Power Platform Boost Podcast

Shiny New $hit (#67)

• Season 1 • Episode 67

Share episode
Speaker 1:

But I'm looking for the next evolution of this where I can connect it to my drive and it will automatically back up some of these chats or, like you said, you have notebooks, right. What happens Suddenly, like you said, legislation, suddenly I can't access stuff. There's a data storage kind of separation, something. Something happens and it's lost in ether, I mean.

Speaker 2:

Yep.

Speaker 1:

Yeah.

Speaker 2:

We should be recording this.

Speaker 1:

We are, because I hit the button, because I've learned. Oh, you did.

Speaker 2:

Yes, I only see it now.

Speaker 1:

Okay, sneaky Because when you started talking, I was like, oh, I'm going to push the button, and I did.

Speaker 2:

Welcome everyone to the Power Platform Boost podcast, your weekly source of news and updates from the world of the Power Platform and the Microsoft community, with host Nick Dolman and Lidika Akebek.

Speaker 1:

Because I expect that when they created Plan Designer or whatever it's called this week, I mean, they didn't just start from scratch. They started from a lot of other solutions being made and a lot of other data models and they backtracked and they helped. You see, the manifest type instruction prompts that they're using for generated pages or code apps, it doesn't come out of thin air. They've spent a lot of time creating that prompt and, as we'll see in some of the news articles that we'll talk about today as well, prompting is now just and, as we've said so many times before, the quality of what you make is really very tied to, tightly tied to the prompts, the quality of the prompts that you're using. So, yeah, and with that, let's just dive into the list of news, because we have a big ton of news to get through, just as always. Yeah, do you want to start at the top with our good friend Thomas Hansen?

Speaker 2:

So Thomas, as sometimes he does, basically just pulled the pin out of a grenade and threw it into LinkedIn. And this is funny because this post doesn't seem to die, it just keeps. I keep seeing comments and things like that poking out from it, and basically it's linking to a post that he did. It's called low Code is Dead-ish, from Canvas Hacks to React Apps and really talking about, you know, basically calling out Canvas apps for the walking dead, because from here we're going to be able to build pro code applications using AI and sort of getting rid of that layer in between, because at the end of the day, like I think my point is, you're making code, at the end of the day it's going to be ones and zeros running on silicon, at the end it's going to be binary code. So this is a way of our makers going to be able to kind of utilize that working with professional developers, like, is Canvas apps going to be a thing? And it is? I think for now it is, but it's like, in the future, is pro code going to take over that? And so, anyways, we'll have the links and I'm sure a lot of people have seen this.

Speaker 2:

There is a ton of different camps appearing in this from the. Yes, you know, this is, this is the future, this is the way it's going to go. To other folks, oh so, no, no, low code is so not dead. Yes, ai is great, but we're going to be able to continue on with power effects and canvas apps and this will never go away. And to you know, the other side of like no, no, ai is, it's generating garbage You're not going to be able to maintain in governance and all the issues that we're talking about. So it's really a great article but really, in terms of thought provoking and basically, you know, thomas's point is his Microsoft is killing low code as we know it. I don't think they're killing it, I think they're revolving it. That's my take on it. Like, yeah, the days of the.

Speaker 2:

I think we're still going to live in a drag and drop world, but I'm not sure if it's going to necessarily generate that YAML code that we're seeing. I think we're going to see more and more of the. Yes, we could do drag and drop, but that's going to be working with AI tools and it's going to be generating probably more pro code. And I'm looking at, like you know, react or TypeScript or some of those existing frameworks that have been well established and I think that's sort of again, that's again my opinion and we don't know. But I think that's sort of again that's again my opinion and we don't know.

Speaker 2:

But we take a look at things like generative pages, code apps, single page applications on power pages. These are all kind of, and the common theme with them is they're generating React code. And I, in my video that I did last week, I dove into it just to try to get my head wrapped around it and it's like, yeah, this is very approachable stuff. It is pro code, but the tools we have now it's not to say I'm not going to suggest that makers are going to be generating, like coding, react directly, but they're going to be using tools that is going to be generating React code and we're seeing this already. That's sort of where the features are.

Speaker 2:

The future, I think, is going and I kind of agree with Thomas on that and in his very last line in his blog post he said low code isn't dead, it just grew up, and I think that's a very powerful statement. And it's growing up, it's evolving. So, yeah, that was, I think, quite the. It was a LinkedIn article, it was a blog post and there's still comments coming in. I haven't been able to keep up on everything like that. I'm not sure of what your thoughts are on this or if you have a different opinion than I do.

Speaker 1:

Oh no, absolutely not. I just yeah, just continuing on what you're saying, because I think this shows that the Power Platform is a platform that is evolving with the times. So you talk about lock-in effects and the platform. I think this is just proving that when everything else around is changing, the platform keeps up. So when wide coding is a thing and role-prompting and using AI co-pilot as a coding buddy, the platform keeps up and enables us to create pages for model-driven apps, new code apps from scratch, using the professional developer tools that the professional React developers have used forever, and also with PowerPages. The same thing.

