[Kevin 00:07]
So Zoe and I recorded a whole load of mini-interviews at the MVP Summit, and we thought rather than build those into a few larger episodes, we'd release those to you in short little snippets like this. So enjoy a set of our interviews live from Seattle with various people from Microsoft and MVPs. Hope you enjoy them.
[Zoe 00:15]
Hi, this is Zoe, we're live in Seattle for MVP Summit. Would you like to introduce yourself?
[Pragati 00:36]
Yeah, so myself Pragati, I'm a data platform MVP, and I work at Avanade. Of course, I'm here with Zoe to attend MVP Summit.
[Zoe 00:44]
Brilliant. Well, thank you so much for joining me. I hope you've had a good week this week. First question, what does CoPilot mean to you?
[Pragati 00:53]
Oh, yeah, I think CoPilot is everywhere. So I think in my view, Copilot could be like something which is really helping me to start developing something. Not from scratch. It kind of gives me like a pre-template already done, on top of which I can create and add more skills. So I really want Copilot to work as an assistance for my existing job.
[Zoe 01:16]
So helping you get started, but allowing you to use your own.
[Pragati 01:19]
Knowledge. Exactly. Helping me to get started, but giving me all the flexibility to customize it at different various levels, I.
[Zoe 01:26]
Guess. Brilliant. So, which co-pilots do you?
[Pragati 01:29]
Use? So, currently, mainly I work in Microsoft Fabric area these days, because I come from Power BI background, but now Power BI is basically Microsoft Fabric. So, like in Fabric, if you know copilot is available at three different areas for us. One is if you are into data engineering side, then you can use it on the notebooks. If you're from the Power BI side, you can use it on the data flow. And then, yeah, apart from that, it's at some other place. I just forgot about it. But yeah, sorry, but yeah. So, but that's the area where, you know, it kind of brings more skills and adds more value to my current job.
[Zoe 02:06]
Brilliant, and how are you actually using it? Like, can you give me examples of, like, specific things you've done where it's really added benefit.
[Pragati 02:13]
Yeah, so, like, I can actually point out two things. So, in one way, where Copilot has really helped me is, like, you know, adding comments to my calculations that I add on top of the semantic model. It kind of just generates like pre-created comments around the measures or calculations that I'm writing in. And then- We all.
[Zoe 02:32]
We all know people who do development, writing comments and notes and explaining.
[Pragati 02:36]
Oh, yeah. That's like the.
[Zoe 02:37]
That's like the most boring, hardest part.
[Pragati 02:39]
Yeah, that's the hardest part and where CoPilot is actually adding value and making my life easier doing that. And making you a.
[Zoe 02:45]
Better developer.
[Pragati 02:46]
A better developer, yeah, yeah, definitely. I mean, it increases productivity, right? I don't have to spend my time anymore in doing that because Copilot is doing that job for me. Then the second place which I would really like to highlight is like as I said, I come from a Power BI development background, but I don't have the coding background skills, like I'm not a data engineer. But in Microsoft Fabric, now I've started using PySpark coding language to write my notebooks with the help of Copilot. I can literally go and ask Copilot, can you tell me this function or how to do this transformation on my data? That's where Copilot helps me and gives me like a code snippet, which I can use and try to run and see what kind of results it's giving. So it's kind of helping me to develop my skills in a way.
[Zoe 03:28]
Brilliant, that sounds really interesting. So next question, is there anything that you're worried about.
[Pragati 03:35]
Oh yeah. I mean, definitely it's CoPilot, right? It's an AI thing that's happening for us. So, I mean, obviously, I mean, CoPilot is not right every time. So sometimes you have to think whatever results CoPilot is giving, you need to make sure you validate it because it's still growing, it's still improving because it will improve based on the inputs you give CoPilot, it will grow over time. So at this stage you really need to make sure that you validate your result because not every time it's going to give the right result, you have to be very careful about what you're doing with it.
[Zoe 04:09]
Yeah, I think that's a really interesting one as well because I've talked to lots of where they give examples of thinking copilot can help junior people get.
[Pragati 04:17]
Up to speed.
[Zoe 04:18]
Quickly. But actually, if you've got a person who's junior, how will they know that it's right? They don't necessarily have the experience.
[Pragati 04:27]
To know the.
[Zoe 04:29]
Code snippet. Or the example I had from another MVP was around PowerShell, where he said that he asked for PowerShell scripts and it was only correct two or three times out of ten. So when you think about that in the context of it helping a junior.
[Pragati 04:44]
Developer.
[Zoe 04:45]
How will they know that it's right before they just apply this stuff and like, oh, Copilot's.
[Pragati 04:50]
Given me it.
[Zoe 04:51]
So it's going to be right, I'm going to use it.
[Pragati 04:53]
Yeah, that's what it is. I mean, that's the risk. You can't trust Copilot blindly, you have to be sensible about what it is giving and if it is the right way to accept what it has given you. So make sure to validate things, possibly.
[Zoe 05:05]
And I guess it is Copilot, not Autopilot, so that human in the loop.
[Pragati 05:09]
Yeah, yeah, exactly, exactly. It's not autopilot, it still gives you results, but you still need to make sure you are working on top of it just to make sure that it's right.
[Zoe 05:21]
So last question, I'd like to know either what you'd like to see next in the world of co-pilots or if you had a magic wand and you could make anything happen that would make.
[Pragati 05:33]
It better.
[Zoe 05:34]
What would that be?
[Pragati 05:36]
I think something that I actually raised to product group team, like, two hours back. I mean, as a consultant, when I deliver end-to-end solution to a customer, the majority part of it is actually documentation, providing the right documentation to the customer about your solution, right? So currently, we see already Copilot everywhere, right? So I would really want Copilot to do some kind of magic or flick a magic wand, but generate some kind of documentation for me, which I can get started with. Because documentation is not the part everyone as a developer likes. No, no.
[Zoe 06:12]
It's.
[Pragati 06:12]
The hardest part?
[Zoe 06:14]
I remember working for an end-user organization a few years ago, and we had someone build some SharePoint website. Yeah, yeah. And my boss said to the third-party company, I'm not paying you to write documentation, I'm paying you to build a.
[Pragati 06:29]
Solution. Yeah, yeah.
[Zoe 06:31]
Is the bit that everybody needs.
[Pragati 06:33]
Exactly. I mean, for me, documentation is the most important part of making a delivery. And that takes the majority of my time as a developer, because I hate doing it. But I still have to do it, because when it comes to customer, you need to make sure their life is easier when you deliver that product. Because after that, it's in their hands. So you don't have to make their life explainable, supportable. It needs to be as detailed as possible. So definitely, I don't want Copilot to do it as detailed as possible. but if CoPilot can just give me like a starting template to start from what I've already developed, then it will be amazing for me. It's going to, yeah, make my life a little bit easier as a developer.
[Zoe 07:11]
Brilliant. Well, thanks so much for joining. I hope you have a wonderful end to the week.
[Pragati 07:16]
Thank you. Thank you, Zoe. Bye. Bye. Bye. Bye. Bye. Bye.
[Zoe 07:22]
Bye.
[Pragati 07:22]
Bye. Bye. Bye. Bye. Bye. Bye.
[Zoe 07:27]
Bye.
[Pragati 07:29]
Bye. you