Ben, we know that people are using these tools to check symptoms, but I suppose what comes next? What comes after that?

I mean, I think it's the transition from chat to agents is I think what's really interesting.

So we're going to go from a world where consumers are just corresponding with large language models, checking their symptoms, understanding their diagnoses, et cetera, to a role where they have a health agent, an always-on health agent, that not only can they ask it questions, but they can connect it to their data, whether it's from their wearables or whether they're granted permission to have access to their results and their their health records, et cetera, and can create a really sort of integrated, always-on health advisor that gives you advice about diet, about exercise, about sleep, about all those things that are actually genuinely related to your underlying health conditions and circumstances.

And then potentially can move from there to make appointments for you, to arrange health care checkups, prompt you to get your immunizations, prompt you and make appointments for you to do things that are important for your health.

I think that sort of always on health advisor function is a really interesting one.

And there's some really interesting questions about whether that's something which in nationalized health systems is provided by that national health system, or is it something which is a private sector competitive choice, or how do you regulate that? How do you manage it? How do you license and assure it? I think there's some challenging questions about how those agents will access private and confidential data and what sort of security and confidentiality will be provided alongside that.

But the future of AI is definitely agentic.

So what are you actually seeing? What examples have you got?