Top Down and Bottom Up DevOps – Ep. 434

Mike, Tommy, and guest Matias discuss the top-down and bottom-up approaches to adopting DevOps in Power BI and Fabric teams. From convincing leadership to grassroots adoption with Git integration, they unpack practical strategies for building a DevOps culture in BI organizations.

News & Announcements

• Introducing MCP Support for Real-Time Intelligence (RTI) — Microsoft released an open-source Model Context Protocol (MCP) server for Fabric Real-Time Intelligence, enabling AI agents to execute KQL queries against Eventhouse and Azure Data Explorer backends. This is the first MCP integration for Fabric, allowing tools like Claude and VS Code to seamlessly interact with real-time data. Support for Digital Twin Builder, Eventstreams, and Activator is coming soon.

• New Item Creation Experience in Fabric — Microsoft is introducing a revamped item creation experience in Fabric that reduces clicks and lets you select the destination workspace and folder upfront. Mike and Tommy discuss how this addresses the pain of disorganized workspaces, with Mike admitting he’s abandoned favorites entirely in favor of just searching the full item list.

• Microsoft Named Leader in Gartner Magic Quadrant for 18th Consecutive Year — For the 18th year running, Microsoft has been positioned as a leader in the Gartner Magic Quadrant for Analytics and Business Intelligence Platforms. While the streak predates Power BI (which has been around for about 10 years), the last several years have been driven by Power BI and now Fabric’s dominance in the space.

Main Discussion: Top Down and Bottom Up DevOps

Mike, Tommy, and returning guest Matias dive deep into the practical realities of adopting DevOps practices in Power BI and Fabric teams. The conversation centers on two distinct approaches: leadership-driven (top down) and grassroots team-driven (bottom up).

The Top Down Approach: Leadership Driving Change

Matias argues that in an ideal world, DevOps adoption flows from the top of the organization downward. Key ideas include:

• Embed engineers in business teams — Rather than keeping IT in a silo, place data engineers directly within the business units they serve. Mike shares a personal experience from his corporate days where an IT resource’s salary was paid by the business unit, creating direct accountability and faster delivery.
• Treat dashboards and reports as products — Adopt a product mindset with a product manager, customer focus, documentation that makes sense to end users, and investment in training.
• Cross-functional teams are essential — You can’t produce meaningful insights with an engineering team “tucked away in the basement” detached from the business.

The Bottom Up Approach: Grassroots Adoption

Tommy argues that DevOps is so technically specific that it’s more likely to be driven bottom-up by passionate practitioners rather than mandated from leadership. The team discusses practical strategies:

• Start with pain points — Ask your team: Have you ever deployed a report with wrong filters? Had someone overwrite your changes? Lost work because SharePoint couldn’t merge? These shared frustrations build the case for change.
• Let people experience the pain — Matias suggests letting teams encounter the consequences of not having source control and automation, then stepping in with solutions when they’re ready.
• Be a savvy leader — When failures occur, use them as teaching moments to introduce DevOps concepts rather than forcing adoption.

The Practical First Steps

The team converges on a clear progression for teams getting started:

1. Move from PBIX to PBIP — This is the fundamental keystone. Understanding that your model and report are a collection of files (not one big zip) is the “aha moment” that unlocks everything else.
2. Replace SharePoint with Git integration — Turn on Git integration in your Fabric workspace. You don’t need to understand branching yet — just commit changes and see the history. This alone gives you rollback capability and change tracking.
3. Introduce an issue tracker — Use Azure Boards or GitHub Issues to formalize requests from business users, track priorities transparently, and create a shared backlog.
4. Then introduce branching — Once Git is comfortable, map branches to business requests and sprint cycles. Feature branches follow naturally from the issue tracker.

The Environment Challenge

Mike raises a critical gap in Microsoft’s current tooling: the difference between branches (cheap and easy) and environments (expensive because they need data). Creating a branch workspace means re-running all compute to populate data, which is a massive friction point. Matias acknowledges this is precisely what his product Navigator focuses on — automating environment management so teams aren’t stuck with manual workarounds that violate DevOps principles.

Looking Forward

The team agrees that DevOps adoption in the BI space is still early — it’s not yet a universal expectation like it is in software engineering. But the trajectory is clear: Git integration, PBIP format, and AI-assisted workflows (like AI-generated commit messages) are lowering the barrier to entry. The SharePoint “safety blanket” era is ending, and teams that start with simple Git integration today will be well-positioned for the more sophisticated DevOps workflows coming tomorrow.

Episode Transcript

Full verbatim transcript — click any timestamp to jump to that moment:

0:26 Good morning and welcome back to the Explicit Measures podcast with Tommy and Hello everyone. Welcome back to the show. Hello Mike. How you doing? I’m doing well Tommy. Thank you very much. All right, we’ve got a another packed hour again today. Some more great topics for you today. today’s topic will be around top down and bottom up DevOps. what we’re going to try and talk about here is when you need to use DevOps in your organization, how does the the business team or the the people developing in PowerBI, how do they communicate up to leadership that we should spend time learning DevOps?

0:59 Like what what are some of the gotchas here that we should be thinking about? And then also maybe you’re leadership and you want to employ or get add DevOps to your teams and they’re not software development teams. They’re they’re in the business, right? So, how do we communicate the value of what DevOps is doing to our organization? So, we’re just going to have some candid conversations around what we are hearing in these conversations and what we think works well and top down and bottom up. Oh, yeah. That being said, Tommy, , let’s go through a quick news items. I think you found an article here. Let’s

1:30 Unpack it. What do we What did you find first, Tommy? Oh, man. We’re going to go from least coolest to coolest. I don’t know. They’re all cool, actually. This is really jam-packed news. So this is the first time I’ve seen for fabric by Microsoft. There is MCP support for real-time intelligence. MCP is what’s known as the model protocol or model context protocol. Yeah. And basically what that allows you to do is in your other tooling like claude or if you use

2:01 Like VS code but there more and more clients or ids are supporting it. basically allows your agent or whatever you’re using that’s using AI to basically get the context of that service. So for example, some really popular MCPs right now are like one’s called like thinking so it makes your agent think before it does stuff. One keeps memory and then others connect to services like notion, GitHub. Now, this one is by Microsoft, which allows

2:32 You in any of your AI tooling that supports MCPs to simply execute KQL queries across it and be able to do some more plug-and-play. So, that’s pretty cool. It’s the first instance we’re seeing that. That’s wild. , again, I know the term MCP. I’m trying to figure out how this works a little bit there. , my understanding, Tommy, again, correct me if I’m wrong here. An MCP server is what you need. So, , you’re talking about there’s a large language model that

