Other - ADI - ChatBot - AgileData Gone in 60 Seconds

Welcome to another Agile Data Gone in 60 Seconds. Today I want to have a little chat to you about this thing called ADI, the chatbot, and what we've built over the last couple of years using ADI and where we see her going in the future. So let's have a quick look at the Agile Data product. So we've come in, we're in the catalog screen, we see a bunch of tiles, a tile effectively being a view or a table of data that we want to use. And we can see on the right-hand side here we have a little fish poking her face out, and she's our chatbot called AD. So if I click on AD from any screen, what I'm going to get back is a bunch of steps. So effectively replicating the menu we've got at the top, bringing it down to the chatbot. And so I can say, you know, let's go to the rules screen, and it will take us to the rules screen the same as if we clicked rules or rules. So a typical way you start off with a chatbot, using it as kind of a menuing system to get you around the product if that's the way you prefer to do it. The next thing we did was we said, okay, we could use ADI to reinforce an action, right? See something that we actually want some feedback on. So for example, if I go down and actually want to click on that tile and maybe I want to go and delete it because it's a draft. So we can use AD to come back. Instead of using pop-ups, we can say, uh, are you sure you want to delete it, right? It's a, an action that we want to actually get confirmation on before you do it. So yep, I'm happy with that, delete it. She's going to go away and tell me that it's been deleted, and then I need to go somewhere else. So maybe I want to go back to the catalog, or I want to go somewhere else in the workflow. Depending on what my next data task will be. So in this case, I'm just going to go back to the catalog. And effectively that tile's gone. The next thing we want to do is see if we can use AD to actually explain some things to our data magicians if they're new to the product. So if I go into the rules screen and effectively open up one of the change rules, in this case it's creating the detail for a product. I may— I've got a bunch of versions so I can see the published version, the previous versions of those rules rules, I may want to modify that rule. I may want to change it. So what I'm going to do is click Modify, and then Adi's going to pop up and she's going to explain what's going to happen if I carry on with this task. So she's going to tell me that if I modify the rule, we're effectively creating a version or a copy of the rule. The, the published version is still active, it's still running as a production rule. Uh, so I'm going to go create this version and then run both in parallel. And then if I actually go and publish the draft, it'll overwrite the, the published one and keep an old version for me. So again, try to bring in some user guidance into some of the more complex tasks that we've made simple, but you still have to understand what's happening from a data point of view. So if I go create draft, she's going to go away and tell me, hey, I've done that. And then I have a choice. I can go to the published version of the rule or the draft version. Typically we go to the draft. So by clicking on that, I've now been brought into the draft version of the rule. So I can see now that I've got draft and publish. So reinforcement of what's going to happen, and then remove some steps to go in and actually do the thing I want. Again, if I go back, maybe I've got a rule there that is already available and I want to go and actually run it. Really simple thing where she's just popping up saying, 'Hey, I've done that, your rule's running,' right? So nice reinforcement again of the task that I've just done. And the last one that we've started working or worked on for now is this idea of using AD as a form of a wizard. So if I wanted to, maybe I wanted to schedule this tile load to happen as a custom sync rather than auto sync. So what I'm able to do is go to custom sync. She's going to come back and tell me when the tile was last synced, so when it was last refreshed with data. Do I want to change it? Yeah, look, I want to go and actually say I want to move it to a weekly schedule. Okay, if you're going to go to weekly schedule, then, you know, do you want it on a Monday or Tuesday or a Wednesday? What time do you want it? And then by clicking that, that would actually change it to that custom sync. Again, if I want to go through and say it's hourly, then she's going to come back and ask me a different question, right? So how often, you know, how many hours every year— 1 hour, every 2 hours— do I want to actually go and run that? So again, that idea of bringing micro actions or interactions into AD as a wizard rather than big pop-up screens is something that we think has value and is a useful way of working. So lots of different ways that we're using AD at the moment. At some stage we will bring in a Q&A function where you can ask AD a question such as, you know, what were my total sales in the southern region in 2017? The other thing we'll do is we'll bring in some form of recommendation algorithm. Them. So when I click on her, uh, we would have tracked what you commonly do as a data task in the product, uh, and would start reordering the tiles for you in terms of the menu options that you use the most often. Maybe color coding them, I don't know, maybe a bit of yellow or something to say, hey, this is what you normally do, uh, can we make it faster for you to get there? A little bit less, uh, mental, uh, cognition required by doing it as a recommendation or a hint for you. So, 80 the Chatbot, the idea is not to make her, uh, Clippy the paperclip, but, uh, make her something that I think Clippy was actually designed to do. So that was another Gone in 60 Seconds, or about 80 the Chatbot. I hope you have a simply magical day.