Why Fabric IQ Will Cause Friction – Ep. 489

Mike and Tommy tackle Microsoft’s new Fabric IQ announcement and its ontology-driven semantic foundation, warning that defining shared business concepts across departments will expose organizational maturity gaps and spark friction. But first, Tommy drops a hot take on MCP servers vs. IDE-based instructions for Power BI semantic model development—arguing that PBIR files and VS Code skills already deliver the same results with less setup and fewer tokens.

News & Announcements

• Introducing Fabric IQ: The Semantic Foundation for Enterprise AI — Microsoft announced Fabric IQ at Ignite, introducing a new semantic foundation within Microsoft Fabric that unifies data, meaning, and actions into a single layer. At its core is the new Ontology item (now in public preview), which connects people, processes, systems, and rules into rich business entities—enabling AI agents and users to reason and act with confidence rather than reconstructing context from scattered definitions.

Main Discussion: Fabric IQ and the Friction Ahead

Beat from the Street: MCPs vs. IDE Instructions

Tommy opens with a hot take on MCP servers for Power BI semantic models. While MCPs like the new Power BI MCP server let you communicate with desktop via XMLA endpoints through a chatbot, Tommy argues he’s getting equivalent results using VS Code or Cloud Code with instructions and PBIR files—without the overhead of running a server.

Key points from the debate:

• Setup speed — Tommy claims he’d be up and running with IDE instructions before Mike finishes setting up an MCP server from scratch
• Token efficiency — Using instructions consumes fewer tokens than MCP round-trips
• Service vs. local — Both agree MCPs shine for service-based tasks (API calls, authentication), while local file manipulation works better with direct IDE instructions
• The landscape is shifting — Three months ago everything was MCP; now instructions are proving just as powerful. Mike notes the AI space is “building on sand” with goalposts constantly moving

What Is Fabric IQ?

Mike breaks down the Fabric IQ announcement from Ignite. It’s not a single button or deployable feature—it’s a branding shift around making your Fabric environment intelligent. The centerpiece is the ontology item, which functions like a graph database connecting business entities (airports, flights, pilots, customers) with relationships the computer can understand.

Tommy provides the computer science definition: ontology is a formal, explicit specification of a shared concept—machine readable, clearly defined, agreed upon by a community, and representing an abstract model of how something works in the real world.

Where the Friction Lives

Tommy identifies three friction zones: cultural, political, and roles/responsibility.

Cultural friction — Organizations have established habits around data and technology. Teams already have their processes, centers of excellence, and self-service definitions. Fabric IQ asks them to change.

Political friction — Departments guard their data ownership. Marketing owns marketing definitions. Sales owns sales metrics. When ontology forces shared definitions, turf wars erupt—especially when bonuses and performance metrics are tied to how data is defined.

Roles and responsibility — Who owns the ontology? Central BI teams aren’t necessarily data owners, but they need to bring teams together. Engineering defines product dimensions, finance sets pricing, marketing creates categories—all contributing to the same entity. Someone has to govern quality and consistency.

The Buy-In Problem

Tommy argues the first level of friction isn’t definitions—it’s convincing organizations why they should do this at all. Fabric IQ isn’t for better Power BI reports. It’s preparation for AI and agentic experiences. Getting a BI team (traditionally associated with dashboards and KPIs) to lead an AI-readiness initiative requires executive sponsorship and a fundamental shift in how organizations view the intelligence function.

Mike pushes back on multiple executive sponsors, arguing that creates competing priorities. He advocates for starting small—experiment within a department, find wins, then expand. The article even notes you can build ontology from existing semantic models, meaning organizations are already doing this work in fragmented form.

Maturity as a Prerequisite

Tommy asks whether organizations need a minimum data maturity level before attempting Fabric IQ. Mike’s response: the ontology will expose your maturity level. If defining basic terms causes infighting, that reveals cultural gaps—not a technology limitation. Organizations with healthy alignment around definitions will move faster; those without it will surface the dysfunction they’ve been papering over.