3:03 You’re talking to, which is typically running on a machine or service, has a series of APIs. Then there’s the MPC server, which is another machine or computer that’s taking commands from the large language model and communicating with that to actual hardware. And so, , I was watching a quick video yesterday of someone, , off of Y Combinator. I I don’t know if you’ve heard of that. Yeah. Like a tech space or something like that. and he was saying he he was talking about his vision for AI and where AI is going. He

3:34 Goes, “The newest hottest programming language is now English. Your your native your native language.” I’m thinking to myself, maybe we’ve gotten to that level. Maybe we’re not, , we’re not we’re no longer writing code for anyone at the lower levels of code. We’re able just to talk to it in English and that’s how we’re going to program and code computers moving forward. This is a really neat new world. And I think this MCP thing is starting to appear even in Microsoft tools. So this is I this is going to be a trend. I think I think we’re going to hear a lot more about this. I’ve been

4:05 Using them almost exclusively with anything I do with the Vibe coding type of thing. But again, it supports flaw desktop as well. So excellent. Nice. All right. Sweet. Any other news items you’d want to cover off here before we jump in? So Microsoft fabric they are just introducing a new item creation experience in fabric. So whenever you want to create a new artifact there’s a lot to choose from. maybe you’re not sure if you’re in the right workspace or in the folder and everything gets cluttered. So this new

4:36 Item creation experience simply just a UI is supposed to pinpoint those pain pain points headon. Basically what it allows you to do is create new items with fewer clicks. You can select the destination workspace and folder from that new creation experience and you can assign it a task within your workspace. So yeah, I think this is an I think this is letting people encourage them to use folders and task flows. I honestly I don’t feel task loads are too valuable to me. It’s very,

5:08 , fictitious or, , again, it’s it looks it’s useful for like the first couple times you use it, but I don’t I don’t really feel like I love it. Of the two items, I really like folders. Folders seems to be a lot better, especially when the or when the workspace gets very disorganized with lots of things in it. I really like the folders experience. I think folders makes a lot of sense here. And the new dialogue box, anything you can do to make it easier for me to create an item. And I think honestly all items no matter what it is should adhere to this new experience. It should just be the same.

5:41 Like one of the things that bothers me right now is I make a data flow, it just makes a data flow and just gives it like data flow one data flow. Mhm. If I click dataf flow and if I click a notebook or something else, I want it to like just force me give it a name. Yes. Where it goes and then and then just use the name I give it. So that way every experience, no matter what I’m creating in Fabric, has the same, , entry folder experience. I think that that makes a lot of sense here. They’re they’re polishing some things here that need to be done. I absolutely love it. So, I’m not sure if that’s rolled out

6:14 As of yet. It didn’t it’s saying we’re going to roll this out, but they didn’t give a time frame. Yeah, it’s it’s it’s just just noting that it’s coming and they didn’t say it’s actually here yet. There is a little animation here that shows you the new dialogue box where you can pick you could pick a different workspace. You can pick the the name. I think you can even add sensitivity labels to things as well potentially here. So, there’s just going to be a better way of creating content. I still think the whole add new item experience is clunky with all the So, here I’ll just

6:46 Say one thing, Tommy. Recently, I have I’ve abandoned all hope in the new item creation experience. I no long Yeah, I no longer favorite anything. And so, I used to go through and favorite a couple of the items that I want to look at, but there’s so many new things happening now. I hate the new item experience. It just drives me nuts. So, I’ve unfavored all the items. I had like a lakehouse. I had a semantic model. a couple things, but it was so annoying because every time I wanted to grab something that wasn’t in my favorites, I had to go over

7:17 To the other list, type in the name of the item I was looking for, and then go find it or just scroll through an infinite long list of items here that I was going to create. I didn’t like that. So now I have removed all of my favorites, and now I only have I only have So now when I click the new item, it goes right to the long list of everything. And then I can just type in what I’m looking for. Semantic, model, notebook. So, I find it’s actually faster now to go type in what I’m looking for and then just go get the item from the full list as opposed to

7:49 Trying to favorite anything cuz it’s just annoying to me. It the whole experience is annoying. I hate to say it. what? I’ve been actually been really utilizing bookmarks in my browser because that’s the best way to navigate. So, for certain folders in a in a workspace or even Yeah. And like because it’s the same thing. I’m like I have it’s it’s always a pain when I know I’m going to very frequent places. So I’m doing that for the folder location which is a that’s interesting. I’m like some of my pipelines that I like to go through and refresh and for the

8:20 Reports too I’m just using a folder on my browser that has yeah I have right now the first four reports then I have workspace location then I have it just it’s been a so much easier. That’s actually really interesting. Tommy, that sounds like a feature that is missing in the product right now because of things can get so buried in folders and in workspaces. , one thing that I and I like I like fing saving my favorites for like reports and having some favorites there. Can you favorite a folder? Yeah.

8:51 Well, I’m fing in my browser. I don’t think you can favorite it in. The UI doesn’t actually give you that, right? Like, so again, I think the idea here is like I have lots of content across many fabric workspaces potentially and I want to get into a folder that’s part of a singular workspace. So it almost feels like I should be able to favorite a folder and then when I log in immediately on that homepage, here’s the things you’ve recently interacted with. Here’s here’s a list of your favored folders. Here’s a list of your favorite reports. Like that

9:22 To me, I think that makes more sense on the home screen. I find the home screen on PowerBI almost useless now. I I don’t think I use it for anything. I completely agree. I completely I never there’s nothing that I do. I never go into I never log into PowerBI and I do nothing on the home screen anymore. It’s it the the recommended cards are not useful. There’s no pictures there. , the recent items may be somewhat useful. Favorites is okay, but it it the whole front UI I need more customization

9:53 On that homepage. It just doesn’t feel right. we got to a point where there’s too much stuff. I I agree. And I think they they’ve been talking about doing the user interface for like the like consumer user interface and an experience for more people building, but I’ve not seen that yet because they have to be different. So for me as a builder, I don’t really care what Jim looked at lately. So yeah. No, the homepage doesn’t seem I don’t like the recommended area, the the the simplified or expanded space. Neither one of them

10:25 Are just like it’s just not it’s not not doing it for me. Awesome. last one here. Tommy, you got one maybe one more announcement. Just a little sentimental here for the 18th consecutive year. Microsoft has been positioned as a leader in the magic Gartner magic quadrant for analytics and business intelligence platforms. Now 18th year is a bit deceiving because PowerBI’s been around for only 10 but it has been by far and away for I

