For someone who's listening who doesn't really know what a self-learning agent is, can you kind of explain what that is and how you set it up?

One of the big problems I think most people who are new to AI are doing is they get into the loop where they build something and it works.

And they'd actually never go and revisit anything.

the output of that thing.

And results aren't always linear, right? So like let's say we have the agent cut off spend after 12 hours.

It might be missing incremental on the other days and you might like see like a 5% lower amount of conversions over the course of a month because there are some that are tripling in.

It's really important for us and kind of everyone to go in and reevaluate the output of what the agent is actually doing.

And then did you have a breakthrough moment kind of where you figured everything out and then were able to really have clarity on what the final project would be?