Unlocking the Secrets to Scaling and Standardizing Power Platform Apps with Ryan Jones

Mark Smith: 0:32
Welcome to the Power 365 show. We're an interview staff at Microsoft across the Power Platform and Dynamics 365 Technology Stack. I hope you'll find this podcast educational and inspire you to do more with this great technology. Now let's get on with the show. Today's guest is from Seattle Washington In the US of A. He works at Microsoft. There's a partner director of product management. You might correct me on that. You can find links to his bio and socials in the show notes for this episode. Welcome to the show, or welcome back to the show, Ryan.

Ryan Jones: 1:07
Thanks so much, Mark, it's great to be here.

Mark Smith: 1:12
Good to have you on the show. This is probably the third time I've interviewed you, but the actual only the second time I've interviewed you on the podcast, the last one being in January 2019. So COVID will depart and it was episode 63. If anyone wants to go back and look at it, and we're at episode 466, so you're really one of our original guests.

Ryan Jones: 1:36
Yeah, I think I got it on the ground floor, right, but no, it's been exciting and it's weird to think about the before times, right, but it's awesome that we're still together in the Microsoft BizApps ecosystem and still chatting together.

Mark Smith: 1:55
Yeah, totally. I mean, back then we were talking about CDS. Now, every now and again when I'm presenting to customers, the common data service slips out. I forget to call it Dataverse, of course, or what it morphed into, and you were the lead on that product back in those days.

Ryan Jones: 2:12
Yeah, yeah, you know, led kind of from the common data service to what we now know as Dataverse. And then over the course of the last year, I did a bit of a tour of duty over at Customer Insights, you know, helping our customers better know, understand and engage their customers. And I felt a calling back to the land of Power Platform. And so here I am in Power Apps, you know, really focusing on scale right and really trying to focus and understand how do we help our admin scale right, like what does it mean for them to be able to manage tens of thousands, hundreds of thousands of apps? And then also like what does it mean for organizations to build mega apps right, like those apps that are gonna support tens of thousands, maybe hundreds of thousands, of monthly active users. You know those are very, very different problem sets that you know. How do we help, you know, suzanne or Robert go in and build their you know screen over the top of their SharePoint list right? Two very, very different problem domains.

Mark Smith: 3:23
You're absolutely right and I'm looking forward to unpacking that with you, because I get frustrated with the people, with folks that their belief only look at the Power Platform from a strategic or a stopgap measure. Right, they wanted to build one app, that's why they got it. They're not looking at it, dare I say, as an ecosystem, they're not looking at it as a. You know, when will we hit our thousandth solution on the platform? When will we hit our 10,000th, 50,000th, you know, and we've got customers out there in the market, right, that are doing those numbers on the Power Platform 50,000 automations. If you look at Slumberj, I think over 20,000 apps deployed on there on the Power Platform from them. So those are realistic numbers. I like that. You talked about hundreds of thousands then, because now I'm talking like, okay, let's peg it up another level Now. Back in the day when you first, you know, came into this team, I remember we're sitting in the office from memory in Atlanta campus having a discussion with you and we talked about the birds nest that you were unpicking from. You know the transition really, from what was the code base of Dynamics CRM back in those days and then across, and I find it interesting that you're back here now, as we're really seeing this scale up vision right Of really going. Hang on a second. This is a solution that the organization should be deployed. You know anything that's in Excel access, anything that you can't buy off the shelf why shouldn't it be going to the Power Platform?

