow.

Ram, one of our Developer Advocates, has shown developers how to mash these two technologies together in cities ranging from San Jose to London, Berlin, Zadar, Paris, and beyond in just the last 12 months.

On the leadership side, I give myself a B here because I know in my bones there’s more we could be doing and it’s on me to set even loftier expectations. I look forward to carrying some of this work forward in 2024.

Integrating AI into our developer experience: D

On the DevRel side, we’ve taken a much more cautious approach to integrating AI directly into the Nylas developer experience. Handing the wheel to an AI model to directly interact with Nylas developers is... well, it’s scary. Developers are the star of the show for a DevRel team, and we want to ensure they are treated well.

In one experiment outside of our team, the Product org’s Growth PM, Saif, has added an AI chat bot from Mendable to our Developer Dashboard. There have been enough positive indicators from that initiative to encourage further exploration.

But I give myself as DevRel lead a D here because I haven’t articulated a comprehensive developer experience AI strategy for the DevRel team itself.

Creating AI-powered product features: N/A

As DevRel lead, product feature roadmapping falls outside of my direct purview (though the team does have some influence). In fact, I wrote the original blog post last year to provide a framework that emphasizes the preceding DevRel areas of responsibility over ideating on product features. As such, I’m not grading myself on this area.

However, over the last year, I’ve had the opportunity to observe the Product and Engineering teams create API features like Nylas Smart Compose which leverage AI models to draft emails. I'm really happy with the measured approach Nylas has taken here—we aren’t hitting developers over the head with the obvious AI plays that developers could easily build themselves.

This has given me further confidence that our Product team will make the right moves to deliver what developers need. DevRel will continue to be there to support how we deliver product to developers around the globe.

Conclusion

This last year's exploration of using AI in DevRel was a blend of achievement and learning opportunities. As I reflect on my own contributions as department lead, the potential for exploration and growth is clearer than it was 12 months ago, but I know we’re all still in for a wild ride.

There’s so much we don’t know at the beginning of this major inflection point in the history of technology—and that uncertain future is the most delicious part. I remain ready to play and to support teams who are as well.