Skip to main content
There is a difference between an agent that works and one that feels like yours. This chapter is about closing that gap — refining behavior, fixing what is off, and understanding what is happening under the hood.

​
Skills are your memory

Ottic does not silently learn your preferences in the background. You make it yours explicitly, and the main tool for that is Skills. Anytime the agent produces exactly the report, audit, or workflow you want, save that prompt as a Skill. It captures the structure, the data sources, and the format — so next time you (or anyone on your team) get the same quality in one click, without re-explaining. Your Skills library is the institutional memory the agent draws on.

​
Refining Routines

After a few runs you will spot things to adjust. “Make the report shorter — top 3 only.” “Stop including branded queries.” “Pull from the production database, not staging.” Each correction is small, but they compound: a Routine that felt 80% right becomes 95% right after a week of tweaks. And if a Routine is not working for you, turn it off, rethink the instructions, and re-enable it. There is no penalty for pausing one.

​
When something is not working

Start with the run history — every run is logged in full detail: what the agent read, what it decided, and what it did. You can also just ask the agent why it did, or skipped, something. If a Routine is not running at all, check three things: is it enabled, is the trigger and timezone correct, and is the right account selected? If it runs but misses things, it is almost always an instruction gap — you told it something slightly different from what you meant. Look for an assumption you left implicit and make it explicit.

​
Auth and permissions

Sometimes a connected service stops working with an auth or permission error. The fix is almost always the same: go to Settings → Connected accounts, find the service, and re-authorize. Think of it like a session expiring — you just sign in again.

​
Getting help

Two options: ask your agent — it can check its own run history and suggest fixes — or use Talk to founder in the menu to reach us directly.

​
The big picture

Ottic starts by working with the tools you connect. You start with small tasks and check the output. Over time, trust builds: you save the wins as Skills, schedule the recurring work as Routines, and the things that used to take you hours start happening automatically. The best way to get there is to keep using it.