Jonathan Reisin Levin
Analyst · TD Cowen
Yes. I would just add, I wouldn't obviously do anything specific in your model, Bill, only in the sense that what we've told you is we've got operating leverage in the business. We've shown continued margin improvement, something on the order of, I don't know, 1,200 basis points or something since 5 years ago, and we've told you we think there's still margin to go. And I think it's all kind of captured in this. So I wouldn't -- it's not like to me a unique modeling exercise, but it's facilitating obviously, that scalability. I think that without offending all of my other wonderful 550 colleagues, my favorite meeting of the month that we have is with me and our Chief Technology Officer, we meet with our top 5 users of Enterprise ChatGPT every single month. And we do that to understand what they're using it for, what's been successful so that we can then have best practices spread out throughout the firm because to Michael's point, it's happening every day and everywhere at the firm and what our job as a leadership team is to make sure that we can figure out then how to scale those best practices across the organization without dampening that entrepreneurial spirit and without dampening that curiosity that we're seeing across the employee base. I could give you numerous examples of it, and none of them would really make any like big sense to you, but they're all little wins. I'll give you a small one for fun. Our tax team just recently built a Custom GPT that enables them to read tax returns and automatically enter data that was used to be manual. As you know, in the industry in private markets, a lot of that no longer done by facts, but still a lot of PDFs and the ability to take PDF information, whether it's on a report on a portfolio company or whether it's a financial statement that used to be manually entered somewhere, but now have that be automatically read and input it into our data lake. All of those things are fascinating usages, the first draft of a marketing deck you might create for a client that can come out of a GPT and your ability then to obviously make it appropriate for your organization, make it accurate because we all know it can hallucinate. There's just exciting use cases kind of all over the place, and we encourage people to think about it not only as a productivity and an efficiency tool, but also think of it as a thought partner and a creative partner and not limit the power of its intellect for lack of a better word. So I just think you're seeing that all across the world, Bill, all across the industry, and it's just obviously such early days.