10:59 Know the last six at least it’s been the leader and the visionary like way as an outlier across all the ones and you’re looking at Tableau and Google and Qu and it’s just another time including with fabric just to stand alone. Yeah. Yeah. I think you’re right on that one, Tommy. I I do agree. I think this Gartner space is not really good. , it’s really good for Microsoft, but like there’s a little bit deceiving there cuz Microsoft has other BI

11:31 Products. It was SQL Server, SSRS, Excel, like whatever business intelligence products you had there. They’ve been a participating in this game for a long time, which is great. Analysis services, wonderful. Love it. It’s been a very mature product over time. Totally enjoy that. having Microsoft be there. Tommy, the last 10 years of this has been really PowerBI and now the last two more years of this has been fabric. So the tools that we know and love that I think really make move the needle here for organizations those are the ones that are really making a difference and I think Microsoft has clearly been

12:03 Pushing themselves themselves further and further away in this magic quadrant. , Microsoft is not allowed to do it, but you can usually find someone on the internet who’s animated every single Gartner chart by position of the dots and you can watch where the dots move over time and you can see their path. and one thing that’s been happening that I think is revealing here is in the BI space, Microsoft has taken off in the upper right hand corner and then Salesforce

12:34 And Tableau has really shrked away from innovative and and leading as much as Microsoft has. Another one that was a really big player I think was Domo for a bit and then even THS spot was also in there, but they’ve also shrunk away compared to the competition. Dude, I I cannot tell you how many of these I tried to play with before in the past. Like when we were looking at different tools, there was one that was I I don’t think it’s even up there anymore, but the days IBM had a shot. But I think

13:05 That the biggest thing is they’ve been a leader in this and a visionary and and the white paper is great. You can download it for free. I would if you’ve never read one, it really really intense. very good reading for going to bed at night when you’re really tired, just can’t sleep and Yeah, exactly. But I I think one of the biggest things is they’ve been the leader in visionary since PowerBI’s days, , it’s not just fabric. Yeah. Which it makes me feel confident that like putting putting us our knowledge, our experience into helping people learn PowerBI is a good

13:37 Is a good thing because Microsoft has continued building on it. They are considered the leader in the space. That’s the the product you want to attach yourself to. I feel like I looked into this one, Tommy. I saw PowerBI. I thought this is going to change the world. This is this is going to change how people do things. And I think I’m I’m glad I leaned into that cuz I was like, “Oh man, , a couple years into this really doubling down on PowerBI, Microsoft could have gone a different direction and I could have been wasting my time on a product that they weren’t going to continue to invest in.” So for 10 years now, they’ve been heavily investing in this product and it’s made a career out of it for me

14:08 For sure. Thousand%. Awesome. All right. Well, I have one quick little note here. , I just want to talk a little beat from the street here. So, as as one who creates a lot of content from YouTube’s long form video. I have a really funny example, Tommy. Okay. , I’ve been I went on another podcast the other day and on the other podcast I was with Armando. , he’s over he runs the the web podcast. , he talks about web web technologies, building your your website for optimiz optimizing for marketing and ads. And

14:39 One of the things that we talked about was we we use a lot of like AI generated features from other tools to help us like make shorts and clips and other things from our long- form videos. And I was joking with him and just talking about well when you use these you you upload your long form video and then these AI tools one of them being Opus clip one that we use helps you chop that up and find like viral moments or things that were said in the in the conversation that were viral moments. And Tommy, I gotta be honest with you. I’m not sure if I’m

15:11 Making more viral moments or the AI is training me to talk in more viral moments. So, think let me give you let me give you what I’m thinking here. When these AIs read or listen to the podcast, they take his transcript and then they score these little clips or these hooks in the video and say this hook, this segment that you said has this, , verality to the the comments like this could go viral, right? And it gives you a score 0 to 100. How how much the AI thinks this is

15:43 A hook, a good hook that people will want to listen to. Tommy, we’ve been doing just the podcast forever and we haven’t we’ve only incorporated the tools here recently. And so as as I was talking with Armando, he’s like, “Well, , in in past podcast episodes he’s had and he uses a different program called Riverside to broadcast.” He said, “I’ve only had like, , four or five video, , viral clips made out of an hour or hour and a half podcast. That’s it. Four or five.” He goes, “The podcast we did with you, the machine without extra

16:14 Prompting made like 20 clips 20 clips of like viral moments.” Oh my gosh. And so my my joke here to him was like, I don’t know cuz I’ve been using Opus Clip for a bit. I’ve been processing the videos, pushing them out on the internet. So I’m I’m looking at our conversations and I’m learning from the AI what makes a viral moment. Like what does that do? So the thing here is I was joking with Armando. said, “Maybe you should gauge the quality of your speakers by how many viral moments were made.” Right? Those

16:45 Are insightful pieces of information that were condensed down to a minute or less. Maybe that’s what you should be. That’s what it’s doing. It’s finding sure a thought that can be used in a minute or less. So I the joke here was going and why this is funny to me is I was teasing with him and and talking about this conversation and I thought, well, wait a minute. I’m trying to write, do the podcast, but yet I’m editing the video using AI, and the AI is telling me what moments are viral or not, which then is probably informing me

17:17 To actually make more viral moments because I’m using the AI to show me the viral video clips, but I’m editing the clips, which is training me, the user, to actually be more viral in what I speak about. It was like, I think I just had an inception. Like, it hurts my brain. Like if you if I didn’t use the AI to tell me what conversations were going to go viral or what video clips were going to go viral, then I wouldn’t change how I talk and I wouldn’t adjust what content I make or how I make it. So I could be in I could be

17:49 Inadvertently right now being trained by the AI to change my behavior on the podcast to say things in a way that would go more viral and then when the AI runs the video, it picks up on those moments like killing me, man. I I when I got to this moment, I was like, I think my head’s hurting. This is crazy. Like, the AI is training me, but I’m giving it content, so it trains on my stuff. Like, it’s this circular loop of like, wow, this could be like a reinforcing , , what do they call it? A control loop, right? We could be making a

18:20 Control loop between the AI and me and what I say and how the AI works and what’s viral, what’s not viral. Like, oh my word, this is crazy. Anyways, only reason I bring that up is because there were some papers that came out recently from Microsoft talking about how AI is reducing the cognitive load on your brain when you do things. And so it’s actually making you less you’re not able to retain knowledge as much. There are some pretty large downsides to using AI. If if you don’t do all the technical details like

