FABCON 2026 Recap Part 1 - Ep.517 - Power BI tips

In Episode 517 of Explicit Measures, Mike Carlo and Tommy Puglia unpack the latest Power BI and Microsoft Fabric topics from the show. You’ll get a quick read on the episode’s biggest ideas, why they matter, and where to dig deeper in the full conversation.

News & Announcements

• No linked announcements were available in the episode description for this post.

Main Discussion

This episode covers the major themes, opinions, and practical lessons Mike and Tommy surfaced during the conversation. The transcript below captures the full verbatim discussion if you want the exact phrasing and context.

• Mike and Tommy react to the episode’s biggest Power BI and Fabric developments and explain what stood out to them.
• They connect product announcements to day-to-day practitioner decisions instead of treating the news as abstract roadmap chatter.
• The conversation highlights where teams can move quickly, where they should slow down, and what tradeoffs deserve attention.
• They share candid perspective from real project work, which gives the discussion more practical value than a headline recap alone.
• The episode mixes tactical advice, opinionated takes, and a few forward-looking predictions about what listeners should watch next.

Looking Forward

If this episode’s topics affect your current Power BI or Fabric plans, use the transcript and linked resources to identify one concrete change you can test with your team this week.

Episode Transcript

0:23 Good morning and welcome back to the explicit measures podcast with Tommy and Michael. Everyone and welcome back to the show. We are now officially live again, no more recorded episodes, the travel travel I guess chaos that has ensued for the last month is now officially over. Hello, Tommy. Do we know how to do this still? still? Right? Are we going to remember how to make this work?, , we’ll we’ll we’ll hope so. What is We’re going to wing it like we’ve been doing for the last 5 years anyways, we’ll probably just wing it. Perfect. Perfect. [snorts] [snorts] All right,, let’s jump into our main topic today. So, the main topic today is

0:53 topic today. So, the main topic today is really going to be all about our reactions to things that came out at Fab Con. Fab Con has a lot of new announcements. There’s a lot of blog posts that came out about it., there is a huge,, all the major updates that came out by major feature summaries., so anyways, it’s just really rich stuff is getting released. I’ll just give you the general overall feeling of the conference. So, two two maybe two observations. First observation, conference was big. We had 8, 000 people

1:23 conference was big. We had 8, 000 people in a single conference. There was a lot of people moving around. I felt like we hit a level of like feverish pace that I have not had in a conference before. Tommy, we we remember the conferences when they were like 2, 000 people, Data Insight Summit when they were really small. 8, 000 people felt like everyone was trying to be everywhere. I was always late to something. I could never get to like the sessions I wanted to quick enough. It it There was just so much going on. It was just difficult to like keep up with everything that was happening, which I get it. It’s 8, 000 people. It’s a lot

1:54 I get it. It’s 8, 000 people. It’s a lot of sessions. It’s a lot going on. And they did a great job on food. Food was easy. You just sometimes in those bigger conferences you have issues getting food and getting food quickly wasn’t an issue. So, the con- The conference, the venue was nice. It was good. I think Tommy, you were there a while ago., did you go back to the one in Fab Con It wasn’t Fab Con. It was like Data Insight Summit in Atlanta a long time ago. The Business Summit. That’s where we met. Yeah, it was like a long time ago in Atlanta. Atlanta. Yeah. Yeah. it was the same conference center. They took up both buildings, building A and

2:25 took up both buildings, building A and building B. I No, sorry. C and building B is what they were in. It was It was awesome. Good conference. That was the first observation. It was big. big. Second ob- observation was there was so much talk around agents and agentic everything. And I think my second observation here just would be it felt like for once,, remember Tommy, the last couple years at Fabric conferences, we’d go and we’d, hang out, but it was always like a lot of like, ah, we’re still trying to fix things. We’re getting stuff done. We’re getting CCD

2:55 getting stuff done. We’re getting CCD fixed. Like there was a lot of These conferences were announcing a lot of like like getting things cleaned up for us so we could actually have like a really smooth experience in Fabric. [clears throat] This year felt like we were done cleaning and fixing to some degree and we were building exciting new integrated features around agents and large language models coming into Fabric and building and creating. So, I I felt like this year was like a a level of excitement I haven’t felt in a number of years just because there was new things appearing and showing up. So, that that was like my my general feel.

3:25 was like my my general feel. Those are my two observations. All right. Anything else, Tommy, you want to note about the topic or episode before we get into the news? How many people did anyone come up to you ask about the podcast? A number people did. Actually, I had a couple people actually respond to me directly inside So, there’s an app that they use at the conference called the Whova, w h o v a app. And if you go to the conference, you can see what event, you conference, you can see what event,, the event notes and there’s know, the event notes and there’s messages in there, all kinds of other

3:55 messages in there, all kinds of other fun things. I got some very wonderful, let’s see if I can pull through up a couple here,, , messages here. There was one here that really struck me and I’ll see if I can call it out here real quick. Stall for me a bit, Tommy. Talk about something. Okay. Man, it’s good to see your face., I’ll tell you what, the cadence of doing our podcast because you and I were talking about we obviously scheduling was a big thing. It’s like maybe we missed a

4:25 It’s like maybe we missed a Thursday, but what, we’ve kept two in going. That being said, I really missed you for all the Fab Con announcements. There were so many. Yeah, there was. There was a lot going on. Yeah, it really is incredible how how much they actually put together here. So, So, Man, I’m not finding the message here. I’m going I’m literally going through message by message here in the in the Whova app. There was a number of really good messages.

4:58 -oh. Did my Did my Oh, no. Okay, we’re good. I think I’m back. Did I come back, Tommy? You’re Yeah, you’re here. You’re here. My graphics card has been wildly weird when I come back from long breaks and vacations and it like things freak out. I think it’s okay now, but it would like the whole screen went blank, came back, and I think it was my graphics card freaking out. Anyways, okay. I think we’re back. All right, let’s get into our main today. Oh, sorry. The news article. So, I have a news article here from X and this is someone posting something about

5:30 something about basically this is a a note from Satya. Did you Did you watch this video, Tommy, at all by chance? Yeah, so basically the agents you’re spot on exactly what the big focus was. Okay, so I’m going to put this link here in the in the chat window of the the the note here and it was an interesting conversation., Rohan Paul, I think was interviewing Satya and they’re having a conversation around agents and what does this look like? like? And I feel like this is really true., typically, Tommy, software has been

6:02 , typically, Tommy, software has been divulged into software is typically baked into business intelligence. And Satya here is basically saying business intelligence is moving from other systems like software that we program in code directly into agents. Agents are now becoming the method of where the business logic and the reasoning starts to exist. What were your thoughts on that, Tommy? Did you Did you resonate with this message in this this podcast?

