About the talk

Inevitably, there will be someone at every company who asks “Why can’t we just use AI to write the docs?”. The Twilio Developer Education team has an entire document that outlines our approach to AI, but for the sake of brevity, the tldr is: Generative AI still carries too much risk at too high a cost to be used as an autonomous customer-facing agent given the expectations for accuracy among our audience. That might leave you wondering if we’re sticking our heads in the sand and pretending like AI isn’t here to stay. That’s not the case at all. Instead, we’re using AI to supercharge our potential as a small team and spend our time working smarter, not harder.

As a team of 11 Developer Educators tasked with supporting over 100 Product Managers across Twilio, we have always had more work than we can possibly do by ourselves. To address this, we created AI Agents that take care of repetitive or time-consuming tasks, freeing us up to focus on more complex work that better supports developers' needs through documentation. In this talk, I’ll share how we decided which agents to build and how we made them more accessible to less-technical Twilions outside of our team. I’ll share some practical ways you can use agents in your own work to simplify your workflows and streamline your authoring process. Once we’ve covered the basics, we can share some ideas we have for taking these agents, and our docs, to the next level.

This talk is for anyone who contributes to or cares about documentation and wants to use AI in smart, thoughtful ways. These agents help our internal teammates and external stakeholders write better docs and help our Developer Educators focus on high priority work and maintaining technical accuracy. There are a ton of short-staffed technical writing teams out there that could similarly implement these agents to enable product and engineering folks to help write docs or make the time from draft to publication much shorter.

About the speaker

Maria Bermudez is a documentation advocate who's just as comfortable wrangling APIs as she is leading hackathons with a side of coffee-fueled ambition. When she's not demystifying tech jargon and reducing backlogs by half before breakfast, she's likely designing intricate corsets or headbanging to her favorite heavy metal bands. If you catch her between sprint planning and crafting GitHub magic, she might just let slip the secret to launching both a killer AI process and a successful food truck!