Starting Points and Practical Advice

Both agree on practical approaches:

• Start small — Pick a department or use case, not the whole organization
• Think strategically — This is a long-term initiative, not a three-month project
• Leverage existing models — Semantic models already define relationships and metrics; ontology extends that foundation
• Focus on decisions — At its core, all of this is about making better decisions with data
• Get community input — Both are eager to hear how organizations are approaching this, recognizing it’s uncharted territory for everyone

Looking Forward

Mike and Tommy see Fabric IQ as an evolutionary step—accelerating conversations organizations should already be having about what makes their business tick. The ontology bridges the gap between mental models of how businesses run and actual data that supports decisions. They plan more episodes diving deeper into ontology implementation, reusing semantic model investments, and the governance frameworks needed to make this work. Mike warns that Microsoft needs to stick the landing on this feature—it can’t go the way of metric sets.

Episode Transcript

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

0:00 Heat. Heat. Morning everyone and welcome back to the

0:31 Explicit measures podcast with Tommy and Mike. Hello everyone and welcome back to the show. Tommy, how you doing? Dude, I’m doing great. We we’re in the middle of December. we have already been doing a lot on our topic today, but it’s good to see your face as always. Yes, this is a recorded episode for those of you who are listening ahead of time. just a quick call out for those of you who are listening in real time. If you’d like to hear these episodes as soon as they come out, please go over to our YouTube channel and become a member. We will publish these meetings or the podcast episodes as soon as we create them for members only. So if you want to

1:03 Have the earliest knowledge around all the episodes that are being recorded and when those things are coming out, we’d love for you to join us on our YouTube memberships channel. That being said, , our main topic for today will be this new feature that was announced at Ignite, which is called Fabric IQ, which encompasses, it’s fabric IQ, I think, is like a general term. There’s not really a single button that you click or something you deploy that is the fabric IQ. I think it’s this mentality of, hey, did that we can have like these semantic models and this thing called an ontology and really

1:35 Unpacking like the data that it lives in your organization? And there’s a lot of tools being developed to make your fabric have an IQ or have some intelligence with it. So I think there’s this branding shift that is happening here and we’re going to unpack this one. We think and Tommy and I were discussing this a little bit offline. This new fabric IQ will cause friction in your business. So be prepared for it. This conversation is going to unpack what friction do we see coming? How will we deal with it? and maybe some potential solutions here around how do you address this and what do you what do we see

2:07 Coming as BI professionals in the space? What’s going to what conversations will you be having? So, this is a a a visionary podcast maybe a little bit. We’re looking into our crystal ball here and trying to figure out a little bit what the future holds for you. At the very least, we have a tagline. if nothing else, it’s controversial. So, we’ll go there. All right. That being said, Tommy, you’ve got a little bit of a a beat from the street here. you want to have some thoughts on something that you’ve been working with. I do. So, it’s figuring out bit of a it’s a bit of a hot take with the MCPs and it’s going to be focused more around

2:39 The MCP semantic models. this is not including here my discussion here around the fabric MCP which is more admin more everything in the service. Okay. So, a lot of people are all gung-ho and a lot of people are also creating MCPs around PowerBI too. One is a semantic model connected live. Did something open on your desktop? Sure. And it’s great, but I would argue, Mike, that because of the developer role of what the PBI format is, what they’re doing with idees and git, it makes a lot

3:14 More sense and I’m finding a lot more utility using IDE or even cloud code with creating a cloud skill around fabric and semantic models than creating MCP itself. So again, I’ll break that down. I can use VS Code. I can use cursor. I can create instructions which is like the manual way to create an MCP. It’s a very in a sense static way outside of an MCP. But all the things I would want to do in semantic model and I’ve done this to create DAX measures to obviously all the

3:46 Descriptions and I’ve allowed that to look at the semantic model files to create DAX to look at my relationships to optimize. Yes, with MCPS I can connect to Claude Mike and I can use the chat in Claude. It’s really cool. But to me, Mike, I’m like, I’m already doing all this stuff with PBIR. So, I’m already needing to know the way of the land working in an MCP or working in the files, the semantic model files. Like, it’s a lot of work to right now.

