We have looked at whether we can use AI to address the administrative backlog. Unfortunately, the way we have the data is not such that we can use it with AI systems. Some of it is handwritten or it requires a lot of analysis to get into a system, so we don't think that AI can be useful for looking at historical files.
We're looking at process automation a little bit, but that's already built into the national additions-to-reserve system that we've looked at.
The other area where AI doesn't help us, and Darlene just referenced this, is with the relationships that are so critical to the advancement of ATRs. We don't yet have AI that can help us with that, unfortunately. That's been a shortcoming, but we are certainly trying to look at how we can be creative.
As AI advances and evolves and as we become better with our digital footprint around what we have data on for reserves, we may find new avenues. We're doing what we have looked at so far and remain open to possibilities, but we think we've maximized its ability currently.