6:33 message in this this podcast?, 100% individually and also to what clients have been at asking for and what a lot of my projects are focused on. But, I’m interested for you, how much of his keynote resonated? I don’t understand. Your projects are focused on just making agents? A few The last few projects that I I’ve started working on or been doing discovery, the number one focus is how can we get our data ready for agents? But I don’t think that’s what Satya is

7:03 But I don’t think that’s what Satya is describing here. He’s not saying getting your data to the agent. He’s talking about business logic. So, he’s talking basically software companies are dead. So, SaaS as we know it is is basically a dead commodity. You would have a business process, you’d have a business logic, you would put that into the software application, and that’s what you would use to distribute the content of that business knowledge from the software to the teams that are using it. What I think he’s trying to convey here though is agents are now becoming that layer of business knowledge. Like the

7:34 layer of business knowledge. Like the programmability of agents is where business knowledge is leaving the app because that’s just now becoming a medium and it’s moving directly into the agent. So, you’re I think what he’s what I’m resonating with here. Let me just describe a bit further. Yeah, yeah. I think he’s talking about like tools like MCP servers. He’s talking about skills and skills are the bundling of knowledge of people so that the agent has context to like what does the person know? What does this information mean? And so, as I was kind information mean? And so, as I was unpacking over here the last couple of unpacking over here the last couple weeks and talking about things with with

8:05 weeks and talking about things with with individuals, there’s so many low-hanging fruits in every business that are just task task-based things. Things that you don’t need to be doing every day or things that you’re just moving information around. These are things that we could get rid of. We don’t have to do those things manually anymore. The goal is to automate them or use agents to help us build the automation that we need to do this more efficiently. Anyways, a really interesting observation. I think if Satya is describing the death of software as a service, we’re going to have to really reimagine what we think computers can do

8:35 reimagine what we think computers can do and how we interact with them. It’s going to be greatly change what we do. Dude, I think it’s time to get into Fab Con 2026. All right, let’s jump into the main topic. Main topic of this. All right, let’s jump into all the major updates. Tommy, let’s I don’t know even know where to start. There’s so much to talk about. Do you want to go just by topic by topic? Do you want to get like overall feeling or what? what? I I always love to start with the main blog article. It’s usually that

9:07 the main blog article. It’s usually that one that’s on the azure. microsoft by Arun. He always has like a high-level overview what Microsoft is doing with Fabric. Okay. And there’s really they always have those themes, right? So, the pillars, what they’re trying to do. And it’s interesting that it started off. So, the article will be in YouTube and all the podcasts. And it’s a very high-level view basically what Microsoft focuses for Fabric data in the future.

9:37 future. And just to highlight here, the data base hub, getting your data state ready for AI. So, I think that’s, there’s something to this article. Like there Microsoft is intentional about this So, Tommy, I want to stop you right there on the database hub cuz I think this is a good point. The very first part of the article they mentioned is introducing Microsoft database hub. This is like the big announcement I think from Fabicon cuz this was emphasized multiple times. But we now have many different databases in fabric. We have KQL, we have SQL SQL

10:08 in fabric. We have KQL, we have SQL SQL fabric. We now have Azure Cosmos, we now have Postgres, you have MySQL, any database out there will

10:17 have MySQL, any database out there will basically be able to rerun inside fabric. What are your thoughts to that, Tommy? What What’s your reaction there to the database hub to manage all these different databases? About time. So, no, I think this is the appropriate place where really it’s my first thought was, “Wow, “Wow, I don’t know any reason to really use an Azure database anymore.” That was my first thought of like or using outside sources where I can use fabric for nearly everything

10:47 I can use fabric for nearly everything and just having the capabilities that I need. need. I want to I want to pick on that comment there, Tommy, for a minute. You said you don’t think it would use an Azure database Azure SQL database or managed database from Azure now. Do you consider yourself a DBA or do you just consider yourself more of like a business user who uses SQL? [sighs] I I would not consider myself a DBA. Okay. Okay. Cuz I would also argue the same point for myself. I’m not a DBA. Like I I’m not going to be doing all the little

11:17 not going to be doing all the little bells and knobs and everything on the SQL database, but I use SQL databases. I know how to run SQL. I know how to make the turn the database on, put tables in, get tables out, do queries on things. I think this is the new medium like So, the same thing that Power BI did for report building, which is everyone can now build reports. Here’s a semantic model everyone can create. We’re I think we’re seeing the same thing for now databases. Everyone can make a database. What do you want to use? MySQL, Postgres, Azure Cosmos DB,

11:47 Azure Cosmos DB, KQL KQL databases, Kusto databases, everything’s going to show up. I think this is going to be the new way things are going to be happening. Yeah, I because honestly like even the the database administrator, right? Or the DBA that you talked about. It’s one of those things where that simplicity like yes, there’s obviously a ton if you’re DBA and you do that. There’s a lot you need to know, but Mike, being able to set up a database or

12:17 Mike, being able to set up a database or lakehouse. Yes. and honestly a lot of the main the main metadata, the main those configuration fabric makes it very front center like from the scheme of her having a schema from being able to see the change history to kind see the change history to optimizing your database. There’s a of optimizing your database. There’s a lot here. Obviously enterprise is a different story, but you talk to most of my clients, you talk to most midsize companies and what was their other

12:48 companies and what was their other option? Maybe they already they had something configured in Azure or maybe they were still using,, SSIS and very legacy systems. Now anyone can start Well, I and I think that’s interesting, Tommy. The other thing here I’ll just also call out here the whole database hub is trying to reduce costs for what you pay for. So, if you notice, Tommy, right there at the very end of that they talk about a there’s a little star asterisk area.

13:18 there’s a little star asterisk area. I think I’ll look at the fine print for this one, but they say beyond unified database management, we’re also including savings plans for databases. So, a new way to save up to 35% compared to pay-as-you-go pricing on very selected services. So, not only is Microsoft trying to get you to bring your databases to fabric, but they’re even trying to give you a substantial running discount to run them inside fabric as well. This will be interesting. So, one of the observations I’ve had, Tommy, around the fabric SQL database is it runs a bit heavy, I think. My database is not very large. I

13:48 think. My database is not very large. I have a couple apps on top of it. The database doesn’t really like to run on an F2 very easily. It maxes out the F2 very quickly. So, I believe the way the databases are fabric SQL databases are running today, they immediately turn on an F6 a 16 cores immediately and that’s what they charge on CU usage 16 cores whenever the database is on. So, even if your database isn’t doing a lot of work, of work, heavy work back and forth, it still requires requires some substantial CU usage to run.

