Generative AI has had a lot of a lot of press coverage recently. It was certainly getting a lot of attention. It's not as fancy or as a, complex as some people may make it sound. It's really just a system by which one generative AI model can talk to others. Best analogy I've got for it is different between having, like, a single individual contributor working on a project, to having a whole team, and having a manager or a project manager organizing who does what. So instead of one model doing its best to answer every every question you give it or every task you give it, an agentic framework tells the primary LLM, the manager effectively, what tools it's got available, what other agents it can call, and what they're supposed to be good at, and what kind of format this supposed to send messages in. These frameworks allow us to provide put specialized to LLMs into different jobs. Ones that are geared to particular pieces within our, for example, within our Copilot tooling, we have tools geared to working out, okay, what's the most relevant table to the customers request to go for information they want to build? Which of our tools is most relevant? Is it separate, separate tool to work out the best components? Another might be to query the pipeline and find out what's there already. All of which become part of the library that Magento framework can can access. For data leaders, it's probably quite important to get a grip on this sort of thing because the trend we can see or we can imagine is that over time, people are gonna start coming to you to build out those tools, to provide capabilities to feed into an agentic system. And that gives you a great deal of modularity that you can plug tools in. You can plug in data requests. It allows a lot of sophistication. Potentially, also can be brought into data processing. But an agentic system might allow us to have more flexible pieces. It might give us more in terms of dedicated tools for responding to that support request, may be reading the the previous customer's previous case history, and understanding their system might be a tool within a framework to allow a more tailored response, for example. They become more extensible. It provides more more options and the like. But don't be frightened by it just because it's the latest buzzword. It's just a different way of setting the LN instructions. And literally, all it needs is the the prompts I give to an LN will involve its history back and forth with other models and a list of what's available to it as tools. So they can build them up very simply.