Speaker 1:

And so I'm happy to see that the YAML kind of layer, like you said, is gone, because that spaghetti thing we didn't really need to begin with, and also I expect that we won't see a lot of investments in the drag and drop designer interfaces going forward, because I think the idea that the citizen dev is going to create an enterprise application is kind of gone, and if someone is going to create an application, it's going to be for the thing that they're going to use it for right then and there, and then it's going to be discarded until next time they'll need something similar.

Speaker 1:

They'll just buy themselves to another application or whatever it is that they need at the moment. We talked about this before when I read Thomas' article. I loved it because it was just echoing. If you listen back to some of our previous episodes, this is exactly what we've been talking about in the previous episodes, I find. So it was really nice to see that he kind of had the same take we did and, yeah, just keeping a kind of a finger on the pulse on this one really.

Speaker 2:

Yeah, and I think you as well. Thomas is our friend, right, so we've had a few conversations offline as well, kind of surrounding a little bit this about the future and things like that. You know a few little rants and whatnot. So yeah, anyways, and this like we're kind of it's interesting kind of leading with his thing, so maybe he'll listen to the podcast a little bit more often.

Speaker 1:

Yeah, because that's also lifelong. All of ours right To get Bridget and Thomas to listen to this podcast. The things we do you wouldn't believe. But it's kind of segues a little bit into another friend of ours, magnus Satter-Suttersen. He wrote a blog I'm sorry, a linkedin article saying that I've written less than 10 of my code in the last three weeks talking about having production code and from, as you know, ai generated production code, um, and how, and his experiences with that um.

Speaker 1:

So this is my one, this is I put this one in here, right, um, and he talks a little bit about what it's like to be a coder versus an architect as well.

Speaker 1:

So you might think that as a as a coder, professional coder you won't be out of a job for a very long time. But also when you try to use ai services for the architecture job, it kind of shows itself into a bit of a loop where it's it's caught in the it depends loop that we also find ourselves in. But as human architects we make decisions and we know from experience that in certain situations and what small little detail that is that makes us make a decision on the decision tree kind of thing and results in the decision, whereas AI is going to be stuck in that loop because it all depends and it can't really find its own way out. And that kind of speaks to our quality as humans in the loop being able to actually make those decisions. So sure it can help us along the way, but definitely not with everything. And he also wrote a tour, didn't he?

Speaker 2:

Let me just yeah, the XRM Bedrockm bedrock yeah yes, it's on github.

Speaker 2:

I from a very, very high level. I didn't, uh, dive too deep into it, but, um, yeah, it looks really cool. I definitely I need to explore that for some other projects I'm working on right now. So, um, yeah, it goes back to, I think, a lot of the you know what I see with coding, and you talked about prompting or what's you know they call prompt engineering or context engineering. A lot of even the stuff that I've done lately it's been.

Speaker 2:

It's like the prompt that you start with makes all the difference in the world in terms of what you're trying to build, and that was this is're trying to build and that was this is part of it too. And this is, I think, is again what we said. This is where solution architects are so important, because it's not like we're telling AI to take over for us. We're basically, instead of writing something in C sharp, we're basically saying here's what I want written. We're basically taking that C sharp, making it another layer up into natural language, but it's generating C sharp. And then the fact that we have this expertise, like Magnus does, in terms of interpreting that code, seeing what it does and verifying that it is doing what it's meant to do verifying that it is secure.

Speaker 1:

But we're doing it a lot faster because it's sort of like as opposed to where you know, building something from scratch something's getting built, but we can still tangibly look at it and make sure it's exactly what we want and then, if not, we go back. Yeah, so the just reading up just to see what the extreme bedrock is. It's a template for uh. So it's a project template to make it easier to work with. Dataverse and asher together, so it shows a modern development paradigm are applicable to Dataverse. This is too coded for me, but better in-code description, for example, to increase adoption, new way of handling web resources, cleanup Dataverse service and deploying data. So yeah, if that's something that rings a bell in your brain, check it out. And well done, magnus.

Speaker 1:

Yeah, because this just reminded me of the post that I saw from Diana Birkelbeck, the PCF lady, as she calls herself.

Speaker 1:

She made a time tracking generative pages with generative pages in Fluidline and in her experience, I think one of the important conclusions she makes is what we said in the beginning your app will only be as good as your prompt was, and that is very important.

Speaker 1:

And also it will not always listen to you. So if you want it to just adhere to FluentLine design framework, you have to be very adamant when you tell it to do that and give a very explicit description on how to do that and give a very explicit description on how to do that, and we saw that in Scott's video when he created this as well. The manifest, the document, the thing that he gives it as an instruction prompt, is like eons long, and there's a reason why he has to do that because it needs to be so specific for it to do what you want it to do, and so I think that just echoes kind of what we're trying to say with all of these posts that we're bringing up is that prompting is so important and shit in, shit out. I think that's the conclusion.