4:19 It’s still a lot of work more or less to set up an MCP, especially connecting to anything that’s a service. it. You can do it, but there’s setup with VS Code, with an IDE. I all I have to do is write instructions and give it I can give it documentation and I’m ready to roll, ready to run. And I’ve found a lot of utility where I don’t have to have the MCP running at all times either. So, let’s let’s for context for people who don’t really understand MCPS, I want to unpack this a little bit so I so I understand what you’re referencing here, Tommy. The the MCP server is basically a little bit of software that runs on your

4:53 Machine, right? So, , Ruie has just recently announced there’s a Microsoft PowerBI fabric. I guess it’s a PowerBI PowerBI MCP server. So, it allows you to talk and communicate directly with PowerBI desktop and it says I’m able to communicate with the XMLA endpoint back and forth between a chatbot in some IDE or some chat system, whatever that may makes sense, right? You have to be able to run the MTPC server. The chatbot has to understand the commands the MPC server can do and then you’re communicating with the

5:26 Chatbot. The chatbot then uses MCP and then makes changes directly to the model, code changes, whatever those things are. So that’s that’s what MCP is doing. And Tommy, what you’re saying is look, if I don’t use the MCP server, meaning I don’t have to have VS Code up and running, I don’t have to have all those that ser installed on my machine to make it work. You’re saying look I could instead use cloud code or VS code which again is a code editor the same program but instead of turning on the MCP server you’re providing additional instructions to just

5:59 Manipulate the files directly. So skip the whole server side just manipulate the file and then you’re able to then make the measures or changes or or things you want directly on top of the files and then desktop’s picking them up and automatically adjusting them. Yeah. No, a thousand%. And to that point too, why would you do that? Well, there’s optimization. I’m using less tokens when I use an MCP. And then two, again, the setup like I can get started. If we both did a race right now and you were going to start with the semantic model VS MCP and I was going to create a

6:32 Project in a sense in VS Code with instructions and skill, I’ll be set up and up and running way before you are if we’re starting from the ground. And I think that’s a that’s a big deal to me where I can still accomplish semantic model. Again, I’m not talking notebooks and the fabric and everything in the fabric service in that project accomplish more or less just as much as the MCP can. I’m going to I’m going to I think I agree with you Tommy. One thing I would just maybe note as I think about or

7:03 Unpacking this, right, the the advantage of the MCP is that there’s someone else smarter who’s saying these are the skills again than me, right? Someone smarter than me that’s saying, look, these are the skills that I think you would want to use or need. I’m going to give you those skills and then you’ll be able to utilize them like, hey, , call down this report, call down the semantic model. , there is some information that’s supplied there that me as the user like I may know I want to make a measure But I may not know exactly how TIMDLE works and the definition of it or how to get that

7:35 Started. So I think in both regards there’s a little bit of knowledge required to use either the MCP or to your point Tommy you’re still providing that knowledge but it’s just being handled inside the instructions. Yeah. For the large language model to then go do what it needs to do. Right. I would I would agree like I think I understand what your point is there. Like yes you’re probably right. you get set up faster. But my counterpoint to this one would be like I need to understand enough about the instructions to be able to create them or find them

8:07 Or put them somewhere that I could use the IDE instead of the MCP server. Does that make sense? Yeah. The only thing I would counteract is I can literally feed a documentation as a PDF or from the website of Tindle and just say create DAX measures. In so many words, it would work. But to your point, I will give you the the point that also too I have to have the experience in VS Code and know how the project works too. Correct. And that’s that’s a difference compared to a chatbot. But where I’m going to where I’m going to agree with you here

8:38 Is if you’re developing things in PowerBI desktop already, making measures, making model changes, like you should already understand or be able to build something in desktop and see what the PBIR format outputs or produces, right? You’re probably going to need to know about this stuff anyways. So if you are an actual developer, an avid developer, I don’t think the jump between making instructions, to your point, Tommy, making instructions and then using the MCP, I don’t think there’s a huge jump in the gaps there. And this is where I feel like AI right now is changing things so rapidly for us

9:11 Because three months ago, everything was about MCP servers. Every everything was trying to get an MCP, everything everywhere. Now we’re learning that, okay, well MCPs are good. They definitely add their purpose, but instructions are potentially just as powerful, if not more powerful than the MCP stuff. So, how , how did the MCPS become like a flash in the pan? Is it is it this like quick hit of it works? I will say again, this is something I was telling me, I’m not sure if I understand where all this fits yet. This is me just unpacking the idea that you’re

9:42 Giving here is there’s also something to say for like the MCPS can handle like authorization of the user. There’s a little bit of extra like code bits there that I think makes this a bit easier as well. Not to say you couldn’t do this with instructions either like that’s potentially also very true too. You could probably get there. But if I think about like the MCP server, let’s say let’s let’s not say we’re using PowerBI desktop locally. Let’s instead say we’re going to be talking to an API in powerbi.com and I’m going to use an MPC server to talk to the APIs directly.