18:53 Write a paper or something like that, you’re not gaining the skill of critically thinking in a way that lets you write papers. So, if you use AI to help you do things, it’s taking away it’s like a use it or lose it thing. That part of your brain muscle is not being worked. because the AI is doing it and you don’t have to worry about anymore. So, like there’s just this is interesting, Tommy. We’re stepping into a very new world here. Anyways, wow. Yeah, let’s move on. All right. Yeah, you’re you’re you’re racking my brain, dude. I After that head scratcher, let’s get into our main talk about PowerBI and fabric. All

19:24 Right, let’s get into our main topic today. Ah, do you hear that, Tommy? I do hear that. All right, we have another caller today. I would love to bring back to the podcast. Matias, welcome back to the podcast. , speaking our AI DevOps genius is now here. and Matias is going to program us and how we think about DevOps a little bit more through this conversation. Welcome Matias. Yeah. thanks for having me. Super excited. We had a great chat a couple of days ago. So I’m I’m looking

19:55 Forward to what’s going to happen now. So we we talked yesterday I guess it was on Tuesday. We talked a little bit about what DevOps is and and what it is not. Actually more of the conversation of what it isn’t so we can rule out some things. It’s not a piece of software you go buy. But at the end of our conversation, we start talking about, well, where’s the proven value? Where do where do things land and what conversations should our organization have around DevOps, the methodology, right? And then what does

20:26 That look like and maybe how do we roll this out and implement it for our teams? Tommy, is there any any other topics or questions we should maybe frame this conversation out with and start with? No, I think the biggest thing is that selling point, right, where you don’t want to try this in a vacuum individually, right? And start making all your workspaces get and that’s your process at an organization that’s never done that before. especially that you’re going to be siloed. So for me, I want to talk about today that that initial step of really bringing that to your team or

20:58 Your organization and getting everyone’s feet, , their first toe in where they’re not overwhelmed, but they actually see the value as well. I think it’s a great starting point. So Matias, give us some conversation here. So there’s like two approaches here. We’re talking a little bit like top down leadership level understands it and my team doesn’t do it. And then we have another approach here from like the bottom up approach where the team says, “Look, we’re having struggles getting content out the door. We need a better way.” And I think a lot of conversations I I’m involved in. I

21:31 Think I’m in a lot more involved in the bottom up approach. Honestly, I think I’m having a lot more conversation with like the technical teams that are building the content and they’re asking me as a consultant or talking to me about like, “Hey, look, should we just turn on Git on the workspace? Can we just do that?” like but our team’s not comfortable with it yet. How do we how do we get our team comfortable with this? We think this is Microsoft is touting this as CI/CD. We need to be using these artifacts now that dataf flows gen 2 is supported. We think we should be using it. How do we get started? So let me kick it over to you

22:03 Matias. What maybe what conversations are you having more frequently here and are there certain talking points we should be thinking about when we talk to bottom up or top down approaches? Yeah, good questions and huge questions, right? , where do we start? I think , a lot of what we discussed Tuesday was around culture, right? DevOps is not culture. Yeah. It’s not software. It’s it’s not technologies. It’s not a product you buy. It’s it’s all about , , organizational , setup and

22:38 Patterns, right? and Tommy just said it. DevOps is the opposite of doing things in isolation, right? DevOps is the opposite of of of of of small tiny teams fighting each other. , DevOps is all about communication, collaboration. so in a way you started off with is it how do we bring it down from the top as opposed to

23:11 How do we bring it up from the bottom in a way the former is much easier right in in an ideal world you would actually have the top of your organization push the devops approach all the way down right because ultimately that’s what’s needed I appreciate this is all very theoretical and and and abstract. So let me let me give you a few specific things here. Stuff that in fact I’ve implemented myself in the past.

23:44 I also mentioned cross functional teams when we talked last time. I think that’s really really key. It’s particularly we’re in a data space which is ultimately all about delivering insights and value to more or less the top of a company. you cannot produce those kinds of insights if you’ve got some engineering team tucked away in the basement sitting there detached

24:15 From the rest of a business. , so one thing I would really want to stress here is think about how can you have how can how can you embed your engineers into the department where they’re supposed to deliver insight. That would that would be one thing, right? So, and really get away from from traditional notion of having it and ops and sales and and and or finance, , all being

24:46 Their separate slices. so one thing which which I’ve done previously was to basically think of data of of dashboards and and reports etc as products which we deliver internally and then and then put on the product hat as if you were as if

25:17 You created, , a a a public product to be sold. , which means a customer focus is going to is got to come first, right? You got to have a product manager. You got to have someone who is as closely embedded with the people ultimately using your stuff as possible. bring a trainer in. invest in documentation. Not documentation, , that’s very techheavy and and and niche, but

25:50 Documentation that actually makes sense to your users. so th those would be some some specific ideas. but ultimately it’s about embracing things like cultural change breaking down barriers. Back to you. I want to I want to react a little bit to your embedding the engineers directly into your business teams. I I had a an an

26:22 Experience in one team when I was working back in my corporate days when a large Fortune 100 company and we had a like a like a data it was like a the IT team had like a data team. There was people that were touching the databases at the time we were using business objects. So there’s a business object developer building universes and creating the data. And so one of the things that was very unique in that situation the salary of the IT individual was paid for by the business. That person sat it

26:55 Was they they had a their their report line was to an IT individual but the salary was paid for by that business unit because they had so many data needs. So that what that meant was is that yes that person had to adhere to all the IT standards of things or what it was doing. They had to make sure that like they were in doing things in a secure way not following IT practices and principles but that user sat inside the business. So this was a commitment that was made

27:28 By the business said look we had data problems. We need help getting data things done. The IT organization is doing their own thing. they they’re getting things done fine but there’s not a resource that’s like directly allocated our projects were coming in they were getting shifted around to multiple people you had other initiatives like we’re redoing the the the ERP system or so they had their own priorities right so they’re like we don’t have budget to get your business stuff their business data done and so the business kept doing a lot of work in like Excel they’re like we need someone to be like a voice an advocate again to

28:01 Your point embedding getting your engineer like we did that. We physically had that person sit on our team in our row and then we worked with them directly to build all the tables and and it made a lot it made a huge improvement like all the needs of the business could be funneled down to this one person. We could have better timelines. They were dedicated on the work that we need to get done. and then that way we weren’t always struggling with the okay we’ll add it to the backlog. Where does that item sit in the large list of items? Because everyone’s need like when it’s just your team or your department your need is always number one priority. like I need