Speaker 2:

But and that was the thing about, like reading Deanna's posts and playing a little bit with generative pages as well Myself one of the I don't know the blockers right now that I see and I'm hoping they fix this, because I saw another post about someone kind of ranting about degenerative pages as well as yes, you put it in the prompt, you can give it all the prompting you want, it will generate that code. But you can't go in and tweak and edit that code directly. If there's something not quite right, you basically have to reprompt and then it goes back and regenerates something. Basically have to reprompt and then it goes back and regenerates something. So which I find okay, a little bit unfortunate, because a lot of times it's like okay, I just want to change that little thing. I see the code, I see where I want to change the code, I know the line I want to change or the thing I want to add. But this process at this, the way it sits this time, and of course everything's constantly evolving we can't not yet go in and edit that. I'd love for it to have the little Visual Studio Code button, like we have in Power Pages, where I can just click and then, boom, I can make those little edits, Because I did see a post and I meant to grab the link and unfortunately, because someone was complaining going.

Speaker 2:

Well, yeah, generated pages, you say it's so great. I tried it, but then I needed to move this and this, this. If I did it in Canvas apps, I would have been able to open and drag and drop and then and everything would have been fixed and like, yeah, fair enough, but if you're able to go and edit the code, you'd be able to do pretty much the same thing just as quickly to fix that particular thing. So, yeah, this is, but again, we're still.

Speaker 2:

It's still in preview, you know. So it's still. We can't move it in solutions. There's a bunch of roadblocks still in place. But I mean, but again, it's still important that we have people playing with it, People like Diana creating these things and posting it online, because this is the type of information Microsoft is going to take back and see this kind of evolve, and I'm hoping that we can at least get to some point where we can begin to apply this into projects, kind of that preview purgatory that we sometimes talk about, because I think that's where the rubber is going to hit the road. Are we going to get genitive pages or is it going to be degenerative pages?

Speaker 1:

so hoping for the genitive pages yeah, and did you see the latest post from yuka or not? Probably not the the latest, but he had a post about this.

Speaker 2:

The shiny new shit. Yes, it's a good segue.

Speaker 1:

Oh, fantastic. I just I love it. So he starts his post with it's like selecting JavaScript libraries in 2024. Trying to figure out how to build with Power Platform in 2025. Everything old is dead, but the new things will be different by next week and it's also in preview, right. So he has a fantastic blog post about this. Apps are dead. Co-pilots the UI for AI. Dendrotip pages will replace Canvas apps. Just chat with AI to build a UI. Canvas apps will unblock code for its development in Power Platform.

Speaker 1:

I love it and also, just like Thomas' post, this is a blew up and it has a lot of comments and, um, yeah, a lot of reactions. So, um and yeah, it's, and it and I feel this personally or professionally as well talking to customers, because it's hard to go in because to a customer conversation now and advocate for canvas apps. I find I find that really difficult. Uh. And in power pages, same thing. Or I wouldn't advocate for you going in and starting with one of the templates and start using that and build power pages on top of that, when I know the spas around the corner, right? So it is one of those purgatory things where you don't want to be an advocate for the traditional way of solving it, but the new way of solving it isn't yet out of preview.

Speaker 1:

So what do you do? But I think you kind of touched on it a little bit where we need to be confident enough to try and to start using the preview stuff without using it in production, and work with the product to always help, push and also give the feedback that Marks of Need, because they do listen to the feedback. We talked about this so many times. We have to keep emphasizing, keep using, keep testing and keep giving your feedback, because that's the only way that we'll be able to push this forward in the right direction, because, trust me, they will move it in the wrong direction if we don't. So it's kind of on us to help this along, I think.

Speaker 2:

Yeah, absolutely. And I mean yeah, to be fair, like it is, he lists off all the and, of course, like you, could like to stir the pot quite like this way a little bit, but it's also again, we've talked about this exactly what you said too. We're in a transition and it's it's easy to dismiss something like if, like plan designers or whatever like they call it, like creating plans or intelligent like I'm doing a session on this and I always struggle what it is I'm calling it this week, but to me it's like when I first looked at it in the first video, I did, like last december it was like, yeah, it couldn't do it but couldn't do a lot of things, but we saw the vision. December it was like, yeah, it couldn't do it but couldn't do a lot of things, but we saw the vision of what it was coming to. And I what I liked about it is it was forcing you to make sure you think ahead a little bit on building your app. You talk about creating big, long manifests.

Speaker 2:

Um, part of you know the plan designer as well. You can't just like, yeah, the prompts, like create me a, create me a, an HR app, just like, yeah, the prompts, like create me a, create me an HR app? Like, yeah, don't start there. Your prompt should be, or your context engineering should be at least a paragraph or two as a starting point. You need to think these things through and then, yes, it's still, you know, still missing a few things, but of course, we saw this week there's new updates, so it's already incrementally adding these things just to make it better. So that's the thing you're seeing the evolution and the cool thing is this is why sometimes things I believe it's a catch-22, right. So I believe this is why sometimes Microsoft will push something into a GA mode, like plan designer, which is technically GA, even though some of us might feel it's not quite ready for prime time, but we to get it's like anything, you need to start somewhere. You need to actually start getting people to use it, to get the feedback to evolve correctly.

