Contents
In this follow-up to my 2024 talk in Sofia, I share what really happens when you embed generative AI into day-to-day documentation work. Over the last two years I’ve moved from used custom GPTs, to adopting Agentic Coding AI models, voice cloning, and online, offline, and Hybrid AI-assisted workflows in a custom-coded Python-based documentation CMS. Trust me when I say that I’ve seen the significant benefits (and hit the numerous brick walls).
We’ll look at where AI genuinely saves time (often lots), can improve consistency, and unlock multiple output formats at once - as well as also looking at where it introduces risk, generating more work than it saves, and other issues with keeping in mind. The session focuses on practical patterns, decision checklists, and concrete examples you can adapt, so you leave not just asking “Can I use AI for this?” but confidently answering the *far* more important question of: “Should I?”
Takeaways
Learn a practical framework, real examples, and pitfalls to decide when AI should (and shouldn’t) be used in your documentation workflow.
Prior knowledge
No prior hands-on AI experience required.
Helpful but optional: basic familiarity with software documentation concepts (topics, versions, workflows) and common authoring tools.