28:33 It now to get it done so I can do data things. Well, if it is then pulling that resource away to do other stuff, you lose that resource and you’re forced to find workarounds to to work with your data. So, I I personally experienced this and felt honestly this was a really useful and impactful way of leveraging a let’s call it the data engineering team and sitting them closer to where the business intelligence was being developed. Mhm. Tommy, any reactions to embedding your engineers? Especially

29:05 When you talk about the team structure, team organization, because I I’m really intrigued about the bottom up approach when we’re going to talk about this because it’s happened to myself where everything was great, everything was dandy, we had the right connections and then someone came and said, “We’re going to reorg, we’re going to remove everyone.” And it’s amazing how much that broke process and also the quality and especially in this conversation when you’re dealing with something. So again, I will say the assumption is so dev

29:37 Tech heavy that’s probably not going to come from the top down to say all right you guys are going to start using git because it’s here’s the thing though as much as I love it and I’m using it. not a global best practice right now for BI teams, right? Like it’s not like a prerequisite to do source control or to do DevOps at all. I or some organizations do it, but it’s not like you go into an an enterprise software

30:08 Team and you’ve never heard of Git and you get laughed out of there. That that’s not going to happen in our space. So we have this service right now or this process methodology that is not a global stamp in our industry yet we’re all seeing the value. So it’s probably going to happen more bottom up. So I want to hear about that grassroots appeal or the work there because that like I said this is not you I I imagine you ask a hundred people

30:39 Doing fabric or powerbi right now. How much of them have heard of GitHub? how much of them have heard of DevOps or have done it. So that that’s where I’m I’m going from. So it really has to come from the people who are passionate about it. And and Tommy, I can’t I’m not going to try to like debate you on your history here because you just you you’ve learned what you’ve learned. You’ve gone through these experiences, you’ve said these things, but , as I think about what DevOps is doing and like the culture of that process, right? Again, Matis, we were talking about that a bit yesterday. I would really hope that

31:10 Shifting my people around or moving people in my organization should not impact my process. And I would argue I would argue if you if the act of moving people around in your organization breaks the methods by which you’re able to produce content, you don’t really have DevOps in place. It was relying too heavily on the people and there wasn’t actually a robust process under the hood. I I would think and this is part of the reason why I think we’re talking

31:41 DevOps here to some degree is again Matias is maybe thinking he’s like I think he’s going to disagree with me here but I I would hope that the DevOps process will would would be able to stand a shift in individuals underneath of it. And so we have documented patterns we have this is how we do things. This is what the tech this is this is how we integrate the technology into what we do in order to produce content. and whether rework happens or not that process I would think to remain relatively the same in order to develop quality products. Maybe

32:15 Your reaction there Matias then we can go into like the grassroots side of things. H well oh you you you introducing or implementing DevOps is not something you’re you’re going to get for free and over or overnight, right? so it so okay coming coming back to the

32:47 Question how do you go about it and how do I convince people I I think there are two completely separate buckets to look at here one is my fellow engineers don’t believe in git or my fellow engineers have nothing automated they’re happy, , pushing buttons and doing everything in a manual way, right? So, that’s one challenge, but it’s very different from trying from from actually believing into

33:20 The the the bigger picture of DevOps which is all about agile delivery and continuous improvement and and wanting to bring this into your organization bottom up, right? So we we need to I think both of them are really important but we need to talk about them differently. Which which one do do you want to tackle first? Shall we shall we do the former the the tech internal? How do I

33:51 Convince my report developers that g is a good thing? Let’s start there. I think that’s Tommy. I think Tommy wants to start with like the grassroots perspective like how do we get the developers on and I think one story that I have found that works really well here is did you need a backup? Have you ever I start presenting problems. Have you ever have you ever done and then you enter the Have you ever deployed a report where it was on the wrong page? You added extra filters in the filter pane and you weren’t supposed to and you had to like quickly revert the report or edit something. Have you ever done that

34:23 Before? and and , have you ever had someone overwrite your changes after you made a day worth of work? Have you ever lost work because you’re using SharePoint as a versioning tool and you can’t merge the changes back together? I think that’s usually some of those conversations I start everyone’s like, “Oh yeah, we I see nodding of heads. People have done it. It has happened.” And I think that’s one of the pieces that I use is like, “Okay, there’s a better way to develop than using this.” , again, this is a culture thing here. We’re trying to get people people have been very well trained to use

34:55 SharePoint as their backup site for PBX files. I think there’s a better way now and I think the Git integration is something we should embrace and train our teams on. Sorry, Matias, what do you think? , one thing we we do have to acknowledge sadly is that there there’s there’s a price here in terms of upskilling, , people changing habits, people having to be convinced to work in different ways, right? So, , that’s always going to be problematic. , and I’m just going to be very open about it. , so you’ve got

35:27 Two ways of going about that, right? You can move to another organization that’s already con already converted or you can do the leg work to convince people, right? and well, maybe a way to achieve that is to let them fail, right? to to let them experience themselves all the drawbacks of not having source control, not having automated processes, not having

35:59 Automated testing in place, right? let people get to the point themselves where they think there’s got to be a better way here, right? rather than trying to force everyone, , either you learn Git or you’re out of the door, right? that’s not going to work because ultimately you it only works if if everyone’s fully on it. So but I think you want to pay an an alternative picture in in

36:30 Which lots of today’s problems don’t exist anymore and and you mentioned things like having chaos around different versions not knowing what’s the latest in production having to stay late on a Friday because , you you you pushed a new report and suddenly everything broke. , there there should be lots of in in fact,

37:02 You remember I mentioned the Phoenix Project, the book. yeah, a couple days ago. if you need examples for that, read the book. This is what it’s all about, right? It’s pretty much showing you in in in in a very nice and readable way, , how businesses fail because they don’t actually embrace those practices. and yeah, so that that would be one suggestion. And I want to just tag on there, Tommy, I want your opinion on

37:33 This one, too. I I think you have to be a savvy leader to be able to do this, right? So, I I think when you say let them fail, I’m I’m I’m on board. I’m okay with that. Like, I’m like there’s going to be times where that someone’s late on a Friday. We need to fix a semantic model. like these issues are going to happen. But I think as the leader in this situation, I need to be mindful of that. And when these occur, have a little birdie in the ear that says, what, there might be a better way for this. You

