Getting Started with Count AI

Hi. I'm going to show you how you can use Count to get insights from your data in minutes and, more importantly, what to do next with them. So, in your canvas, you will see this AI agent box and if you've come into a work space that already has database connections or catalog set up, then you can just go ahead and start asking your questions. If you're new here, I'd suggest the easiest way to get started is to just drag a CSV file onto the canvas. Now my data is ready. I can go ahead and ask a question in this box or I can use one of the suggestions that the agent has provided. The agent will start analysing my data and all of its reasoning will appear in this panel on the right hand side. So, I can review everything that the agent has done by expanding on any of these steps, But importantly, I'm not just left with a chatbot output. I also have these assets that have been added to the canvas for me. So let's have a look at what we have. Here I can see an overview of the insights that the agent has found, and then underneath you can see they have presented some frames that represent these key insights with more detail, some supporting tables and supporting visuals. Crucially, all these elements are editable and you can do as you please with them. So, if I were to click on a visual, for example, I can see in the visual editor exactly how this has been created and I can make edits. So, maybe I want to add in more detail here and bring plan level detail into this. Similarly, where the agent has relied on SQL for its analysis, I can also review that code. I can edit it. I can export it and do as I please. So between the conversation thread and these editable assets that are provided to us, we really can test the lineage of where this information has come from, how it's formed, and trust the data that we're working with. So what happens next? Well, there's a few options. We can keep iterating and working with the same agent. So to get our conversation thread back, I'm going to simply select the agent or the frame that it's created, And it's already suggested some next questions we might want to ask. We can also just ask for this to be put into a different format. So we might we could ask for a dashboard view or a slide deck if we just want to present this. It's also possible to add multiple agents to a canvas at the same time. Perhaps we'd like to preserve this initial analysis. I'm going to use this agent icon and add a new agent to the canvas. Here you'll see the data sources available. I have some catalogs that I could ask questions of, and this refers to the cells that are on my canvas. If I want to base my analysis on objects that are already on the canvas, for example, more CSVs or the outputs of another agent's queries, I can use this pointer icon here and simply select the cells that I'm interested in and just add those to the context. Now I can ask my question based on the output of these objects. We believe the real unlock to better analysis is through collaboration. By using our whiteboard tools, you can tag other users and bring them in to work real time in the Canvas with you. Decisions can happen faster because everyone's working together in one place. They can bring their own ideas and perspective and verify the outputs too. And when you're ready to present your work, we have a presentation mode and this tidies things up for meetings. So I could select here which frames or slides I would like to present. I can toggle through these and use them for my meetings. So really count doesn't just speed up getting from data to insights that you can trust, but it gives you this single place for your team to work together and really allow them to focus on making data driven decisions. I hope that's helpful. I hope you enjoy exploring your data in count. Thank you.