14:21 some substantial CU usage to run. I think there’s plans in the work to make that more efficient. I think Microsoft was announcing some more efficiencies coming to fabric SQL databases. We’ll see what that looks like when it comes out, but I wish they would be a little bit more scalable like scale smaller and could scale larger as part of this and I again I think as Microsoft sees adoption, they’ll easily add that as well. Okay, enough for the database one. Second second note in Arun’s article, what’s the topic there, Tommy? Getting your data state ready for AI with fabric. There we go. Okay.

14:52 AI with fabric. There we go. Okay. And this is now now we’re finally seeing the announcements and the releases on fabric IQ or really and foundry IQ and now something called the work IQ. But the amount of updates that they had on the fabric IQ is probably the major feature it sounded like to me based on blog articles and buzz I was seeing online. I don’t know if that’s your impression as well. I think it yes, Tommy. I think I’m going to agree that fabric IQ, foundry IQ, like there’s this IQ language is coming

15:24 like there’s this IQ language is coming up. I’m I’m not sure if I understand exactly what they’re trying to discuss in this IQ space. there definitely is this vibe of we need to give agents more information, more more knowledge about how our business works, right? What is a customer? Where does the customer data live? What is a sale look like? Where does the sales data live? So, these things called ontologies are coming out. We’re talking about making additional instructions. There’s a lot of net new things that are

15:54 There’s a lot of net new things that are coming. And again, it sounds like a lot of this is just being initially released. It’s a lot of it’s in preview. it’s a little bit heavy to run right now. It seems a bit CU intensive in general. general. Yes. but looks like, the work the plan are in works to make this more efficient. I’m going to again, there’s a cycle to software, right? Step one is get the idea out there. Make it work. Just run it, right? That’s step one. Step two comes back and says, “Okay, great. Now that it’s working, let’s make it a bit more efficient to run, right? More more economical.” So, I think that’s

16:25 economical.” So, I think that’s where we’re at right now is we’re kind where we’re at right now is we’re in this initial like releasing phase of in this initial like releasing phase of all these features. I’m a little bit vague on Tommy what even some of these sessions. There’s there’s so many IQs. It feels a little bit like “Hey, we’ve made IQ.” It feels like co-pilot. Like, “Hey, it’s like co-pilot everything.” It’s just like the a word they use across many different services and they’re pulling it all together. So, I think the proof will be in the pudding here of what this really means. means. I saw a lot of claims around fabric IQ adding value to agents.

16:55 adding value to agents. I didn’t really see any hard numbers that helped me understand why the IQ was required to make the agent more effective. So, that was one thing that I felt like I didn’t quite get the connection between like, “Okay, you tell me we need this ontology thing. How much better does it make my agent perform? Does it reduce the amount of consumption it’s using? Does it reduce the amount of tokens it’s using? I didn’t quite get the jump between why and the results.

17:25 why and the results. I understood the results of it’s supposed to make it easier for it to find all the data. I get that. But outside of that, I didn’t really I didn’t really feel it. Yeah, there’s a lot obviously on the focus on the semantic intelligence layer. layer. But also talking about the piece a lot there wasn’t any correct me if I’m wrong. There weren’t anything new with the fabric IQ, right? In terms of like a new feature that they added. These were still have the graph in Microsoft fabric. Oh, no,

17:56 the graph in Microsoft fabric. Oh, no, the fabric planning. That’s the that’s the feature that’s part of fabric IQ. Yeah, so so that one is very specific and and again I I was like, “Oh my goodness, Microsoft releases planning thing.” My initial thought was, “Oh boy, Lumel’s going to be angry about this one.” Like Lumel’s going to be mad that Microsoft just replaced their product. Well, come to find out Microsoft approached Lumel and said, “Hey, we just want to use your product and put it into the actual main product of of fabric.” And so that it’s actually Lumel’s product. So, Lumel’s product is now part

18:26 product. So, Lumel’s product is now part of the fabric ecosystem. Instead of buying a Lumel license, you just consume CU usage and the CU usage is what consumes that planner portion of the the planner item now in the workspace. But again, now that’s that’s taking over the whole planning experience, right? Here’s a table of data. You want to edit them, you want to modify them, you can do so in a planner type mode. So, planning is something I got to play with. I haven’t had a chance to really play with the planner that much. So, that’s a new item I need to go explore or and use a bit more.

18:56 explore or and use a bit more. Have you played with it, Tommy? I have not played it with yet, but it does seem like one of the cooler features because it’s basically what I think 18 custom visuals have tried to do in the past. Oh, yeah. Probably more than more than that at this point. And I think this takes the wind out of the sale of any company who’s looking to do write back to some degree. Here’s a table with tape with information in it. Why do you need write back now? This is this is what you want. Here’s Here’s the information. Edit the information in real time and

19:26 Edit the information in real time and have it update the database. Like this this makes sense to me. Yeah, so what and I think for a lot of people who have not heard of fabric planning, it’s simply allowing you to do budgets, forecast, targets, and scenario models directly in fabric and on fabric data. So, literally being able to do forecasting, operating on,, editing the different numbers. It’s connecting to semantic

19:56 numbers. It’s connecting to semantic models, Fabric SQL, and OneLake. There’s no exporting data. There’s no Excel file where you’re trying to do your budget and forecast and then then try to build a report connected to that data. It’s all built into Fabric in the UI. So, there’s no no code self-serve experience and it works obviously and has pretty good reporting because the interactive dashboard dashboard to get this IBCS compliant in terms of the financial visualizations. Yep.

20:26 the financial visualizations. Yep. That’s one that I would wish people would glean or grab onto a bit more is this IBCS

20:32 grab onto a bit more is this IBCS standards of visuals. I think that’s a very big win for people to to use and see. I think that’s very useful for people to like we should be we should be using already known language around data and leveraging it in where we are in our existing reports and dashboards, which it’s very difficult to get IBCS level visualizations out of the standard Power BI desktop. So, I think this is interesting. Fabric IQ I’m I’m going to see where this goes, Tommy. This is one I think we’re going to need to do some learn Fabric experiences around this. I’m not sure I’ve seen a lot like I

21:03 I’m not sure I’ve seen a lot like I think the community is excited about Fabric IQ. But I’m not sure we have enough examples of like what’s effective and how to really use it and I found Fabric IQ useful in this scenario and here’s when you should use it, right? I’m not sure I’ve seen enough of that yet to really help me like put my my stamp of approval or really get behind it yet. I want to I need to see some more proof in the pudding here. Make sure that this is actually going to be something that will be useful.