10:14 There’s how do you authenticate? Where does the instructions for that go? Does that is that something that the MCP makes easier for us versus having to write instructions? So that’s those are there’s like little edge cases here that I’m thinking through that are like that’s where that I would agree with you 100%. Anything service based MC MCPs really shine because then I can have a conversation. It’s like hey Claude I’m trying to take a look at all the notebooks. Can you tell me how many notebooks I have and which ones have to deal with X Y and Z? I don’t have to do that search. I can have that conversation around it. Anything

10:46 Locally, I’m going to do instructions. Anything in the service, I would use an MCP at this point. So, interesting. I think this is going to continue to shift. I think we’re we’re we’re building on top of sand at this point and the sand keeps shifting very quickly on us here. , this is one area that I feel like in AI I’ve just been feeling very behind because as soon as I learn something or try to get my head around running or building on top of something already something new has come out and it’s already sh the goalposts keep moving further and further into the future and the AIs are getting much more capable. Like every time I’m looking at

11:17 Another AI, I’m looking over here at Chat GPT. It’s doing amazing things. Oh, look at this. , Gemini shows up and is extremely efficient. Oh, look at this. something from GR just appeared and it’s super efficient. I just saw a a a graph the other day. I don’t know if I bookmarked it or not, but I was just looking at the number of tokens you can send tokens and then efficiency based on those tokens by model. And so all the normal models are following this normal trend. And there was one model that was incredibly efficient but had a lot less token usage for whatever it was

11:49 Doing. So I don’t know I can’t I wish I could recall the graph. It’s it’s probably somewhere on Twitter at this point. , but I don’t know where that was. But I I think the next, remember how we talked a lot about this Tommy about build it, make it run, and then optimize it, right? Phase two, right? I feel like we’re starting to enter the stage of AIS to starting to be optimized. I don’t know about that. I I think we’re I think we’re getting I we’re seeing it’s not going to be fully optimized. I think we’re starting to see like again to the point of the MCP,

12:21 Right? We can use MCPS. These are great things. They had they had a lot of functionality but at the expense of a lot more tokens. Well, now we’re saying, well, maybe that’s not the best way because we’re optimizing on the wrong thing. We’re optimizing on ease of use as opposed to optimizing on spend and cost. So now, , the price of those tokens, the price of using MCPS has gone up because they’re more expensive. And now we’re saying, well, maybe there’s a better way. Maybe we could just throw this stuff to your point in instructions, bring down my cost, and get the same result. And it it’s funny though, too, you mentioned like trying to keep up on things. My original method of like

12:54 Learning how to learn when I’ve been doing PowerBI RSS feeds follow the right people on on the social medias was really great. I can save and feed that information. Y but with the introduction of especially the in our world of AI now I cannot tell you how many times I would like check my phone go come on like a new model came out like came out with Opus 4.5. what does that mean? Should you use that? And we never had this many decisions to make at one time. Now again I’m still dealing with the individual. I would actually agree with you. So funny you say about where we’re heading and I think if you

13:26 Work in an organization too you’ve probably heard someone say to you we need AI and you’re probably next your response if you’re a logical person goes for what we can use AI for a lot of things where’s your problem and like no we just need AI and we’re still at that phase in the sense of talking actually about yes may maybe 95% of organizations who are trying to adopt AI right now according to Forester are failing. I I don’t disagree with that statement. What what I’m

14:00 I think it’s really limiting how do I say this one? I’m going to unpack another thought here. There is so many startups and there’s so many variety of ideas of happening right now and I I don’t know I wasn’t I was around when the internet boom occurred like in the ‘9s but I wasn’t like adamantly involved in the tech in the space to figure out like what was going on there like so the internet changed fundamentally how we do things and it’s has forever changed since we’ve how how we’ve been working and building on things I think AI is having a similar effect in the same way it’s it’s this

14:34 But then now I’m a bit aware of like what’s happening in the space. I’m a bit more in tune with it. So to your point, Tommy, I I agree with you. AI is probably not being adopted very well in the organizations, but I feel like there’s too many different varieties of tools. And what you’re going to start finding is the tools that are going to win are going to have a certain design, a UI, a pattern. There’s there’s something something is going to start consolidating or into like a thing that everyone will want to use. And what will happen is all

15:06 These like peripheral ideas of like super creative very neat but didn’t quite have the right functionality or was too expensive to run. All those ideas are going to like fall away and what we’re going to get is this distilled effect of here’s how we use AIS. And for me right now I don’t feel like Microsoft is building AI in a way that makes it easy for me to use it. I’m still really confused about like all the different AIs that Microsoft’s producing and how to incorporate that into my standard workflows.