38:04 Know, I think if we looked at and so I feel like there’s a little bit of like a, , you need to be mindful that the failures will happen when the failures do occur and you as a leader step in and say, “Okay, Tommy, I need you to work a little bit later. The semantic model’s down. The sales team can’t get the numbers. Can you help me out here?” And while Tommy you’re doing the extra effort and you’re you’re experiencing the pain that we’re eventually going to try and reduce as the leader you also after this a retro a moment here take the opportunity to voice some what there’s

38:35 Probably a better way we should be doing this right we should probably be thinking about git and then and then the engineers like yeah that was not fun I did not I did not enjoy that git would probably saw this or let’s add some some process to what we’re doing here let’s and then you could start weaving in those conversations around DevOps and start seeding the idea even before you’re really ready to make the time commitment the tech debt right to to really just implement and go with it. Tommy, your reaction? I like that, but this one’s a

39:07 You make a good point. The only thing I I would feel like there’s a lot to introduce here. It’s like, oh, what could have worked better here? A whole infrastructure solution in from soup to nuts. like what part I feel like I want to introduce some of the parts in in pieces, right? Rather than the the whole package of like we’re going to do source control, there’s automated testing. Oh, it’s great. And no, no, there’s this idea with the you work on a development and then you have to publish it, but then I can check the changes. There’s a lot to unpack there. And I feel like I would want to like take some

39:40 Of those pieces there and introduce them slowly things. It’s like, “Oh, you also like sesame chicken.” And like rather than saying, “We’re going to go to the entire Chinese restaurant.” Like, “Well, we have rice. Rice is good.” Well, what would you start with, Tommy? Like if if if this is too big to eat all at once. Yeah. What’s the easiest thing that in your mind where would like if this the issue happens, problems arise, something? Yeah. What do you start with? I would go with the one that probably is the most frequent and it’s what we talked about on Tuesday where it’s a change that someone made. I have to go

40:13 Back and open up PowerBI desktop and what did you do and what did I do and you was like hey I know this is going to sound like it’s a developer thing it’s not however what if I told you that any change that any of your the team made you can revert to the line thing to the DAX measure without even opening up PowerBI desktop and you can revert each push thing and then just introduce that way and so you’re you’re basically talking the first thing you’re going to do is going introduc I to me it’s the natural place to start.

40:47 But I’m I’m happy to hear if there’s maybe a better introduction too. Well, I would be even more specific, right? so let’s say we’re in a PowerBI world. that first step would be to move away from PBX, right? I enable PBIP, enable PBI projects. and introduce everyone to this is effectively the the same

41:19 The same output in terms of what you work on in desktop I like that but it looks different and most importantly it allows you to track your changes once you’ve done that it’s easy to go the next step and think about a bit of CI/CD maybe right that’s the stuff you can and automate in terms of deployment. , once you’ve once you’ve got that in place, you can think about having different environments. Once you’ve got that in place, you can think about doing

41:51 Parallel work by having feature branches. , and and then you can really convince your business because you’re going to be able to scale in terms of ORC and development, right? so the story writes itself almost and you can start very easy very simple with tools that are there today going to PBIP is very easy it’s just safe as in in PBI desktop but it opens

42:25 A very long road of potential improvements one other note I feel like I’ve been fighting here recently is the resistance of the business to move away from SharePoint. I I don’t know if you’ve seen this yet, but Microsoft has done a I feel like Microsoft has done a really good job of like put all your files in SharePoint. That’s how we version that’s how we bring in that content and keep snippets or versions of your reports. And I feel right now the conversation even with leader level,

42:57 Right? Oh man, we need to move away from SharePoint. We we got to embrace the PBIP format. They know PBIP is the future, the way to go. But getting them to rethink, well, can I store all the PBIP files back in SharePoint? That’s the question, right? And you’re like, well, that solution is not really made for to your point MS the direction we want to go like different environments, parallel work with branching like this. The the strategy should be move us more towards that direction and we need to step a little

43:29 Bit more away from the SharePoint. the number of times a conversation that’s what we do that’s how we store things yes they understand but we need to bring the team knowledge up a bit so that they understand the PBIP format and I think that’s that was really a good point like that’s so fundamental like once we understand what PBIP is and the fact that the model and the report are a collection of files and they’re not one big zip file that that’s an aha moment for people right there that to me That’s

44:01 A that’s the that’s the fundamental without that that’s the keystone for the whole process. If you can’t grasp that part, you might as well just ignore everything else. It’s not going to be worth it, right? , I think like that’s and I think Microsoft is aggressively moving towards the PBIP format in everything. It’s going to be everywhere shortly. And the there’s a fantastic analogy here to to to the rest of fabric, right? Because in a way obviously fabric meaning everything that’s not

44:33 PowerBI traditionally you do everything in a browser right yes it’s it’s just UI and and you and and you click buttons so in a way the native fabric experience is like the the PowerBI PBX experience in the sense that it’s a black box you don’t really see what’s happening. when you do action A or action B, you you can’t really tell what’s changed between version one and version two. So the

45:05 Move from PBX to PBIP in PowerBI is in fabric turning on Git integration and suddenly seeing all these JSON files that are behind your items and being able to see, , you’ve changed your data flow. you made a commit and pushed it into Azure DevOps or or GitHub and suddenly you see there are a few lines of JSON that have actually changed right so that that would be I think a perfect equivalent here and again pretty easy first step I like

45:38 That analogy actually I wrote this down my suggestion here was and as you’re thinking about this Matias I was unpacking it a bit as well the first step should be let’s replace SharePoint with workspace synchronization in Git, you don’t need to see all the files. There’s a nice pretty like I I really like that experience the synchronization and I’ve so let me let me caveat this with some things as well. I’ve had some people tell me the git the get integration doesn’t always work 100% right all the time and they get into

46:09 Some weird scenarios every so often. So I’ve I’ve seen situations where it gets a bit off. However, what I would say though is my experience has been here just turning on get integration to the workspace and then seeing items that are committed or not committed in the workspace is invaluable. that that to me so by doing this I feel like I’m now comfortable letting teams build reports in the service if they want right to me like I not not maybe wouldn’t love it

46:42 But that’s one of the things I’ I’ve had a huge I’ve been voicing this since when PowerBI came out never never ever let people build reports in the service because you can’t you can’t get it out you can’t download it there’s no way to to get the the the information like if you screw it up you have to rebuild like that was that was the mentality with building things in the web. Well, now with the git integration, you’re at least getting the parts of that report and everything’s being stored. So now, yes, the workspace is having some changing of artifacts. They’re, , adopting or mutating