Speaker 2:

Um, generative pages is another one that I think when it does go ga, a lot of people will probably say it's too early, but again, we need it in something like maybe not mission critical scenarios, but some production scenarios, to actually get the telemetry, the feedback, to make sure that these things are working. I know we're belaboring this point. So you see something like your shiny new shit. It's easy to read through these things going okay, this is all garbage. I'm just going to go back to creating my canvas apps and my model driven apps, like I did for the past five years. It's like yeah, no, you need like. Yes, it is crappy now, but it is. It's on a path and it's, whether we like it or not, if you're, you're gonna have to get on board eventually. Um with it.

Speaker 1:

So you say that, but I had someone bitching about the new interface for ppac the other day. I was like, yeah, they're still gonna remove the old people. I'm like what? You're still on the old, what you haven't. The toggle is like yeah, and it's going away. I was like really, really, you're just doing the old clunky, how it takes forever. Okay, okay, you know, people are like that.

Speaker 2:

It's gone and it's gone. I just looked back now. There's no way back to the classic it is. I think it's today's the day.

Speaker 1:

Yeah, that's right. He's probably crawled up on the floor crying by now, but that's another thing. I have like five different segues in my head as you were talking, as I'm trying to get back to my. Okay, so we need to separate two things. So something is in preview, it's not ready for production yet, but something like plan designer, what it is that you create, it's a data model, it's a flow, it's an app, it's a site, but those things that you create, they are GA. So when the product is released out of preview and into GA meaning you can use it in production, something like Plan Designer and you touched on this, I think, two episodes ago where you need to separate between a product and a feature. So when something's a feature and it's not, you can use it in production because what's kind of put into production is ready and has been ready for a while and is battle tested. That's another thing. So bringing plan designer or planning power apps, or whatever it's called this week out of G into GA and still evolving, it isn't that big of a thing for me.

Speaker 1:

And just think about Canvas apps, for instance. This week. You look at the Pulse. The advocacy team puts up this Power Platform, power Apps, pulse thing. The two things that came out this week was the user they had an acronym, udfs the user defined functions, which is you can create your own function, global functions and then you can call them using other functions. So it's kind of a way to and to be fair. To be honest, I didn't realize that that was in GA yet and I kind of expected that Canvas apps were able to handle that already, in my ignorance, because you know why, wouldn't it? Um, but that just goes to show you. I mean, how long have we had canvas apps? Like three years, four years, five years?

Speaker 2:

I don't know oh, longer, and it's still yeah, and it's still evolving, right.

Speaker 1:

So the fact that plan in power apps because that's the other new thing that came out on the polls this week was that now Plan in Power Apps is able to connect to existing tables in Dataverse. You ask it to create that HR thing that you made before, it would create a candidate as a new table. Now you can explicitly say for candidates, use contacts, and it will recognize contact from the standard data model and connect to that instead. I think that if they waited until they had all of this sorted out before they released something out of preview, we would be kind of falling behind. This is why this platform is actually pushing the bar and staying relevant and up to date. This is the price we have to pay for that to be possible.

Speaker 2:

Yeah, it would be like. The thing is, if it's always in preview, it would end up being abandoned where and we do see some of that, unfortunately because something's been in preview for so long that it just doesn't get the fee and eventually dies in the vine. Because they look at the back end telemetry and they're like going, oh well, no one's using this, so why are we investing money on it? Well, no one's using it because it's not GA yet. But then it's funny too, because, like when they, when they announce something like, oh well, plan designer now uses, you know, tables that are already existing, or you can specify the tables, and it's sort of like I don't know if that's something to celebrate or that's something that, ok, that should have been there a long time ago. Something to celebrate or that's something that, okay, that should have been there a long time ago. So, but it, there's a balance there, right, it's sort of like I don't know it's, uh, it's, I don't know. I, this is.

Speaker 2:

I mean, I, I sit here like I'm not, I'm not one creating the. I know the work that gets involved in creating these features more than anyone, because I was definitely heavily involved in it on the, on the sidelines creating the documentation a couple years ago. So I get the, the stress and the and the, the focus and uh, where the pressures come from and it's not always from mvps writing a blog post bitching about something. It comes from other pressures as well. Um, but the the cool thing with with uh, with that, with all right, where I was going with this. Yeah, there are some features that, yeah, the plan designer, but you said about the features that we assume that we're, we're um already ga, there's. Sometimes I'll go doing something and I'm looking at the, I pop over the documentation. I'm like, why is the preview tag still there?

Speaker 1:

like I thought this was I thought this was good to go.

Speaker 2:

uh, yeah, so yeah 100%, yeah.

Speaker 1:

And also I mean maybe all the development juices went into the new MCP wizard thingy.

Speaker 2:

I saw that this morning you posted that. I didn't know that that was a new one for me. That's awesome.

