 I think that there are two particular opportunities where I'm excited about the use of these new technologies for lending. The first is we've seen some interesting new innovations in agriculture and small business value chains, which were previously very economically active but underserved financially segments. And with the increasing availability of information both administrative transactional data and alternative data such as their social media profiles on social media marketplaces and things, these businesses are becoming better known. And so from my work in East Africa, I've seen a wide range of different firms that are taking loosely dispersed small businesses like the kiosks in your villages and connecting them to tighter value chains and are able to offer lower cost access to wholesale products, wholesale financing. In the case of farmers, we see advances on future sales from the next season. We see loans being used to bridge the gap to send your children to school while you're waiting for the next harvest and the next payment because farmers income is lumpy. And so those are really exciting for underserved economic segments. And then the other thing I'm excited about is what access to more alternative information and artificial intelligence is doing for retail banking. We're seeing innovations in markets where they have open banking concepts where firms are helping consumers to be smarter consumers. So looking at their financial history and giving them advice on when they should switch banks to save money or where they should be putting their investment funds. And so some of those automated advice and facilitated switching products I think are going to be great for competition because DFS has some market concentration issues in some jurisdictions. I think the diversity and speed are the two challenges. So diversity in terms of you have many new entrants to the financial technology sector that come under a wide range of regulatory jurisdictions and in some cases are not regulated at all. If you look at digital lenders in many markets, there's not the proper supervisory infrastructure to have equal oversight on all different lenders. So one app you download comes with certain consumer protections and another does not just because of regulatory definitions. Now that's starting to be addressed, but it's still something where there's clear issues that need to be done to update market coverage to reflect the modern economy. And then the second aspect, for example, we see rising concerns around how data is being used, things like biases and algorithmic scoring models which can harm already vulnerable or marginalized populations. And keeping a pace with really fascinating technologies in the private sector is hard for traditional supervisory institutions. And so we'll never get to the place where I think regulators will have as sophisticated of an artificial intelligence or understanding of algorithms and scoring models. But at a minimum, I think we need to invest in more staff and resources which can at least understand these models and can look under the hood a bit. Because I think we can't just take these innovations at face value. We have to become experts ourselves.