47:13 Over time as you build them and create them, but at least there’s a a timestamp where I can commit the changes using the UI. I’m not writing code. I’m just writing a commit message. Here’s the message. Here’s the item I changed. Save. And so I think starting with that mentality in replacing the SharePoint experience is an easy win. Like that that should be a no-brainer. We should be able to teach that. That’s a good first step. And then once you’ve gone there, you can go f further down

47:44 The environments, branching, all these other more elaborate things. , but that’s very fundamental to even getting things started. I love that. Very, very good. So Tommy, you were gonna say something. No, I I I wanted to dive in and I I cannot believe how quick time has gone. This has been great. So, let’s say we we have a team set up and get we have our source control fabric set up and things are going smoothly, right? But as we talked about Tuesday, that’s a a misconception that DevOps is just source control. So, when does that next

48:16 Step take place? Because obviously we could we could either dive into those pitfalls that would happen when people use source control and how to get people to really gung-ho about git. But at some point right there’s also that like all right we’ve been doing development prod or in my me and Mike’s case we just push the main all the time thing. But to start introducing I only work in production Tommy work in Maine.

48:48 I only work I only develop on main. I am the branch that I am the main branch. Oh, I love it. Oh, it’s hilarious. But at some point, right, there’s going to be this conversation like, all right, we’re going to start talking about really more than just source control because I think if you just stop there, it is fine, but you have added more steps to the process. there’s a better quality but you have not made it faster or at that point

49:21 I would I would introduce something else first actually and earlier when we were talking about educating people horizontally versus vertically within the business as in how do you get your peers on board as opposed to how do you get your receivers you your your end users on board. if we just briefly focus on on the latter for a bit. from my point

49:52 Of view, the key here is to introduce an issue tracker. something like Azure boards or or GitHub actions to basically say to the people who want the dashboards and the semantic models and the reports from you you may have sent me teams messages or or emails previously if you wanted something why don’t we formalize a little bit why don’t we actually put

50:23 That in in a piece of software where we actually track each request and where we then have a regular meeting with you where we jointly prioritize what we’re going to be working on. and where you have full visibility about what’s in our backlog. You have full visibility about how long it takes us to deliver that. and we have complete transparency around what what you value in terms of your priorities. ,

50:55 Once you’ve done that, and I’m coming back to your point, Tommy, now then you can very neatly tie that into the git world because if you then think about regular releases and and planned releases and features that contribute to those releases, that actually maps to your branching model. That that’s how I would do it. Then I would say, , well, we’ve got our main branch here. By definition, that’s in production. Anything that’s that’s committed to

51:26 Main is in production. So whilst we work on new stuff, , logically we have to introduce a new branch. So why don’t we then create a new branch every time we start a new sprint and why don’t we let every developer own their own branch whilst they work on their specific slice and then integrate all of those branches into that release branch and then ultimately when when everything has been

51:57 Signed off that in that release branch is moved into main right so I think this is this is how you get first of all full circle you user request all the way to release into production. But this is also at the same time how you introduce a branching model to the developers because you’re not introducing branching for its own sake. You’re introducing branching because it maps directly to what the business is actually asking for. I like that very much. I and this is one area that I think is a little bit clunky right now

52:30 In the fabric experience. So I agree with you Matias. I think I I think the tracking issues at least getting a list of things that you need to get done, right? Stuff breaks and that way you can I think a lot of times we we solve problems but we don’t actually know the history of like where the problem came from, what happened, what was the root cause, what what was the what was the issue, right? And so by having an issue tracker and being able to write down these items, we have a better feel, I guess, for some of these things. One of the things that I find just is clunky today and I think this is

53:02 What Batiz you’re probably trying to solve in your world as well with with Navad is it’s not always easy to get a branch from a main workspace created and and my issue like so I feel like Microsoft has done a good job of like creating the items the physical items like oh I’m going to take a branch from we have our sales workspace and then that’s on the main branch and I’m trying to create something from that a branch that’s different where I want to modify something to report or change the semantic model. The items themselves

53:34 Seem to go along very well in Microsoft’s processes. But what doesn’t what doesn’t do very well is the data part. So, I feel like what I what I’m doing right now in my in in the ways that m the the way Microsoft has implemented these DevOps like processes is I have to take a copy of a workspace, make a whole brand new workspace for my branch, my testing, my my changes, and then I have to run the compute again to go get the data so I can do the

54:05 Operations cuz I’m going to bring all the content down, but I’m actually going to I’m not actually going to bring the the data with it, right? And I feel like there’s a really missed opportunity here if I again, this is I’m going I’m going too deep here, so just bear with me. I really think there’s a better way to do this where I should be able to think about my workspace like an encyclopedia. And and here’s what I’m going to go through analogy here. Every single item that I build in my in my workspace might have, , there’s

54:36 Some items that depend on the lakehouse. There’s data flows that are pushing data around in the lakehouse, but the data is already there. What if my changes on the report or the semantic model side, I don’t need the I if I’m going to go look up the word any word, , zebra, I don’t know, I’m just picking a letter. I can grab the Z’s and go look through the encyclopedia and for all of our younger audience, this is a these are books and you would go get them from a shelf. , we don’t I don’t there’s not an encyclopedia in my house anymore, but like I go pick the thing that I want to

55:07 Go change. I want to go look up. I know what I’m looking for. I grab generally the area that I’m going to go change, pull that item. So, I want to pull just the semantic model or just the report and I want to remotely while I’m doing my branch, I want to remotely connect back to the original semantic model. So, like the semantic model needs to be a parameter that I can point back to main. I’m not going to change the semantic model. All I’m doing is modifying the report or even even more differently right it’s a it’s a

55:38 Lakehouse right I have direct lake maybe enabled right I want to pull the semantic model down still being pointed to the dev branch of the data but then I can just work only on my little artifact separately make my changes update my measures da da da da and then push that back into and then when I push it back in this is where the automation makes sense it automatically recognizes oh you were working in a in a branch, you’re actually trying to connect to this lakehouse that’s supposed to be in the main branch. Oh, we’ll just switch out

56:09 That connection for you and then put the book back on the shelf. We replace the connection and then now your change has been made to the semantic model even though I’m I’m I actually physically have to change the connection between workspaces potentially. There there’s a whole data switching story here that I don’t feel is very well developed from the Microsoft side. And I think this is where Matias, you’re putting a lot of effort right now and where the community at large is going to be building tooling around to making this easier. Does that make sense? Am I do you see

56:40 The same thing or did I just totally derail this conversation and go on a random tangent here? Well, so what you introduced actually is a new concept, the concept of environment which is different which is different from the concept of branch, right? Yes. , so branches are cheap, right? You can have a million branches and and and create and destroy them and it cost nothing. Environment in a data world, that’s where the challenge is because you need data, right? And so, correct.

