Yes. I'd say that's a really great question and an element that I forgot to share in answering Keith's question. So we think about it in the realm of context versus core. And for us, core is what we talked about, which is we are all about delivering done-for-you experiences with AI, data and HI to help you from lead to cash and to help you from credit building to wealth building. And so all of our investments over the years with the proprietary data, data models, our domain-specific AI models, which, by the way, to a lesser extent, is actually LLMs. The majority of our AI capabilities is actually knowledge engineering and machine learning. And of course, we've built out our Intuit financial large language models that really works in a very complementary way to deliver these experiences with confidence where there's a lot of liability for customers, which is why there's so much demand. And if you think about our category, demand is high and supply is short because there's not too many that do what we do end to end. That's core. And when you think about what context is the part of the alignment that we have, like, for instance, with Anthropic, where for mid-market customers, if there is a customer like a construction company that is looking to look at their project plan, look at their lean waivers, look at their subcontractor payments -- and then they're looking to actually understand the combination of those and the impact it has to their cash flow. Our customer now on our platform, because remember, it's our AI models, data models and HI that drive the end experience can now -- they don't know what they're using, but they're using some of the capabilities with Claude, Cowork to actually be able to create a dashboard to see the long tail of things that, that construction company needs to see. And by the way, every construction company wants to see different things. Every roofer wants to see different things. Every architect wants to see different things. That's context for us, right? Because we don't want to go out and build the long tail of things, but it actually allows us to disrupt industry-specific verticals. That's an example of context. Another example of context is for Anthropic and OpenAI, context for them when a customer is in their app is what's the customer's intent. That's core to them. And once they identify what the customer wants, then it becomes very much our skills, our experiences is what the customer use and they're in our platform. So this is back to where I started. This is very much about context versus core. And from our perspective and the LLM providers, it's actually a very clear cut how we are partnering to deliver experiences for customers. So hopefully, that helps, but that's really the way we're executing the experiences and how the models actually work.