Using Copilot for Model-Driven Power Apps

hey power rappers this is Brian knight from pragmatic works and in today's video we're going to show you how co-pilot is going to integrate with your future apps model driven app specifically in this video but also in canvas apps so stay tuned [Music] we're going to show you how do we integrate co-pilot into our model driven apps so that comes in three steps we're going to do we're first of all going to turn the feature on at an environment level second we're going to turn it on at a table level and then lastly we're going to turn it on at a column level to say what columns we want to be involved in a search so what it looks like is something like this when you open up your mile driven app or canvas app there's a component for canvas apps also I can basically ask questions of my data and you'll see that we'll play around with this after we get it all configured but you can ask questions here and it will come back with answers chat gbt or co-pilot is integrated in other ways as well they have the ability to help you build applications they have the ability also if you're in Dynamics to ask questions of your opportunities but here we're going to ask specific questions about our data these questions be really hard to answer unless you spend a lot of time building reports so let's start with step one step one is to turn it on at the environment level so I'm going to hit the gearbox up at the top right I'll go to the admin Center now for this you'll need a course admin rights to do this I'm going to go to my developer rights I say admin rights you need admin rights for the specific environment to do this okay so you don't necessarily need Global admin rights or anything like that I'll go to environments these are the environments where I'm an admin of and then after I choose my developer environment let me find one of these guys okay there it is I'm then going to turn it on in the developer environment it comes in two steps we have to do here our first step is to hit settings up top and we're going to go to product and we're going to go to behavior first and turn on the the refresh rate of this site how often are we getting updates in other words now that's the Timeless recording maybe if you're watching this a year from now in 2024 you may not need this step this is do you want to get the latest and greatest stuff this is a preview feature at time and it's recording in September of 2023. the next thing we'll do is we'll go to Features after that but let's start with the behavior and I'm going to make sure that I have set this uh model this small different app release channel to monthly channels I'm getting more up-to-date refreshes of it by default it'll be set with twice a year in your case but monthly means you're getting a little more the the features and latest and greatest then I'll go back to settings then go back back to product again and then the features next so I'm going to flip on these two settings you're seeing here allow canvas editors to access it and then also get AI power from all different apps as well I'm turning on both uh to give me the full co-pilot experience the one we're really focusing on now is this guy right here but both are kind of nice nice to have so while you're here you can also turn on data for searching if you type a search parameters in it will search multiple tables at one point as a little bonus and you can do as well after you're done go to the very bottom and save that so now you've turned it on at the environment level Now to turn it on the table level we have to turn on two things track changes on our table and it had this table participate in search results also so Find Your solution it has your table inside of it once you're in a solution go to your table and then go to properties of that table once here go to Advanced options and ensure that track changes is turned on Once you turn it on you cannot turn it off it's mainly used for doing things like change tracking and those kind of things and then down below appear in search results those two check boxes need to be turned on at the time that's recording for this to work all right that's step two done our final step is to turn it on at a view level so if I go over to views from my table this table is for a soup kitchen in my case I'm looking for the view called quick find view so it'll be quick find active whatever your table name is so when you select this this is what what happens when you're doing quick find searches and you'll notice in my case it's going to return like the event numbers not very not very awesome but we can actually change that and and we can say custom and only and show me hey what kitchen is actually getting the most amount of soups and those kind of things but what you really need right now is not necessarily the quick find form what you really need is on the right side is where you see find by and I'm going to hit this little edit right here and check what columns do I want to get extra intelligence to here so I'll go ahead and select the kitchen and maybe the gallons of soup and uh as well as the pounds of cheese and pounds of meat and there's one more I want as well and that will be total adults and total children served and to go meals let's go with that total meals there we go I'll just pick a few of these here and I'll get the volunteer leader as well now is only a certain amount of columns you can do this with before it eventually has issues once you're done hit save and publish and now these columns are now participating if you want to use dataverse searching you can easily do that as well