57:12 , , in in a in a in an evolved DevOps setup, , you’re going to have environments , the most famous one being production, right? the sacred environment. but then you’re going to have pre-production environments. Some of them are shared. Some of them are individual. Some of them are long lived. Some of them are ephemeral as we say in the DevOps speech. Yep. and so this this is this this is ultimately where

57:46 Smart tuning comes in and this as you rightly say this is precisely what we at Navigator focus on massively because there is there aren’t any good solutions in place today. agreed and yeah so working with branches and working in a in in a in a parallel and collaborative way introduces a new problem for data pro projects.

58:17 You can solve that problem with a lot of manual effort which is what you are forced to do if you just work with with out of the box Microsoft product solutions today or you use and and then that’s obviously a massive turnoff right anything that’s manual is completely against defob principles or you find the right tooling to help you automate that setup as much as

58:49 Possible. so those are those are the choices you’ve got today. But ultimately you need to have a solution for for that problem. you can’t there’s no way to eliminate it. you have to address it. Good way to think about that. Yeah. Tommy, let’s let’s maybe wrap up things here, but we we’ve almost pushed everyone to about an hour now. This is we wrap up the time like that time of the podcast. Thank you all very much for listening. Matias, this has been another great wealth of thinking and unpacking and like talking about these challenges. We have some

59:21 People over here on LinkedIn which we can’t see in the chat. People are are commenting on the ability like the SharePoint safety blanket, right? That that’s like a comment that’s there that that’s what we’re used to using and we maybe have to move away from our safety blanket into other places. Tommy, what would be maybe your final thoughts here? What your wrap up here as you think about this conversation today? I think you have to think in the long term especially if you’re in an organization that has not been introduced to this or just does not perceive BI being part of u u devops but I think there’s the a lot of sprinkles

59:54 And I think the biggest one is just the fabric of git enablement and just starting there what hap what’s going to happen naturally from that is one of someone has to take up the reigns of understanding what devops is or is that’s going to fail miserably and so you’re listening here and it’s going to be you and understanding some of the best practices and the methodologies and what you’re actually trying to achieve that way. But just getting everyone getting comfortable with that initial step, there’s that bit of that , , learning hump, but more just

60:25 Like I don’t know what these commits are. Once people are comfortable there, then the other concepts are going to come a lot more naturally. Matias, any final thoughts as you wrap the the conversation up here for you today? H well since since you mentioned git commits I I think moving forward there’s huge potential in terms of AI to help us creating good comm commits and particularly good commit messages and summarizing what we’ve done right this is this is that in

60:56 In in the git world today this is one of the big issues that every developer every engineer is struggling with to some degree and it’s easily something we can at least to some degree delegate to an AI. So, , and some of the tools I’m following are doing some great work in that space. So, yeah, this is definitely hope and a really good use of AI as well. I think my final

61:27 Thought here really centers around this idea of the the SharePoint safety blanket, right? There there’s there’s I really liked your your idea or your concepts, Matias. You’re talking about when there are pain points as as a from the top down side, leaders need to be mindful of tracking when issues occur, putting them in a place where we can discuss and think about better ways of doing development. really love the idea of the data as a product, right? We’re building products, supporting products for organization. If we treat it as such, you’re going to

61:58 Make your product better. You’re going to figure out how to deliver it faster. These are the things. These are values and core values we want to instill in our team. So for the top down approach, I love that seeing when pain points occur, working with the team to instill or or build out these new DevOps like processes, I think is definitely the right way to go. And then also from the bottom up approach, I I I really do subscribe to the hey, we may not have all the details of environments branching figured out at this point, but at a minimum, it’s not a lot of extra weight or work to just start with

62:29 Move away from SharePoint. Let’s turn on get integration in our fabric workspaces or premium. It’s actually just not just fabric. It’s it’s premium per user. Anything premium related, premium per user if you’re paying per user or any of the fabric or pKus as well. just turn on get just see how that works for your team and for some of your more missionritical workspaces I think this is a no-brainer this gives you the ability to roll back users can see when things are out of date and the the definition of them have changed based on a standard Git integration and the UI is simple it’s straightforward this thing’s

63:02 Different checking your change no problem but I think just that alone adds a lot of value to the team especially we’re talking these simple things of your point Tommy a lot of People overwrite other people’s changes accidentally and when you have more people working on the same stuff, you’re going to come into this place where you’re going to have conflicts. Last one in wins only works in smaller teams. You need to think about this a little bit differently. So with that being said, wonderful discussion today. Lots of talk around top up, bottom down, top down, bottom up. there we go. I got it

63:34 Right there. and and just some hopefully some good thoughts for your organization as you think about DevOps, as you think about building things and a a more robust system, a better data product. Thank you for listening today. We really appreciate your ears. We know you could be doing a lot of other things, , washing your hair, actually getting work done for once, going for a run, being active or something, anything else you could, but you you choose to spend the time with us and we are very grateful for your time. Thank you very much, Matias. Also, we appreciate your time as well. Thank you for spending a good old hour out of your day talking with us more about DevOps.

64:06 More topics to come around this space. I think there’s a lot more to unpack here. We start with a really good intention of talking through a lot of things in these in these podcasts and we like the conversation just drifts and we get a lot of really good thoughts but not necessarily hit all of the topics. So, we’ll keep unpacking here as we go and teasing this out. Please be able to share this podcast if you like this one. We really appreciate you sharing the podcast as well. Tommy, where else can you find the podcast? You can find us on Apple, Spotify, wherever you get your podcast, make sure to subscribe and leave a rating. It helps us out a ton.

64:38 You have a question, idea, or a topic that you want us to talk about in a future episode. Head over to powerbi.tips podcast and leave your name and a great question. And finally, join us live every Tuesday and Thursday, 7:30 a.m. Central, and join the conversation on all of PowerBI tips social media channels. Thank you all so much, and we’ll see you next time. Thank you.

Thank You

Want to catch us live? Join every Tuesday and Thursday at 7:30 AM Central on YouTube and LinkedIn.

Got a question? Head to powerbi.tips/empodcast and submit your topic ideas.

Listen on Spotify, Apple Podcasts, or wherever you get your podcasts.