21:34 something that will be useful. All right, moving on. Another announcement that Arun alluded to here was unifying your data set with the Microsoft OneLake. So, every AI initiates initiates understanding where your data lives. They now have this more consolidated story around mirroring Fabric, more shortcut transformations like basically a calculated column. You could now one of the the items that were as at the conference that was actually really well received was shortcuts to SharePoint.

22:06 shortcuts to SharePoint. Right from OneLake. Yeah, I yes, yes, yes, yes. [laughter] That’s amazing. What a UI such a better UI experience for SharePoint connectors. So, it’s always been very difficult to get SharePoint level things into a lake house or lake house tables, but like this like auto updating mirroring for SharePoint seems to make a lot of sense, right? It make I’m going to SharePoint. It’s a user-friendly space where I can make lists and I can have documents and I can have I I collect information. Why can’t I get the same information

22:36 Why can’t I get the same information directly into OneLake? That that that really makes sense to me. So, I think the fact that they’re making this a lot more efficient here is very very useful. Anything else that stood out to you, Tommy, in the OneLake area of any features there that you resonated with? The the only one was I know there were you can change your data feed on mirror data. So, just a a lot more expansions on what you can do with mirror data, which I always love., nothing terribly I still wish I saw

23:06 , nothing terribly I still wish I saw in OneLake the OneLake hub the ability to customize that governance side of things. but but that’s okay. Honestly, with all things that’s happened, that’s okay. Oh, and also the OneLake file explorer is now GA as well. Yes, so that again I don’t really use it all a ton. Do you do you use that, Tommy, regularly? not not at all. Not at all. There is now in a OneLake MCP.

23:36 in a OneLake MCP. I don’t know if you saw that. I didn’t see that announcement, actually. Yes, so the OneLake analytics search API so you can use the API, but there’s also in the built-in Fabric core MPC MCP server. So, AI agents can actually locate any Fabric assets part of any workflow., that makes sense cuz right you want to be able to talk to your MPC server and say, “Hey, go find me the file or the data or something that’s like in the OneLake somewhere, a table, right?” That would make total sense that you’d want to give access to

24:07 sense that you’d want to give access to some MPC server. Speaking of MPC servers, I’m actually going to skip ahead here, Tommy, on some announcements here. here., just because this is I think I think this one’s pretty relevant. I had a hard time getting my head around what is the Fabric remote MCP server. So, Tommy Okay. And when when you use an agent, right? So, let me just give you some context here. When you use an agent, you use an agent on your local computer either like VS Code and there’s like some large language model. There was a word that came out at the conference that was a lot more prevalent and I’m seeing

24:37 was a lot more prevalent and I’m seeing a lot more of it on the internet in general. It’s called the word called it’s called the harness. Do what by Okay. Okay. So, this was a language word that people started using. So, let me let me just paint some pictures here. The harness is so you have a large language model. It could be Google Gemini 4. It could be ChatGPT 5. 2. It could be Opus or Sonnet, whatever the whatever the models are, right? Those are the models that run. Well, the models need some

25:07 Well, the models need some wrapper around them like either like a chat interface or a VS Code [clears throat] or Claude Code, which is like a like a IDE that’s like you install it on your machine and then it like uses the large language model in concert with whatever that solution is doing. So, when I look at all this and say, “Okay, this is interesting.” The harness is the software that lives around the large language model, right? So, I would explain this as VS Code is the harness for large language models

25:39 is the harness for large language models that are in your chat window. That’s the harness. How does how does the large language model interact with your code, your code base? Like that’s that’s the harnessing the piece of software that helps you work with the agent, okay? Part of the harness is this thing called the MCP server. It literally runs a little mini server. It’s allowing tools and things to connect to that server and then talk directly to an API or something like that through this MPC server. So, large language model sits in a harness. Harness has an MCP

26:10 sits in a harness. Harness has an MCP server, runs a server, and then you can talk to, whatever the solution is, right? Fabric MCP server, Power BI MCP server on your your your desktop application. It can use that. So, the reason I bring all this up is because now there’s a remote MCP server and that means Microsoft will run the MCP thing in their Fabric for you. So, you just connect to that and then your harness, your VS Code, doesn’t have to run the server locally on your machine.

26:41 run the server locally on your machine. I think this is done in lieu of some companies won’t turn on MCP servers for their organization. And so this the remote server is like, “Hey, I have a large language model. I need to talk to the Fabric MCP server, but we don’t like our organization doesn’t allow us to turn on MCP servers.” It’s a that’s a setting you can physically turn off like in VS Code or or there’s there’s settings that organizationally you can turn these things off on your computers. So, because of that is a security risk, Microsoft’s like, “Well, we’ll just bring it in house and that way their MPC server talks directly to

27:12 way their MPC server talks directly to the Fabric stuff.” This is an area that I think will be very useful to us. I think it makes the MCP server act like an API, basically. It’s just a call,, “Hey, large language model, talk to this thing in Fabric.” And it just handles all the tool calls for you there automatically. So, I still got to figure a bit more about how it works, but I’m really excited about seeing this feature. I think this is going to open up a lot of doors for more organizations. This is actually very common because in a lot of tools that you have probably I know Claude has it, Notion they have connectors, which are

27:44 Notion they have connectors, which are basically basically they have you can build your own custom connector in them, but there’s a lot of built-in basically MCP servers that you don’t have to install. Correct. Yes. And I think, Tommy, you’re finding a lot of social media posts. So, when we look at like Tommy did a little bit of research here, he used a little bit of AI to help us figure out what’s hot in topics here, right? Right. The first one is Fabric IQ. Everyone was talking about Fabric IQ. It’s like the big announcement. Behind that was like operation agents, which we haven’t really touched on yet. We’ll have to get to that one next. And then there’s this Fabric remote MCP

28:15 then there’s this Fabric remote MCP server. This is like third in the list, basically, of of like that’s pretty up there on why? The because it’s it’s Microsoft Entra ID. No API keys, no service app that you have to create. Correct. Correct. No no principal. It’s just literally just log in, which I think is a huge part here. So, let me put the article here on the chat so that way you can see the where this is coming from. There’s there is an MCP introducing the Fabric

28:45 introducing the Fabric public preview of MCP server. I don’t know if it’s officially been announced yet or if you can actually go use it right now today, but this was announced on the remote server has been announced. There’s there’s an original MCP server The local one is out. out. But the the the remote one is just been announced. There’s no actual like pointing to a place where you can go use it yet. So, it’s it’s coming soon. Yeah, that is a coming soon, but

29:15 Yeah, that is a coming soon, but honestly, man, the fact that I could just be in Claude and and rather than having to try and install something because I know a lot of people who created Fabric MCPs like in the community. But there’s a lot of configuration to do so. Not that you can’t do that. But to your point, Mike a lot of people don’t have access or permissions to go to certain places on their computer to be able to install or configure a local MCP. Well, think about it. Like there’s a lot of

