Number 3 is AI struggles with genuinely novel situations, new situations that it can't account for and has to adapt.

Yeah.

And, and this is an interesting one.

And a-again, uh, you know, I, I, I think it goes back to how well the AI has been trained or how, uh, what kind of data the AI is working with.

But if you think about the inferencing capabilities are getting better and better.

You know, if you take the example, for example, of fraud detection, the AI may be able to identify most situations, but a truly novel fraud, like an email scheme, perhaps might not be caught by the AI.

So I think it has to be used in a way that it provides value to you.

And so maybe you could take us through that one.

It's the boss agent.

The boss agent.

And, and the interesting thing is that the, in terms of the design, the, each sub-agent can't talk to the other sub-agent.

So you can't just have a sub-agent that deci- uh, that comes up with the exposures and sends that over to the netting cycle.

Um, it's gotta, it's gotta go through the boss.

It's gotta go through the orchestrator.

And this is the big advantage of a, um, multi-agent system versus what I would call a monolithic or a black box system, right? Because with a black box system, you can have this entire end-to-end process running, but you don't know where, if something goes wrong, or, or say the AI made a decision, it, it decided that this was, these were the exposures, um, that the company will have for the next three months.

Yeah, maybe give an inventory and, and touch on a few that you think would be of particular interest.