Ryan Jones: 5:04
I think that's exactly right. You know, kind of our goal in our aim is for technical and business decision makers alike to feel comfortable, comfortable that they can standardize on the Power Platform as their starting point for where they build apps. You know, and I don't know, like I don't delude myself in that, like the Power Platform is perfect for every app out there, but it's probably a pretty good and a pretty logical starting place, right, and in that context, you know, we think a lot about how, like you know, you mentioned some of our customers that have tens of thousands of apps. Like, if I wake up every morning as an administrator tens of thousands of apps or hundreds of thousands of apps I'm gonna ask a very, very different set of questions. When I come to, you know, the Power Platform Admin Center or what have you, then I would, when I'm measuring, you know, like my 10 or not measuring, but looking at my 10 dynamics apps, you know, the metaphor that I've used a lot when this probably comes a little bit from my Azure background, is have you ever heard of the cattle versus pets analogy when it comes to managing infrastructure? Okay, so, like, I'll tell you a little story. Like, think about it like back in the day, you know, all of this probably did a little bit of like IT systems administration, right, and if you think back to when you had you know that handful of servers that were in you know your rack at your office or at home or what have you, you know, and I don't know if you're a Star Wars guy or a Lord of the Rings guy or what have you, but you probably named your servers with like clever names, right, like maybe it was named Nambu or maybe it was named Mordor or something like that, right, but, like you know, we're nerds, like we named these things, but they were almost like pets, right, you knew each one of them, you knew exactly what they do, and that's kind of like what we saw for folks that were like doing administration of like dynamics, environments, or maybe even folks that are administering Power Platform, but where they're still focused on Power Platform for like very narrow or specific use cases, and so we call that pets, right, because these animals, these environments, are things that you know and you name and that you take care of and you feed, but then where we see things going is in these worlds of tens of thousands or hundreds of thousands of apps. You're not going to name every environment, right, because someone's going to come in, they're going to get an environment, they're going to go to a party on it, and so we think of it as transitioning to cattle, right? Still animals, still very important things. I love cattle I think I ate some cattle last night for dinner but you know very, very different, very, very different in how you think of them, right, and you know some of this thinking. You know you see through, like the announcements that we've had over the past. You know months around. You know managed environments or power platforms, power pipelines, how do we help people with the tools to do things in a more automated way across larger number environments? But that's really, really central and core to our thinking right now. And I mean even to extend the cattle analogy a little bit further, right, when you think about cattle, cattle come in different shapes and sizes, right? Dairy cow is very, very different from a bull, right? You don't want to try to milk a bull, okay, and so when we think about how people may want to treat their, they may want to treat their production environments versus their development environments, versus their environments that support a hero 10,000, 100,000 now app versus the environments that they use to support, you know, their citizen development. Those are going to be different sets of rules, different sets of needs, and we give a sneak peek of this, actually, in our upcoming wave vision video. I think we published that here a couple of days ago, but it's something that's super, super interesting to us and you know it's all about. You know how do we do more cattle, how do we do fewer pets?

Mark Smith: 9:33
I like it. I like it. The Red Star was definitely a name of one of my servers, and then I've used Matrix. My current machine is Nebuchadnezzar from the Matrix.

Ryan Jones: 9:46
Do you have the little heroes and ones that like go down in the background?

Mark Smith: 9:51
I don't have that on the screen, it's just that's. It's. You know, machine name on the network. I know you know Nebuchadnezzar is drawing a lot of data at the moment, or, you know, using a lot of CPU and, of course, you know, with AI now it's going to be doing a lot more. As I download models and stuff to practice offline with. Tell me about where does the story of the COE starter kit right, which is developed by the cat team at Microsoft, and then managed environments? I see that you know in the last I think it's about 18 months now managed environments has been in play. Where. What is the thinking? What's the delineation? Do you need both? What are your thoughts or how do you explain this to customers, the two areas being that they are both heavily administrative type tools to use with the power platform.