29:45 think about it. Like there’s a lot of like like I feel like a lot of what’s coming out with agents is a lot of pro dev stuff. Like we’re we’re we’re right now we’re getting a lot of like net new things that are like pro developer stuff., set up this agent, use this desktop VS code, talk to this agent, you desktop VS code, talk to this agent,, use CLI on thing like it hasn’t know, use CLI on thing like it hasn’t been simplified yet. Like so we’re right now in the everything’s exciting, it’s heavy technology, it’s super complicated, I guess, to some degree. It hasn’t been distilled down to like simplified for like the broader audience

30:17 simplified for like the broader audience yet. I think it will get there, but we’re still in the phase of like it’s just brand new technology and and therefore it’s a bit more rough around the edges. Is that your impression too, Tommy? Yes, yeah., you literally are my segue. Speaking of developer experiences, oh my goodness. The The love of Git, we have a ton of updates when it comes to Okay, this is going to make you happy, Tommy. You ready for this? start. Yes, okay. There’s three

30:48 this? start. Yes, okay. There’s three previews. previews. Branched workspaces with Git integration, selective branching with Git integration, comparing code changes with Git integration., there’s a ton ton of stuff going on here. Yeah, so you said branch workspaces, right? Because I’m trying to Yeah, so selective branching, branch workspaces. Yep. and then also the notebook integration with any selective resources. And I think we’re just touching the surface still. Yes. There’s a ton of stuff here. So

31:20 Yes. There’s a ton of stuff here. So what what what excites you most about this Git integration, Tommy? So I think the biggest one for me right now is the idea for the branch workspaces. It was always always kind, it felt very much like a sometimes a square peg in a round hole when it was working with branches and workspaces because they were not really made for each other. Even though they, they don’t they have somewhat of a familiar feel, it was not something that was very intuit- intuitive and also

31:52 that was very intuit- intuitive and also it got more confusing than if you didn’t do it. Mhm. Now with branch workspaces, it there’s a very seamless way to simply create a formal relationship between whatever your source workspace is and any workspaces that are created from branching out or creating a branch. And it’s more than simply just having your workspace connected to a branch in a repo. repo. you can actually have any selective branch branch out only the items that they need rather than the entire

32:23 they need rather than the entire workspace. Okay, that to me is I think a really big move here because this has been a pain point of mine cuz right right now it’s like all or nothing kind right now it’s like all or nothing today. It’s been If you want to of today. It’s been If you want to create a workspace branch, you have to take everything with you. So I think this is starting to become a better story. story., honestly, Tommy, like if there This is huge. People got to start learning this. I People got to start learning this., this is a this is a pure play from mean, this is a this is a pure play from the software world, especially if you’re talking about agents. I I’m I’m very pro

32:54 talking about agents. I I’m I’m very pro on GitHub in general, but especially when you start working with agents and agents are going to appear and that’s like that is the way people will interact with their code and build things moving forward, you’ve got to have checkpoints on this stuff. Like this is a non-negotiable for me at this point. point. And Mike, I I even put in our agenda and I’ll save it for another day, but is there any excuse to not by default for anyone who’s a Power BI developer to be using Git?

33:29 Yeah, I’m going to say at this point there’s so there’s so much usefulness with it and it’s so lightweight to get started either in just a workspace is at a minimum just synchronizing your workspace with Git. Like that’s like the bare minimum and I think that’s very understandable. Microsoft has provided very easy user interface on workspaces to make this work. It starts there and as you get comfortable with that, you can really grow your knowledge into like deep Git integration across the board. I I think it’s a no-brainer, Tommy. Like I and the fact that Microsoft’s continually

33:59 fact that Microsoft’s continually investing in this and building new patterns around Git integration and how this works, yes, 100% this makes sense to me. to me. So yeah, I think the branching right now I I I actually have three different teams that I’ve worked with that I’ve had to meetings about like should we use Git where I was like, well, based on what you’re doing, it was very much a caution, but now this is one of those like, hey, this is available. Your biggest fears and concerns

34:29 biggest fears and concerns are now answered. So I love that too because again, the biggest thing for me when trying to branch out was you have to kept copy everything, which I hated. So like so now this is this is true collaboration to for me, too. Where I need to just a few people to work on a certain items. They don’t need everything. So a huge part also for agents, too. love it. So thank you. Also the Fabric CLI command line interface is has

35:02 Fabric CLI command line interface is has more advanced features as well here. It’s generally available this the library integration for CICD for full workspace deployments directly from the command line. It also supports AI agents. Well, that I think this is a, I honestly think that’s where this is going, right? It it If you can just if you can just tell your agent to go do things and it just works, like now we’re talking, Tommy. Like

35:32 now we’re talking, Tommy. Like like that’s this is where you don’t even really need to know what Git is doing. You just need to talk to your agent and say, okay, go get the main branch, go commit my changes, go push the changes up. The agents are now getting very aware of like how to work with this stuff as well. I’m using Git heavily with my agents that I’m using in our organization and so I feel like that’s also a big win for us as well. So having that automate a lot of our work there Yeah. Yeah. in concert with agents is the way to go. All right. What other news do we have?

36:02 What other news do we have? know this isn’t officially from Microsoft, but I know you follow Kurt Buhler Data Goblins, but he seems to be starting his own Microsoft developer experience, too, because I’ve been following him on Twitter and just lately [snorts] and he has a PBI-R command line tool. Yes. Yeah, so that’s See that? So the Well, let me give a little context to this one as well, right? So at the at the Fabric conference, I did a full day pre-con session around agentic

36:33 full day pre-con session around agentic coding coding with Mathias Tierbach. And so when we did this presentation, everything was around using the Power BI MCP modeling server. So this is a Power BI MCP server that you use with VS code, you pick your agent of choice and you go talk to the data model, make folders, folders, document things, move stuff around, have it build things in the model. Immensely useful, right? 100% valuable here. On top of that, we talked about building skills and things, okay? All of this is

37:04 skills and things, okay? All of this is very helpful, but the most important question people were asking us was, hey, by the way, how do we edit the report side? What does the report look like? How can I take this modeling chat experience and translate this to build reports with something? And so this didn’t exist and Kurt was starting to dabble in this a little bit. So what Kurt and Kurt and a gentleman that he’s working with to develop this, they’re producing their own MCP server