Speaker 1:

Yeah, I love it. The whole new thing is now the new MCP connector. The whole new thing is now the new MCP connector. Cp Connector Creation Wizard is now a public preview, which is a lot of fun. So what you do is, when you have an agent and you go in and you choose new tool, then you can choose the MCP and then there's a guided process that helps you no code, just config to create a new MCP that allows you to connect to an external data source. It just I have to have an endpoint url and then you can connect and, uh, generate that custom connector. And that's amazing news for all the no coders that think that mcp has been a bit too debby because you need a bit of tooling and you need a bit of this and that, and then you have to install things that NPM something, something they don't really understand. But this helps you do that.

Speaker 2:

Yeah, from a more config perspective, yeah, and there's a lot you can do Like this is the thing.

Speaker 2:

Like we're still trying to wrap our heads, I think, around what MCP server can do, but you imagine you can now interact with Dataverse through, like they give the example of the cloud desktop for some reason. But again, it's about how all of these other tools are beginning to very much interact so you can truly choose the best tool for the best job of what you're working on. You don't necessarily have to be necessarily tied to a Microsoft tool although like they're evolving, like all the rest but it does kind of open up a lot of integration scenarios and things like that. You know there's still some, you know some limitations and everything like that, and I do. I did ask the question to the team in terms of performance and things like that. Is it, how is it comparable to, like, connecting through Dataverse to an API or something? And they claim that it's comparable. But we'll see where the proof is in the pudding when we actually begin to see these in real life scenarios and that kind of thing.

Speaker 1:

Yeah, and then it kind of becomes down to the right tool for the right job type of thing, right? So if you just want to connect to data purely and then use a data-versed API, but if you need more kind of rag or you need something more around the data retrieval, you need more intelligence in that, then maybe the MCP is a better option, Because there is a, you have the concept. It's not just the API, it's the tooling and the resources around it that you get with MCP. That you don't get with the API, of course, which only has the one thing. So it's it depends, See. And also this reminds me because did you check out the latest episode for the local revolution with Eliza?

Speaker 2:

I haven't, Sorry. I did see the link, but I did want to check that out because, yeah, I've been collaborating with Eliza and a few other things.

Speaker 1:

Yeah, no, that's why I thought maybe you've seen it.

Speaker 1:

Yeah, it really opened my eyes to what you can do with agent to agent oragent or not. Open my eyes, but it was a fun setup, right. So you have another fellow MVP, sudeep Ghatak I think that's right, sudeep Ghatak, showing how he set up an agent-to-agent no, sorry, a multi-agent orchestration type scenario for HR, where you have one agent that is responsible for onboarding new people. You have a list of new candidates in SharePoint and you have a list of devices, and then you have a list and then you have a logging system in Dataverse and then it created a Canvas app on top of it so all the agents they report or create log entries into the Dataverse table. And then he had a Canvas app on top that showed the SharePoint lists, the Dataverse table. And then also he had an agent that was connected to hostelscom for booking restorations, right. So it was such a fun thing where he had just a trigger and the one agent goes off, so he just had the different candidates on the list and he'd choose so onboard or hire or something, and that agent then kicks off, hires the new person, assigns new devices to that person, finds hotel booking and kind of allocates all of that, and all of them always log their interaction between one another and he gave the agents of course he did specific names and you have to be very clear and concise with your naming and your instructions and he shows all the instructions and names and what he does and very clear and concise and tidy, which is part of the success with this, because then the agents understand what the other agents are doing and he had different profile pictures or icons for the different agents so then you can see in database afterwards this agent said that to this one and this one said that to this.

Speaker 1:

It became like a process. If you've given those agents real names and pictures, that would look freakishly like colleagues. I mean, the line is so thin on this one and I just had a conversation out in the open space here earlier today with someone working with an working um at a company doing h hr systems and HR. For some reason HR seemed to be the favorite use case scenario for placing humans with agents. Isn't that weird, considering that's kind of the most human-to-human interaction thing we have. If it was finance or you know anything but HR, but it's because of the LLMs, I mean it fakes. Because of the LLMs, I mean it fakes human interactions so well, yeah.

Speaker 2:

And I think HR, like I think finance, is the next one to get hit, because HR and finance they're very procedural type things. So you go into HR. Here's like HR, you know the you know for employee onboarding every company you know, and when I had my company we had this as well. We actually got an HR professional to build this for us. We had an onboarding manual and it's a binder and you go through these steps and a little bit of that AI flavor into it to kind of deal with a little bit of these exceptions or the particular set of data or situation like okay, this person is living in the next province over, so their tax implications are different, and blah, blah, blah. So having these agents kind of take care of all that in an HR scenario makes sense. I think finance is the same way Invoice processing, payables, all of that again very procedural. So I guess maybe there's a little bit of nervousness there because you're dealing with money and transactions.

Speaker 2:

But again, this is sort of the fact that agents can be employees. Like I'm looking at that for my one person business. I'm also looking at okay, I can do things as a solopreneur that I not necessarily could have done before because I can now invoke agents to do certain things that will take it off my plate. That's part of what I'm exploring right now, which is it's. It's interesting, also a little bit scary as well, but I think this also opens up the door for a lot more other things. So HR they can deal with, you know, they can kind of deal with more of those more important kind of human to human interactions and sort of leave the procedural stuff to the agents to do that kind of work and again. So talk about agents and things like that. Did you finish the agent academy, the recruit section, yet?