all you have to do is go to the solution that holds these tables somebody who's to publish right now and once you find a solution that holds the tables you'll see a dataverse search search option right there so to show this real quickly I'll go to overview here then you'll see columns indexed and just go to manage the index here and just check any table that you want to participate in dataverse searching that's not required but we're doing right now but it is kind of nice what it does is it gives you an extra box up top so traditionally you only see a little quick view form here quick find form but it also gives you this box right here where you can search in number of tables in one click we are now ready to use this dataverse now so I'm going to do a hard refresh of my application there we go and I'll start to ask some questions I want to find out some things that will be a little tougher to find this is all soup kitchen data so I want to find out the average pounds of cheese served I'll hit enter but you can see on average we're serving 3.13 pounds of cheese now as you can imagine this would be a really tricky thing to solve you have to export this to Excel or create reports or create a chart to solve this problem but now able to do this in an ad hoc fashion now what happens if I make this query a little more complex let's say average pounds of cheese served and I'll say grouped by kitchen all right so you'll see the kitchens that we have right here the pounds of cheese I think is this guy right here and we'll hit go now also while this is searching you can give you these searches a thumbs up or thumbs down occasionally the co-pilot will get a little bit wonky with this results in this case you can see that uh it did come back with good results uh let's see Orange Park as far by serving more cheese and Middleburg is and uh and of course the other kitchens aren't showing up here so smart enough to eliminate records that don't have any cheese in this case if I had asked it a more natural way and said uh she served by kitchen it might get a little confused there so half the challenge is figuring out how to to phrase these questions in a way that co-pilot might be able to recognize that's the challenge with with any kind of AI right it's just becoming smart on asking those questions in this case it's funny because it did not work as easily uh when I was practicing it so it's already gotten better since I practiced this just a few a few days ago it's already getting smarter as I did this that question will give me back a uh no results so I give a little thumbs up in this case it's also important that as you do this if things aren't working the way you expect it it will it could potentially work down the road also so make sure you use these little thumbs ups and thumbs Downs to let Microsoft keep on honing your data I wonder what else I can ask it how about I go and say which volunteer organization has the most rows now notice I'm asking such a way most rows a little awkward right before when I asked this question it didn't know how to do that but you see Hibernia has the most most times they've they've volunteered in this case but if you asked you know different ways like if you ask it like which volunteer organization has served the most you might get a different answer again it's getting sharper and sharper as time goes on this this again was not the way a couple a couple days ago so it's in that little bit of time it's gotten smarter and smarter in here so let's ask one more question here which kitchen has served the most meals all right from this time it's hallucinating right the kitchen is not Hibernia it's basing contextual knowledge based on my questions I've asked earlier and it's applying that context to hear so it does not know you know that I want kitchen now this column right here so it's getting a little bit wonky in this case right Hibernia is not a kitchen so I want to ask at things like uh which kitchen uh has the most Rose Maybe and maybe I get a little better answer now in this case so it's as you can see it came back so fast I was like that doesn't feel right so this is a way you can kind of kind of better refine it in this case and as you can see now Orange Park has the most uh times they've served with 21 rows so when all else fails try asking Rose versus some more english-like kind of questions at the time of this recording this is only available in the U.S however it is going to be rolled out internationally of course as you can imagine uh very very soon so when your watch is recording and your region it might have already been deployed out to those regions also all right I hope you enjoyed this video on Microsoft co-pilots integration to dataverse if you want to see videos like this also more like this please do subscribe to this Channel and you'll find more training like this at pragmaticworks.com where we have hackathons virtual mentoring boot camps and also have recorded classes as well have a great day and thanks again for watching

Channel: Pragmatic Works

Views: 3,444

Keywords: power apps, powerapps, copilot, dataverse, power platform, copilot for power apps, microsoft power apps, artificial intelligence, learn powerapps, low code, powerapps tutorial, microsoft copilot, conversational ai, microsoft 365, copilot for makers, copilot for powerapps, power apps beginner, power apps copilot, app development, create a power app with copilot, model driven powerapps, copilot power apps, model driven app, model driven apps, pragmatic works, power bi copilot

Id: zIYN5AVFCpA

Length: 10min 32sec (632 seconds)

Published: Thu Sep 21 2023