37:34 own MCP server or skills library, basically. It looks like it’s a bunch of skills essentially. But they’re building skills around this so you can actually use like a command line interface and say, build this report. This is what I want. Describe the report pages that you want to produce and it will just build out most of them for you. The The format for PBI-R, that the report format, is well documented, but it’s also quite difficult to get it to work correctly. So this is where a lot of

38:05 correctly. So this is where a lot of instructions and examples are being given directly to the report layer so that the agent can understand, oh, this is the file structure, this is how it’s built, here’s the interchange between the different files and things. And so he’s been releasing recently and I’ve been talking with him a lot on Twitter right now or X a lot about what is happening in this report building building AI agentic space, which I think this is going to be the future of things, Tommy. I think very quickly here people are going to expect to build reports with AI

38:37 going to expect to build reports with AI and agentic experiences as opposed to just going in and brute forcing clicking on buttons. And one of the posts Kurt recently made on X was every time you build a report, there’s a very heavy click tax. Like clicking on things a lot. There’s a huge click tax on the the field the the settings panel. The settings pane is just so much clicking to get anything done formatting wise. wise. And so that is reason to go one, push into theme

39:08 is reason to go one, push into theme files cuz then you don’t have to do with the clicking, you can get standard off the box or off the shelf visuals much faster, but then two, agents can do all that set up for you. I don’t like this, remove the title. Like you just describe what you want to the visual and they can go find the properties and then make those adjustments for you directly inside the PBI-R files. So I really like this format. I think this is going to be great. It’s it’s really neat there. I’ll see if I can go find the article, Tommy, if I have the link to it. it’s probably a GitHub page. I’ll keep going on the next topic and I’ll I’ll go find it. So yeah, it’s it’s this latest updates.

39:39 So yeah, it’s it’s this latest updates. He has the PBI-R and also a repo on Power BI agentic development. All right. So, So, the major themes here, I know we’re we’re trying to run through everything or just the highlights here. Just a couple things. Yeah, yeah. Not everything. No, no, but the Fabric IQ, the the developer experience. Yep. It seemed to me, Mike, at least what we’re highlighting here, the major focus,

40:09 the major focus, of course, is obviously the agentic side, but the the major themes of where Microsoft was investing their time, and also what they’re working on, is a lot, Mike, on the developer and being able to work in Fabric outside of Fabric. Yep. Correct. Seems to be a big part here, and I I want to get your thoughts on based on the other the sessions at FabCon outside of just the keynotes. when it comes to what we can do with Git, what we’re being

40:39 we can do with Git, what we’re being able to do with it from the agentic aside of things, the ton on user-defined functions that they updated. There’s simple to view on the web. how did you feel the sessions were in terms of who the audience was for? I’ll I’ll Great question on that, Tommy. [snorts]

41:03 [snorts] I would say a lot of this was just information around AI and agents, and where they’re going to be integrated into the Fabric space. So, so largely to this point, Tommy, I feel like there’s a lot of lot of agentic things happening. Agents are showing up. We have people going to want to chat with things. How do we get access to them? Where do they live? Up until this point, it felt like Fabric didn’t have a story around where you plug in the agent, right? It was more about Copilot, and then trying using Copilot in various experiences. I don’t find that very helpful, honestly. Wherever they were putting Copilot wasn’t very useful for me.

41:33 Copilot wasn’t very useful for me. It feels like we’ve shifted gears a little bit, and moved away from Copilot to give you insights to agents to help you build you build the creator experience. And to me, that’s where I think more power comes from these agentic experiences, where I don’t want an agent. I don’t want to It seems incredibly wasteful for me to continually ask the agent for, “What’s What’s this month’s sales? What’s next What What’s the projection for next month?”, that’s

42:04 , that’s that’s well-known information, and I should say, “Hey, agent, I need a visual representation of this data model that gets me this year’s sales versus last year’s sales month by month. I should be able to describe what I want, and from the description, it builds me the simple to run, to run, cost-effective thing that I can use over and over again. Does that make sense, Tommy? Like, I don’t want to spend Yeah. like 100 CUs per question, whereas I could run it for like peanuts,

42:35 whereas I could run it for like peanuts, like like couple CUs, if I built a semantic model and report, and just had the report show up. So, I think there’s this idea of like, are you trying to use the agent to do regular tasks over and over again? That’s not an effective use of the agent. And I feel like that was the the message here, which was at Fabric conference, that shifted for me. It was less about use the agent to get insights on data. Instead, it was more about build your Fabric environment with agents. Like, to

43:05 Fabric environment with agents. Like, to me, that makes a lot more sense. Yeah, and it seems like we’re a lot able to Do you see that more for the admin or Say you’re in a team, and with all these things coming out now, if you were the data czar, you had a BI team, does everyone have agents? Obviously, Git is a no-brainer. But, this very much feels like that focus on the individual though. Like, for me, my own company, this is all great.

43:35 great. Yep. How much are you giving out to a team right away? The MCP server, being able to kind [snorts] of allow them to start just be doing more work from chatting, so to speak, rather than the actual development. I see what you’re saying. Okay. So, already, I think a lot of people So, let me go let me go back to a session that I went to attend one of them. Sujata from the part Microsoft product

44:05 Sujata from the part Microsoft product team basically did a session of like, expert versus AI. That’s what the the name of the session was. And so, the principle was, let’s give AI to a novice user, and see how well they perform against an expert building things inside anything with agentic experiences. Report building, model building, like there’s let’s take two examples, and how does this work? So, So, what I felt like here in the session was, if you give AI to a novice user, a beginner user,

44:35 beginner user, you get a you get a a higher level of performance out of them., more value. You get you get useful output from them. But, is it the same quality of output as you would give an expert with the same agent? agent? And so, what we what the the discovery was is experts can do a lot more with agents, because they already know like how the model should work, what measures we should be building. Do we want a simple model, or do we want all the measures? Like, there’s some there’s some handcrafted pieces of this where you

45:05 handcrafted pieces of this where you need people involved to like shape the direction of where the agent is going, okay? okay? But, the general principle was you can get like a novice user maybe a two or three X usefulness output on that novice user. But, when you go to an expert and give the expert that same agent, you get like a 10 X output. It’s like way more effective, and the time to produce the same results drastically decreases. So, this is where I would argue

45:35 So, this is where I would argue the the real value of this comes from. So, Tommy, you’re asking,, am I giving this to my team? Okay, that’s a very weighted answer. My first question is, do we have Git enabled? [clears throat] Right? If we’re if we’re if we’re beyond that hurdle, then yes, we’re going to start moving more towards this. Remote desktops, we’re going to give PBIR. Internally to my company, we’ve been for months working on building the skills to use this. So, I do think this is an educational space. I have found personally immense amount of value from using agents to help build