Ryan Jones: 10:42
I think it's a great question, and I think, actually, sometimes people say do I, do I think about managed environments or do I think about COE? And I actually I don't think it's an or type of thing, and I'll tell you why Because of how many amazing features that you know, like I or my team aspire to build in managed environments. The thing is, a lot of what's in the COE starter kit actually came from the community, right, yeah, and so I think the thing is we will continue to see, once again, no matter how awesome I hope that managed environments become over the years, we're still we still have this thriving and active community that are constantly building and adding things into the COE starter kit. Now, I also think that there are a bunch of capabilities that are in the COE starter kit that I would love to come into the product Right, for example, a number of capabilities that historically had only been available through the ALM accelerator you can now do through pipelines, right, and so I think that these things will continue to exist and I think they're both still very, very important. So I would never say, or I would say, and and I think the other thing I will say is the ingenuity of our admin community is pretty stinking amazing, right, and so I don't want to be so air again as to say, hey, the best tools for managing power with platform come out of Redmond or Seattle, right? I think that in many cases, we're going to see people cook up their own tools for their opinionated way that they want to oversee their cattle, if you will. You know, it's just going to be a little bit different, and we want to make sure that we support that.

Mark Smith: 12:41
Nice, nice, tell us then, or explain to the audience, what managed environments is, what's the thinking, what's the strategy behind it.

Ryan Jones: 12:52
Yeah, so there was this like three word catchphrase that we had, and, oh my gosh, I can't believe it. All of a sudden it's escaping me. It's all about, like more control, more visibility and less effort, because, you know, kind of like a rancher right, you're a rancher, you got one rancher and you got tens of thousands and hundreds of thousands of cattle that you got to oversee, and that's really what managed environments are all about, and so we provide a set of capabilities, you know, out of box. The first thing is and you'll laugh, you know, because we're all drowning in email, right, but what we start is actually a weekly digest email, right, it's like, hey, across your power, platform, estate. These are the types of things that are going on, right, like here's some apps that haven't been touched in a long while and maybe you want to, you know, clean them up. Kind of like back in the day, you know how we used to get emails for, like, hey, this SharePoint site hadn't been used. Might want to delete that.

Mark Smith: 13:51
Yeah.

Ryan Jones: 13:52
You know those sorts of things, you know from there. We also, you know, provide capabilities around maker onboarding as well, where, when makers come and try to use power platform for the first time, it may be that you know, as an administrator, you know these are people building apps for the first time, you want them to read some documentation, you want them to complete some training, these sorts of things, and so, with maker onboarding, what you're able to do is you're able to present an experience in the context of the maker portal the first time makers build things, so that you can help them get started, start them out on the right foot. The other thing is pipelines is actually a capability of managed environments. Pipelines is our approach at providing a very low barrier to entry way of moving configurations between different environments. You'll notice I didn't say ALM there, because the thing is, most of our makers have no clue how to spell ALM. To be honest, I want these folks build maps. I don't want them to have to learn all the intricacies of ALM. Bunch of other capabilities as well, things like sharing limits, where administrators can say, hey, apps that are built within these environments, they can only be shared up to a certain threshold before. I need to know about it, because admins don't want to get surprised about an app that organically grew within their organization and now supports 10,000 monthly active users. And they get a phone call at 3 AM on Sunday morning because someone can't punch their time card or something like that. All sorts of goodies in there, mark.

Mark Smith: 15:40
Is there anything around connectors? Oh gosh let me see there are some things it just popped into my mind, and I know that there's a lot in the COE Starter Kit on that side, but anyhow, let's continue on. When it comes to the setup and configuration of managed environments, there's one of the common things that I hear and I feel it's once again from a the wrong mindset that comes out in the market is that managed environments there's a licensing cost that works on premium licensing and, for whatever reason I don't know whether it's been the journey over the last five years there's this fear of licensing rather than understanding, and I find it interesting. And the reason I find it interesting I came from a world where dynamics licensing was quite a bit higher than power platform licensing, right, and so when the power platform licensing come out, I'm like this is wonderful. I mean, the most expensive option is wonderful compared to my history of what I had to sell licensing wise. But now we come with people that have had seated via their M3, five environments and they're thinking about I'm looking into finance institutions that are exiled back into Power Apps and I'm just like what are you doing? You're putting a. It's not even a nicer front end in most situations because none of them know anything about UI and so it's not necessarily a pretty experience, it's just a shambles. And I feel that the discussion around return on investment net present value for investing and the kind of discussions that are not happening enough. Because if you realize that the amount of benefit you get with managed environments for example, a man of mine, keith Watling, one of the things he said you know you're pretty much eliminating one FTE from your admin team because that robustness is built in, rather than you having to check everything all the time. And you know as, when you get to those types of scales, how do we, how do you handle that discussion with potential customers around the? You know, oh shit, the licensing. And you know, in our mind we're like man, it's negligible in the scheme of things.

