A lot of people are talking about AI frameworks and agents. You don't need either to get started. I began with nothing but a Claude subscription, and even that takes real work off my plate.

My core areas are data and databases, that's where I know my way around. Outside of that, my knowledge gets thinner, for example with the operating system or with issuing certificates. That puts me in the same position as many DBAs, because no database runs on its own. I understand how the systems work, but I lack the details. Sometimes it's the exact command, sometimes the log that holds the cause. This is exactly the gap AI fills.

Live in my lab, I show everyday tasks around SQL Server. An instance won't start after a certificate change, and the error message doesn't help. A disk is full, and nobody knows why. Which certificates expire soon, and who is actually a member of the sysadmin role on all my instances? And then there's the inherited script that nobody ever documented.

The decisions always stay with me. AI doesn't replace my understanding. It saves me typing and searching through documentation. I read the generated code and trim it down, and I show that live too: what gets cut from the answer matters just as much as what stays. Finally, a look at where the journey leads if you keep at it: Claude now runs the Ansible deployments in my lab by itself, with the same subscription I started with.

This session is for DBAs who wonder where to start with AI without turning it into a project right away. No prior AI knowledge is needed. After this session, you'll know how to start with a task of your own in small steps and how to check the AI's answers before you use them.