Speaker 1:

No, sorry, I have another time to do that.

Speaker 2:

Okay, so I wrapped it up yesterday actually nice again, a nice little. So, again, I know we talked about it this last episode again, but if you for some reason missed that or whatever and you want to get ramped up in agents, this content is by far some of the best I've ever seen. And shout out to the team. Of course, like I know, I I gave some feedback to eliza, a couple little things I ran into, and of course she got back to me, gave some feedback to Eliza, a couple of little things I ran into, and of course she got back to me right away to say, yeah, yeah, I'm going to make those fixes, whatever it to me, because the way it breaks down the explanation and then gives you that hands on stuff to try, yeah, I'm looking forward to the next two sections and I've sent that link to a few people that have been like, oh, we're trying to figure out agents and this, and that Like, here, go through this, this is, yeah, so sorry that was a segue there.

Speaker 1:

No, no, it's perfect. Yeah, absolutely love it. So I'm it's on the books for the next this week, actually to get to get through it. So. But you know what it's like? I have all this free time on my hands and then suddenly there's something coming in left, right and center. The list of to-dos and things to dive into is just too long. I need to learn how to say no to things, I think, but I'm going to try to finish that for sure. So, in terms of other news, I see there's a few items here that we haven't talked about. That was on your list. So you do want to dive?

Speaker 2:

into one of the first ones here. Yeah, so a little one about our friend joe gill in in ireland. Uh, again this. This to me was just sort of a you know, as much as we talk about ai, co-pilots, all the the new shiny shit, let's look at some of the the day-to-day stuff of people solving problems. I think that's also a core of this, this podcast and for class so title for sure oh, shiny new shit. Oh yeah, I've already written it down easy, so thank you, you got oh yeah for sure.

Speaker 2:

Um, joe joe gill, he wrote some of power apps, real-time updates. Um, basically, and it's true, he said power apps and power pages specifically, they don't support receiving information real time and they were kind of rely on polling, meaning they got to go and check something. Is there new information? Is there new information with a checker, with humans, whatever? But anyways, he's implemented a method using WebSockets to have a persistent two way connection using SignalR and Socketio for Nodejs. So this is very code heavy but it is pretty cool. He supplies the code and his use case is interesting. It is about parking fines, so I'm kind of guessing he might be experiencing this directly. There might be a story here. Joe, did you get a few parking tickets lately, maybe? But anyways, let's go with that.

Speaker 1:

Story with us, please. We'd love to see the pictures and hear that story. We're going to see them at Nordic, I'm sure, so we'll get the whole bang there, hopefully, yeah, oh yeah, cool, okay cool, oh, that's awesome, yeah, so.

Speaker 2:

And then another one no-transcript. This is truly not junk. And then there's other stuff that I've gotten I'll get in my regular inbox, like this is this is garbage. Why am getting this? Like you know, the nigerian prince you know that has that 15 million dollars sitting in a holding account with my only I can somehow cash it in I thought that was just me oh, maybe, oh, maybe you're lucky like so, but anyways, um, our friend, polina cold she's, uh does a.

Speaker 2:

She's actually blogging a lot lately on customer insights and journey, which, I will admit, I am not an expert in. I know what it does, but that's basically the extent of it. But she had a really interesting article on why emails fails to deliver and what you can do. So she made an email bounce cheat sheet to help those people working with customer insights and journeys make better email email campaign capabilities and things like that and some really good, really good tips and tricks and things like that to really make sure that if you're firing off emails, that they make sure they get to the right people in the right time into their inboxes, not into their spam. So thanks, polina, for that. Like I love all your content and I you know, sometimes I wish I knew more about Customer Insights and Journeys just to understand a little bit more. But yeah, she does amazing content, so check out that post plus all her other ones. She's one of these. She's been an MVP now, I think, for one or two years now and just amazing content.

Speaker 1:

Yeah, absolutely, and she does amazing sessions as well. So if you get a chance to see her at one of the upcoming events I know that she does a lot of events, especially in Europe then make sure to attend one of her sessions. It's always gold, absolutely. And speaking of small little helpful things, I saw a post from Radovan Santa I need to practice these names but that was about auto-categorization. So you have a Canvas app where he just had some sort of data entry and you wanted an agent to automatically categorize that entry form and I find so this is one of those things where I had an expense sheet type automation, where I would upload an image and I wanted an agent to kind of pull the information from that image and just automatically edit by it, but then also the expense type, right? So how do you determine what type it is? And then, of course, you use ai to do that. So you have an agent going in, you tell it all the basic details about the data entry and that will automatically choose the right category for you.