Ryan Jones: 18:07
Yeah, I mean, I think that one of we're seeing a sea shift right now in that, and I feel a little bit bad answering the question this way, but I'm gonna go for it anyway. You always love surprises in these discussions, right Like I don't wanna use a metaphor to explain it. So when I joined Microsoft I joined Microsoft in 2010, 2011, right as we were trying to get Azure off the ground, and I remember I would go to these executive briefings where I'd try to convince people and you'll laugh that they should consider using the cloud. Yeah, and the thing was, if I fast forwarded four or five years later, if I had an EBC, an executive briefing- with a customer who was looking for me to convince them that they should use the cloud. I didn't need to really invest too much time in that conversation because that person their replacement would be sitting in that chair six to 12 months later. Okay, and so I think that we are very, very rapidly getting to that place. When it comes to low code, okay, we're getting to the place where people understand the business value of low code and their ability to build apps more quickly, their ability to have more people build apps, for them to have oversight over all of those apps and, with all the things that we're doing with co-pilot, for those apps to be infused with AI capabilities, and so, increasingly, the discussion is shifting from should I low code? To how do I low code, and it's an exciting time, and I know I kind of didn't answer your question exactly. It's just a lot of the value I think at this point is landing as the low code category becomes more mature and we don't have to belabor that as much if that makes sense.

Mark Smith: 20:21
Yeah, it definitely has to happen quickly, because within five years, the low code will be dropped and it'll be no code Right. We will be literally saying to AI if it's got access to all my datasets, it's got connectors to everything, I'll be asking it to build everything I need and don't have to think about it, it'll be best practice, et cetera. So that's exciting times. You talked about AI then for a second, and I love discussing AI and how people are practically using it. So just peg that one in your mind, because I want to come back at the end and ask you how are you personally using AI in your day to day? How is it starting since November? How are things changed? But tell me about AI and governance. What are you thinking around that story there at Microsoft?

Ryan Jones: 21:09
Yeah. So I'm going to go back to my cattle analogy. If you think about how ranchers identify cattle, I grew up in Kentucky, so there are lots of farms there, and so normally there's a little tag on the cow's ear or what? Have you a few other variants that get used 100%, okay, cool. So the thing is, the difference in the cattle world versus the app world is the rancher generally understands when they have new caps, right, and they tag them, so on and so forth. But in this more pervasive low code world, where citizen developers are building apps all over the place, we have capabilities to inform admins that apps are being built. But one of the things that we've heard and I know this sounds silly and simple, but admins don't know what this app does, or what that app does, or what a connection does, or what a workflow does, or, in the context of pipelines, what a release does, and so one of the things that we're doing is we're actually using AI to look inside the box of the app, inside the box of the workflow, inside the diffs that are included in a release, and we'll be like, oh, this is an app that helps you manage your accounts and contacts. Or, in this release, we've added the new sort by, you know, age, contact view or something like that. Sorry, probably not the best example, but those are the sorts of things that, all of a sudden, if you think about the admin who's trying to approve a bunch of releases and pipelines, or is trying to understand across their entire state what apps exist and how they're being used, they're these simple ways that we think actually pretty dramatically make the administrators life much easier. Because in the past, when you do those things, someone you know sends you a release to approve. And what are your options? One, you trust what they wrote. Two is, more often than not, they put nothing in their release description, at which point. Number three, you have to fire up the app and figure out what the heck is in it. And let's be honest, man, we all got better things to do. We got soccer games to catch or beach walks or, you know, beers to drink, right. So I think that you know, once again, it's just a great example of a simple way that we can use AI to pretty dramatically improve the lives of our administrators.