46:05 of value from using agents to help build models, manipulate models, create models. I’m seeing people now blogging a lot about, “Hey, I’d had two models, and I need to merge them down to one. I used an agent to look at both of them independently, document them, and build a a new model that is the merge of these two other models.” Great. Like, excellent use case. And Right. that saves a ton of time, right? But, it needs like an expert to come in and say, “Here’s what I’m going to be pulling together. Here’s how I’m going to be building this system, right?” So, I I really do feel like this is

46:36 really do feel like this is this is a really useful tool once you understand what you’re doing in in Power BI [snorts] [snorts] already, right? And if you’re a novice user, it’s going to make you very effective, but you’re going to still have need to learn some skills around like, what is modeling? How do we serve the model? What’s good practices in modeling that you give models to your end users? Like, you can learn this stuff as you go. I think it’s I think it’s learnable. But, I’m definitely all in on the using

47:06 But, I’m definitely all in on the using AIs to build stuff. Like, it we’re going to do it. I’m going to teach everyone I can. I’m drastically moving my company away from pure consulting on Fabric and Power BI, and it’s more becoming agents. Like, we’re the AI integrators, right? We want to integrate AI into your company, and start leveraging like, these tools exist. Is your team equipped to use them? What do you need to put in place to leverage these tools? What safeguards should we be thinking about these tools? And it’s all focusing more on AI and

47:36 And it’s all focusing more on AI and agentic spaces, cuz I think if you don’t move with the way the the market’s going here, you’re going to get left behind. I’m curious about what onboarding or training looks like, right? Because when we deal with all these agentic things, because you would definitely agree that the the agentic developer, the agentic user is That’s a skill. That’s an That is a hard skill, no different than DAX, or using Microsoft Excel, or using Power

48:07 or using Microsoft Excel, or using Power BI. Being able to use AI effectively. Sure. Sure. Okay., it it is a skill, right? But, in addition to that, the tools that are being produced, right? So, if you go to the Fabric MCP server, one of the things we taught in our class, which was use the Fabric MCP server. server. Ask the MPC server what it can do. The agent’s really good about reasoning about what it knows how and understands how to use and do. So, the first thing you need to do is, what What is this thing? How do I learn

48:37 what What is this thing? How do I learn it? it? What better way to learn than on the job using the actual tool to do the things, right? So, this is where skills and, MCP servers, they’re already documented well enough for the agent to understand. Like, the agent has to know how to use How do I say this? The MCP server and the skill, whatever you’re using inside an AI agentic experience, it has to know how to use the tool. Like, the AI has to. So, if you don’t build if you don’t build an MCP server

49:07 build if you don’t build an MCP server or a skill that the agent can understand, there’s no way it can explain it to the user. So, by default, you’re getting like this self-trained experience inside just using the tools. Again, it’s it’s now, what do I teach my people? I teach them, “Hey, there’s this MCP server. Here’s how to set it up. Go install this, and run this to get it going. And then ask it, “What can you do? What tasks should I be using?” The agent reads the tools and

49:39 using?” The agent reads the tools and says, “Here’s the tools I have. Here’s how we would use them.” So, the agent is basically self-teaching the user how to leverage the tool. Mike, I’m going to give you one of my I see I’m secretly IT and conservative takes here on what you just said because sure, sure, you can have a team just start asking what can you do, but one thing I’ve been finding out just in all the schemes of different projects I’ve been working on Fabric and outside of Fabric is

50:11 of Fabric is like really setting up the environment, but also creating my own skills, too., specific for what I’m doing. And I don’t think a lot of people know that. that. And that’s for me an expectation I’m getting more and more or rather I’m beginning to lean more and more in the direction of if you’re going to be using our MC Ps and going to be using we’re going to give you a subscription for Claude and using this agentic stuff, you

50:42 Claude and using this agentic stuff, you better know how to set up AI correctly both from creating your own rules, what it’s looking at. I’m seeing that more and more important because you can can not that you can mess it up, but you’re still dealing with tokens. You’re still dealing with a lot if you’re just starting from, from a high level, it’s a lot of generalities especially when it’s going to be with this this I know Microsoft has released a lot of like agentic skills out there. Do a lot of people know that?

51:12 of people know that? Even though it’s available. So, anyways, that’s a as we’re moving to a new phase of

51:19 new phase of, having these things available and not just now they’re new, now they’re just part of the workflow just like I expect Power BI desktop to be working on my computer, Microsoft is providing the AI in that same fashion. Well, Well, I’m thinking more and more about what that onboarding looks like is looking like. Okay. And that’s only because how many developer experiences and individual experiences I’m seeing from Fab Con. Mhm. Mhm. Well, already so to be clear

51:49 Well, already so to be clear here, right? Microsoft is setting the pace as to what this is doing, right? There there’s already there’s already an AI assets library of skills that are already being developed for Fabric CLI core and Fabric CLI PLP Power BI as well. So, there’s already a lot of things that are already starting to be evolved from Microsoft and they’re trying to open source this as well, right? This is this is a real feature that Microsoft is trying to produce here and I’ll put the link here in the in the

52:19 and I’ll put the link here in the in the chat window as well. So, where do you learn this stuff from? How do you get better at this, right? This is one of the reasons why you want to listen to the podcast is because this is the information like we are in it, we’re looking at this stuff, we’re figuring out what’s useful, what’s not useful, and we’re trying to direct your attention towards the things that are that are going to be needed in your organization and business moving forward. So, I really do think that the skills area is actually really important here. And Microsoft is already investing in building them. They’ve noted they’re going to be making more of them across different experiences through Fabric, through reports, through

52:50 through Fabric, through reports, through modeling. They’re going to continue investing in this and I think you’re going to we’re going to like where this is going. So, I’m putting the link in the chat window here as well. This is the skills that Microsoft is producing in their Microsoft Fabric CLI library., and then there’s a CLI core and a CLI Power BI where there’s additional skills being showing up here now that you can go ahead and use. So, there’s there’s going to be a lot of additional scripts and references and what the skills can do,, what operations are available to the skill. This this is really I think

53:21 skill. This this is really I think effective for making your agent useful, right? Without these things, your agent isn’t powerful. And one of the things Tommy I’ll just point out here, too. You asked the question like how do we how do we get to this place? This is one of the a couple of notes you made here. How do I make a skill? How do I get it to do a series of tasks? Like these are things that people need to learn. This is something brand new. And I’m finding already I’m working with my team, I’m working with companies. We

53:51 my team, I’m working with companies. We are working with them to help them say, “Look, here’s how you start developing your own skills. Here’s how you build your workflow.” And And I guess one thing I would just comment here is the the ability to make a skill is a skill in itself, right? You have to know [snorts] like people have to learn this. It’s a lesson people have to learn. And it’s a bit of an artwork. And the skills are these living instructions that probably should be revisited every couple months and updated and modified and improved.