Speaker 1:

So this is just a small little ones, the, you know, the linkedin, the type of things that you click through, just showing the diagram. Yeah, yeah, yeah, there are documents so showing kind of the diagrams and little illustrations of what he called at what time and what the services did and how. It kind of got back to the Canvas app that he used to create this. So small little thing, small little tip, but also unlocking. If you're kind of looking at doing this but didn't know where to start, this is a a good way, good blueprint to adopt just to try it out yeah, awesome.

Speaker 2:

No, that looks cool. This is exactly what I'm. One of the things I'm working on right now is just to uh manage my expenses when I'm, because I'm gonna be traveling again.

Speaker 1:

Then I'm gonna get into all this expense and following up and the whole bit so what if you could just have a folder with images and every time there's a new image, AI goes in and creates a new data verse entry just shows a date and amount and automatically categorize it for you. And then, once you have two or three or more on a certain date, it kind of pulls it together and goes did you go to Vegas by any chance? Cause this is an expense report. All of these lines in it, oh, that's the dream. This is an expense report and all of these lines in it, oh, that's the dream. And then, if I could just and I actually I almost don't tell anyone, but I kind of the other day tried to hack my way into kind of pushing data into our own internal finance and operations system. I removed all my tracks because I couldn't do it, but I tried.

Speaker 2:

Good, that's perfect. So you don't even need to think like to me. I just want to have a receipt, take a picture, Don't need to worry about it. Off to the accountant.

Speaker 1:

Fantastic.

Speaker 2:

I think we're close.

Speaker 1:

Oh no, it's doable. I mean, if we can book hotel rooms for people. I mean, come on.

Speaker 2:

Cool, I'm just seeing what else we I know we've hopped all over the list I do want to talk a little bit about actually. First this one about the child response to a parent wants to make it faster by. Johan, that was yours. You want to talk about that? That's a technical one.

Speaker 1:

It's a small one. It's on the same lanes as the things we've talked about already, because it's fun to talk about these huge AI existential crisis things, but then we kind of also want to bring it down to the nitty gritty little details, and this is one of those things where Johannes Anstrum he have a LinkedIn post. It's also one of those carousels that talks about how you can actually make child responses faster. So you have parent and child power, automated flows, and usually what you do and which I do all the time is that you call a child flow. That child flow runs all the way to the end and then you have the response to parent action. But actually what he does is he responds to the parent at once.

Speaker 1:

That's the first action in the child. What happens is the parent gets noticed, oh, the child is done and it continues. It doesn't wait anymore because it will wait for the child to respond. So then they move simultaneously down and finishes. So the child will move independently of the parent and they'll finish on their own time. So it's a way to make your child kind of parent flow dynamic go a bit faster, so it'll be more snappy. Of course, you won't get the result back if you do that. So, and you won't know if it fails or if it finishes. So it's just for certain use cases, but I thought it was worth mentioning.

Speaker 2:

Yeah, that's definitely a great pattern. I definitely probably will steal, borrow or be inspired by that.

Speaker 1:

Yeah, for certain scenarios it's very good, and also something else I came across that I wanted to mention, if I can, something called msportalsio which was new to me, but it's a collection of links to all the Microsoft portals. There's a few links on this list. I'll tell you they're not few, but this is a list of all of those. So yeah, just a little collection there for you.

Speaker 2:

Yeah, no, that's a great resource. I'm bookmarking that right as we speak, because that is actually. I do have a bookmark already. So yeah, for sure.

Speaker 1:

There you go.

Speaker 2:

Good, good resource. I did want to talk a little bit about a new podcast with Mark and Meg Smith. I think everybody knows who Mark Smith is. You probably, if you know Mark, you probably have met Meg or know Meg as well, because she's working a lot with Mark on things like the 90-Day Mentoring Challenge plus some other initiatives as well, and they have started something called the AI Advantage Podcast. Now, everybody knows Mark Smith. As the you know, he's been doing a Power Platform Dynamics 365 CRM podcast for quite a long time maybe not as long as someone like Marcus, but yeah, but still, mark's well known for his podcasting stuff.

Speaker 2:

So they're doing a podcast together. It's interesting. It's all about their sort of their journey into AI when they started using it. They're different. It's nice because they have very contrasting approaches to it. If anybody has met Mark, mark is already kind of his name's at the top of the list for the next AI implant that they're going to do. That Meg's a lot more kind of on the opposite side, a little bit more cautious, a little bit more scared, but trying things out and realizing the journey as well. So this is a so it's an interesting podcast. They're going to have a whole series of these. I did reach out to both of them and kind of congratulate them on this and they said they said they're stealing our format by just having the conversation between the two of them. They're not doing interviews and things like that. So I took that as a compliment to us because that's sort of like our format there, which of course we stole from lisa and megan. But anyways, it's all.

Speaker 1:

It's making your own. Oh yeah, definitely, that's what we're all about so check out the link.