Mark Smith: 23:51
I like it. I like it, so will it bite me in the ass, ai, if I release this? Will it bite me in the ass, yeah.

Ryan Jones: 23:59
I'll see if we can put that name in the yeah, yeah, yeah, I'll put that before the market.

Mark Smith: 24:04
Yeah, see if we get what you know. Yeah, exactly, Will it? Will it pass that muster? Um, yeah, talking about cows and air tags, I've put them in. I was born up on a dairy farm, so I know exactly what you mean. I know the latest ones all all, all RFID enabled right now days. They're not just a number that you need to look at, they they auto scan and stuff. It's pretty cool. Um, tell me, um, what else? What else do we need to know about managed environments, pipelines, AI, governance that we haven't covered today?

Ryan Jones: 24:36
So I think one of the things that was really interesting as we started, you know, shipping AI capabilities is we had a lot of customers ask a lot of questions around like well, I need, I need you know, this list of controls. You know the the length of my arm. And when we started going through that list of controls, you know some of the things that folks would say were like hey, I need you to tell me that when I use AI, my data isn't going to, like show up in another tenant, right, yeah? Or people would ask us like hey, I need to know that my data, my business data, is not being used to train your AI models. Or you know some some other things really around just how, how folks want to make sure that we're handling their data in a particular way when they use AI. And I think what was really interesting is it was a bit of an awkward conversation because I was asking these customers like well, your data is your data and never leaves your 10 boundary. We do not use your data to train our models. How am I supposed to build a feature that allows you to govern things that are in line? with their fairly conservative data policies to begin with. Yeah, one of the places that we've done is we've invested very heavily in documenting how Microsoft approaches responsible and compliant use of AI. The jury is still out if I'm going to need to add an admin setting with one of those little radio button switch things yeah. Locked to only keep data in my tenant and only don't use my data for learning or what have you. But our joke was are we going to put those controls in there, have them set to on and then locked in that way? Because it's a place where, in many cases once again, the way customers want us to help them use AI it's what we already do. I think that that's one of the biggest things that we've seen with the use of AI. Then let's see pipelines. I think what's really just been interesting there is seeing how customers are able to democratize some of these patterns around doing your development separately from your production stuff. Those are things that used to be I mean, you've been in MVP for a long time right. Those are things that were, in the past, difficult to figure out. Now we see just regular makers now able to build apps that get shuffled between their development environment and their production environments and, like I said, they don't even have to know how to spell ALM, and we think that's pretty huge and pretty awesome.

Mark Smith: 27:40
So true, it's interesting that whole discussion on sharing of data, because telemetry data is so important, because it informs so many decisions and I don't know if there's some separation, because I think there's always a need, even for admins, if you like, to get signal back from what is working and what is not working and what just went out of control, and so you can either shut it down and control it. So I think that whole telemetry and whether people are considering telemetry as AI, maybe it informs AI, but yeah, it's an interesting conversation, I think the more open it is, it's not like it's using the data from within a table the content that the system, but actually how the system is performing. Are you exposing a connection that you don't want to expose from a security standpoint? Those type of things I think yeah is so important. In closing, tell me about your personal use of AI. What are you playing with? What are you doing outside of Microsoft stuff to really bring your level of use of this? You know, since large language models have become available everywhere, it seems, what are you doing?