54:23 updated and modified and improved. But I don’t have to physically write them. And like this is the aha moment. Like I don’t have to write the skill myself directly. I can actually go tell the agent, “Hey, do this, do this, do this, do this, do this. No, don’t do that, don’t do that. Yes, build this.” Right? I’ve done this before where I’ve actually asked the agent to do something very specific and I’ve given a sequence of commands and then after it was done, the trick here is ask the agent to stop, reflect on what you have learned. So, actually physically telling the agent to pause,

54:55 physically telling the agent to pause, “Based on what we just did, what did you learn, agent?” And it would describe what it learned. Okay, and then once you have that learning, that lessons learned, you then say, “Great. Now I want you could you again you can shape the learning experience like hey, and I want you to add this and, you can you can basically adjust that while the agent’s thinking with it. Then you go ahead and create the skill and say, “Okay, agent, now create yourself a skill for this activity.” And so, I’m finding a lot more of my work being done is I need to do something in very metered ways. Do one, do two, do three, stop, ask the

55:27 one, do two, do three, stop, ask the agent to reflect on what it learned, then ask it to build the skill and you get much better skills being written, right? And when you use a skill and something happens wrong and it didn’t do it correctly and you need to fix something, again, pause, ask the agent, “What did you learn in our working flow here of this?” And then go update your skill with what the lessons you learned, right? So, these things are going to have to be like modified and continue to be evolving. Is you’re smiling, Tommy. I’m assuming you’re doing the same thing. Yes., honestly, I’m at the point

55:58 Yes., honestly, I’m at the point now where I’m if I’m starting a new project or, I always start with, telling the like agent, “I’m work I’m going to start working on something like this. I have some resources. Yep. Can you help create some skills for me before we begin?” Yep. And it just becomes more and more essential. And I even had Claude code build me a command line tool called Skill Vault that can actually sync all my skills across the different

56:28 sync all my skills across the different folders specifically Copilot and all those in different computers. Yes. I have not released that, but I should. But, just because it is the skills are so essential. So, you’re touching on a point, Tommy, that I think is going to be the next problem. You’ve already you’ve already identified it, right? I’ve got many different tools. I’m in Fabric notebooks. I’m in VS Code on my desktop. I might have multiple computers. That’s just you. That’s just one person. That’s just Tommy trying to figure out like where do you put all your skill how do you centralize your skills into a place?

56:58 skills into a place? Well, multiply that now times 5, 10, 15 users that are developing things. Like who’s building what skills? Does my skill And this is also like at a at a corporation level, do we need to have different skills per engineer? Is that useful? Like does Tommy have a slightly different skill than what Mike does or is there like a base skill that we modify for how we work? Like I don’t know yet. Like this is this is all still very like very like flexible in like what we can actually build and develop here. So, I think there’s going to need to be a lot of conversation, Tommy, as we get into this

57:29 conversation, Tommy, as we get into this more., more., one of the things that I’m really interested in around one of these developer experiences. This was announced at Fabric conference, but it’s talking about the developer experience around agent skills for Fabric. Yes. This is another one. It’s in it’s kind This is another one. It’s in it’s like in preview. It hasn’t a skill of like in preview. It hasn’t a skill toolkit basically., you need to understand what this means more. Like what does this look like in Fabric? So, I I want to unpack this a bit more. I think there’s going to be a problem very shortly here where how do we collect all the skills that our company needs? Who manages those skills? Do

58:01 needs? Who manages those skills? Do skills go through reviews like everyone else? Is there an automatic way to review and vet the skills are clean and not being weird or doing weird injection things. Like we’re going to need a system to vet this stuff. So, stay tuned. I’ve got stuff coming down the pipe that are going to be improving that in with all of our workloads. So, I think this is going to be a really exciting space to build net new things. Anyway, so I I know we’re getting near time and I think we’re probably going to devote next week. Okay. So,

58:32 Okay. So, wrap this up, Tommy. give a shout out real quick to some Power BI because Power BI is still a thing. thing., but there are now in Power BI, if I could find the right one. Oh, yes. Modern visual defaults. I think we’ve talked about that before. That was not the one. Ah, custom totals. I thought I wanted to give that a little shout out as well.

59:02 give that a little shout out as well. Some people might have not seen this, but, Now currently in preview. If you’ve ever been on LinkedIn before and happened to follow people who talked about Power BI, you may notice one time or another some people who disagreed with the way totals were totaled. Well, Microsoft has finally listened to that. So, I just thought that was a little nice possessed touch to it. [laughter] And it’s literally it’s literally on the table. You right click it and say custom

59:32 table. You right click it and say custom sum and it just adds the items in the table that you can already observe and see. I think this is make sense. This is actually like an in visual calculation it feels like it like a a visual calc Yeah. which makes sense to do it that way. So anyways, really good useful case there. I’m going to put the link to specifically that item. So there’s a post that has a link to this one. So here’s the custom visual totals. I think that’s what they’re going to what they’re going to call it now. this solves a major pain point for us which is the totals in the visual should always equal the sum of those items that

60:02 always equal the sum of those items that are shown. None of this filter context across the entire model. I think this just needs to exist for people who want to simplify how they build DAX for table. it just makes more sense. So that being said, awesome. thank you very much for listening to the this episode of the podcast. podcast. We hope you got some good information around what’s coming down the pipe for Microsoft Fabric. There’s so much to unpack. We will definitely have some more episodes on this and we’re going to highlight some more features that we are really excited about and as we integrate and play with them, Tommy and I are

60:32 and play with them, Tommy and I are going to have to do some more videos around how to use them. what what are we finding that it’s valuable and you we finding that it’s valuable and, what is a pain to use and we’ll know, what is a pain to use and we’ll give you our honest opinions as well as we dig into these features and start using them as they come out in preview. That being said, Tommy where else can you find the podcast? You can find us on Apple, Spotify, wherever your podcast. Make sure to subscribe and leave a rating. It helps us out a ton. Do 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, leave your name and a great question, and finally join us live every Tuesday

61:03 and finally join us live every Tuesday and Thursday at a. m. Central and join the conversation on all powerbi. tips social media channels. Awesome. Thank you all so much and we’ll see you next time. Play. Explicit Measures pump [music] it up PR high. Tommy and Mike lighting up the sky. Dance to the data laughs in the mix. Fabric and AI get your fix. Explicit Measures [music] [music] drop the beat now. Podcast kings feel the crowd.

61:34 the crowd. [music] [music] Explicit Measures

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.