Speaker 2:

The link is there, it's on spotify and all the other podcasts. So, again, it's like you know, there's plenty of uh, plenty of things to do when you're out on your run or doing your gardening, you, you're off to the gym or or that kind of thing. In terms of podcasts, a lot of great stuff coming out to help you kind of stay grounded and inspired and not be all doom and gloom versus the. You know, ai is going to save the world to. Ai is going to destroy the world by 2050. If you've watched the latest Diary of a CEO podcast. But anyways, we'll continue on from there. So, yeah, so check that out. And then, yeah, we saw you saw you posted it here, but I saw that as well Mark Christie did some arts and crafts. You want to talk about that?

Speaker 1:

I think Mark is already bored by staying in Norway because we are a weird little introverted bunch. So he decided to make a poster out of his lanyards from different events. So he posted a big picture of the MVP and then kind of that outcut of the MVP logo.

Speaker 1:

He had all his lanyards lined up and it was really cool and colorful and then you could all see all the events that he went to. So I thought that was a nice use of the lanyards that we all have hanging, like you do, on the wall behind you or the door behind you, which I also have on on in my office as well. So that was just a good, creative thing and also sorry.

Speaker 2:

Yeah, I was about to say if any of our listeners have done creative things with either the lanyards or the cards or the t-shirts that we get or any of the swag, that's a little bit different. Um, like you talk about, we're talking about polina earlier. Polina is a big knitter. I I learned this about her, uh, this spring. Um, so if you're doing stuff a little bit like like this, like the arts and crafts stuff, share it with us and we we definitely love to highlight it well, yeah, yeah, 100%.

Speaker 1:

We absolutely love that. And, pauline, I'm just saying maybe your duck is going to be a bit cold in the wintertime, so maybe it should get a sweater, not to? And speaking of ducks, this is disgusting. This is like my episode. I saw something that I wanted to ask you about. So on my LinkedIn feed, suddenly there was something called Rubber Duck Tuesday by GitHub and it was a kind of a live thingy. So I looked into it and it was one of those. I tried to read up on it and it seems like there's just this scheduled thing happens every week Open table, anyone can join in, it's live and it's recorded and you can ask anything. Use anyone as a rubber duck for anything and just, yeah, open mic. Oh my God, awesome. Have you seen that before?

Speaker 2:

No, I saw the link when you put it in there. I had not seen that. But yeah, I just sort of checked it out. It was like live last Thursday was the last one.

Speaker 1:

So, yeah, sort of checked it out. It was like live last thursday was our last one. So, yeah, interesting, it's so cool, very, yeah, I know. Oh, yeah, thursday, not tuesday, sorry, um, that's my mistake. So anyone have anything and they needed someone to rubber duck it with, I guess you just, yeah, join that call and ask away. Hopefully, if that helps you, let us know because that would be a lot of fun. All right, so, um, this I just wanted to mention, in terms of um events before we go, that nordic summit is right around the corner.

Speaker 1:

If you haven't got your ticket yet, please do. Uh, we're both going to be there. Uh, I'm going to be interviewing people, so I'm going to run around with a microphone talking to all kinds of people. We're trying to get to do a live recording of the Booth podcast while we're there. We're going to do a recap anyways. So something's going to come out of Nordic Summit for Booth, anyhow, and then we're just going to be there and have lots of fun and you're going to do the closing keynote, which I'm so excited about, which is going to be so much fun.

Speaker 2:

Yep, yeah, I'm actually starting my tour tomorrow. I'm flying out tomorrow afternoon actually, and first stop Finland collab days, then Baltic Summit, which will be my first time in Poland. I'm looking forward to like getting to know the people in that community a little bit more, and then an Experts Live UK and then from there Nordic Summit. So a lot of back-to-back events. But, yeah, it's going to be. It's going to be definitely a very fun trip. Really looking forward to it and doing a Nordic Summit yes, closing keynote, which I'm yeah, it's going to be exciting. It's going to be looking for very friendly faces in the crowd because it's kind of a very personal topic for me and I'm looking forward to that support as well.

Speaker 2:

But I'm also doing a session with Manju Gharar Gharar on Power Pages and we've been collaborating back and forth and Manju is just so enthusiastic it's very infectious. So looking forward to doing that session with her. She gets so excited of these things within Power Pages. So, yeah, it's going to be a fun one. So, yeah, and get your tickets. People like don't wait, because Guru, two years ago she was chasing people out the door that didn't have a ticket, so get your ticket.

Speaker 1:

Yeah, when we were in Copenhagen, we sold out and she had to turn people away at the door. So get your ticket. Don't just fly in and think you can walk through the door. Get your ticket now. It's September 19th to 20th. So, yeah, definitely looking forward to see you there and I think that's a wrap up for this episode. Thank you so much for listening and we'll catch you on the next one, see ya. Thanks for listening and if you liked this episode, please make sure to share it with your friends and colleagues in the community. Make sure to leave a rating and review your favorite streaming service and makes it easier for others to find us. Follow us on the social media platforms and make sure you don't miss an episode. Thanks for listening to the Power Platform Boost podcast with your hosts, ulrika Akerbeck and Nick Dahlman, and see you next time for your timely boost of Power Platform news. We'll see you next time.

People on this episode

Head Shot

Nick Doelman

Host
Head Shot

Ulrikke Akerbæk

Host