Ryan Jones: 29:00
Oh gosh. So in my personal life, ai the place I use it most often, so you'll laugh. We live in a house that was built in like the 50s or 60s and then there was a major renovation in the mid 90s, and when they did that renovation, for some reason our kitchen has light switches in six different locations, wow. And so the most practical use of AI in our household is we use ALEXA I'm styling it so that she won't wake up to control the lights throughout our home to the point that we don't even touch the light switches anymore. Also, we're in the process right now of designing a new home out in ocean shores, our little beach place, and there we become accustomed to kind of the automation enough that we're wanting to build more of that into the home from the start. And so I've been doing a lot of embedded systems programming and I have been using some of the tools out there for code generation. For that it's been interesting, I will say, because sometimes it absolutely blows my mind with how amazingly correct it is, and other times it blows my mind with how just, completely and unbelievably wrong it is as well. So one of our core principles actually is human in the loop when it comes to how we use AI at Microsoft. I think even in my personal life I see how important it is that we always have humans in the loop, because I mean, granted most of the time that embedded systems code wouldn't even compile, but I can't imagine a world where my lights start turning on and off by themselves and the window shades start flapping up and down while I'm in the shower or whatever. It'd be pretty wild, so really important to keep that human in the loop while we're using the LLLibs to generate code.

Mark Smith: 31:16
You mean, while we're training them right, Because at some point they're going to decide we don't want humans in the loop we're going to switch off the humans and we'll do our own thing. I mean, you know, I'm looking forward to it. I think it's going to create a very utopian future. I'm very hopeful of where this is all going to go. I think it's exciting times. The next five years are just going to be mind blowing, yeah.

Ryan Jones: 31:42
Yeah, I think we are on the precipice of something amazing.

Mark Smith: 31:47
Yeah, I just bought my first robot and I have about an acre and a half property Steep, quite steep, and it now out there mowing and, honestly, my property is starting to look like a park now. It's just immaculate. It does full hatch patterns. It has a 10 centimeter accuracy on the lines. It's just amazing. Quality job. And you talked about integration Can fully integrate to my home automation system. It's just incredible. And this is V1 of the solution. It's just amazing.

Ryan Jones: 32:25
I will say that I have thought about getting one of the robo mowers for our place out at the beach, just because we're not out there all the time and I'm too cheap to hire somebody to mow the yard. But I've also thought well, we have bear that sleep in the yard sometimes and I'm a little worried that the bear tried to destroy the lawn mower or vice versa.

Mark Smith: 32:49
I think it'll be fine because it's all got sensors, it's got sonar on the front. It can detect animals and stuff like that. No drama, check it out, luba. It just came. It's on amazoncom and the three models How's it spelled? But what are L-U-B-A?

Ryan Jones: 33:06
OK, Luba, I'll check it out Luba-B-A.

Mark Smith: 33:09
And the difference from all the other mowing technologies out there. This was built by a robotics company, where all the other robo lawn mowers were built by traditional lawn mowing companies that then are trying to learn robotics Right, and so it's a quantum leap, and the price point is not too bad, about 2 and 1 half thousand US.

Ryan Jones: 33:28
So it's like for a unit.

Mark Smith: 33:29
It's like the Tesla lawn mowers, yeah yeah, and it even looks like it. It's all white shelled and it's amazing. Honestly, I'm just like I go out at night just to see the lights down the field, because it can go. It doesn't matter. You know, it's super quiet and it just does its job and I thought it would not handle like the dew on the grass at night. No problem at all. Love it. Just great. Yeah, I think.

Ryan Jones: 33:54
I'm going to be stopping Anyhow with that I will end.

Mark Smith: 33:58
Ryan thanks again for coming on the show.

Ryan Jones: 34:00
Hey, thanks, Mark, it's a great chat, dude.

Mark Smith: 34:04
Hey, thanks for listening. I'm your host business application MbP Mark Smith, otherwise known as the NZ365 guy. If there's a guest you'd like to see on the show from Microsoft, please message me on LinkedIn. If you want to be a supporter of the show, please check out buymeacoffeecom forward slash. Nz365 guy, how will you create on the power platform today? Ciao.