15:37 Yeah. And I think the I would agree disagree a little. The biggest point is I think for the individual AI is pretty pretty advanced. But for an organization to adopt it, that’s where the the problems lie. Let’s test your how old you are really within the internet. By the way, you you got me curious. What was your AOL name? AOL. , I did have an AOL instant messenger had aim. Instant message. Aim. That’s what I meant. That’s what I meant. Yeah, AOL. I did have an AOL instant messenger and I don’t remember what it was.

16:09 I I I don’t I don’t know. Honestly, I don’t remember. It’s been so long ago I don’t even remember the name of it. What was your AOL instant messenger name? Mine was Fighting Vatican 24 because like the Fighting Irish, , college team thing. Vatican Catholic. So, it made sense thing. So, that was it. I think I got a Yankee one too, but I don’t remember the whole It probably had something to do with Derek Jeter. , but Jeter Yankees. Jeter Yankee something. Yeah. Yeah. Yankees for life maybe something like that. [laughter] Well, that one still holds true. Yeah. Yeah.

16:40 Well, I would say appropriate. Yeah. So, we’re I do know I definitely had a AOL Instant Messenger name or or title. I just can’t remember what it was. It escapes me at this point. That was And that was incredible when that came out like wait, I can message everyone now. Yeah. anywhere in the world. This is crazy. Yeah. Awesome. Well, that being said, I think that’s enough for our just information or your beat from the street here. MCP versus IDE. I think there’s something there, Tommy. we’ll have to watch this space in the future to see if this continues to evolve.

17:11 I think MCPS have a good place. I’m not sure if they’re the most efficient way of using your models and we’ll see what the technology changes into. all right, moving on to our main topic today. Let’s talk about Fabric IQ. this is a blog announcement that was introducing Fabric IQ to the semantic foundation for the enterprise. So this is the article from Microsoft. It’ll be in the description of this video as well. So if you want to go read the full article from Microsoft, it was published on November 19th and starts talking about the introduction of fabric IQ.

17:45 And really the message here is going from unified data to unified intelligence. Intelligence is this language they’re talking about here. One of the key notes they’re talking about is this thing called ontology, which the fabric IQ is this concept. Ontology is the program or the item in the workspace that you can actually use that builds the the concepts of your business, the people, the processes, the systems, rules, and actions. All of that became becomes part of the ontology piece. Let’s pause there. What do you think,

18:16 Tommy? We’re probably gonna have to give a few definitions here of ontology, I think, for the next few months, just for people who are listening. But this is something we’ve been actually devoting December to, Mike. We’ve already talked about the idea of data governance and and the age of fabric IQ. We’re talking about basically building up ontology from either ground up or the bottom. But let’s start again with probably the best way with a definition on what ontology is. And we’re going to use we’re not going to use the philosophy the philosophical one. We’re going to use

18:48 One actually generated in computer science and information science. And it’s a formal explicit specification of something that’s a shared concept. So something that’s formal, it’s machine readable or structured. It’s explicit. It’s clearly defined based on the concepts and its constraints. It’s agreed upon by a community or an organization. the definitions, the categor categorization, and it can be con conceptualized an abstract model of how something works in the real world. And this is where I’m really excited, Mike, as we dive into today’s

19:22 Topic and we really talk about the friction is this came from a conversation you and I are having offline too. We’re just talking about what’s this going to affect. And you realize though too and for me the biggest point when I think about fabric IQ is this cannot be created in a box. This can’t be created siloed by the organization. I don’t think any BI team Mike is going to go and work on this for three months. Go hey guess what we developed the ontology for the entire organization. That’s just not possible. And go ahead and try it. We’ll see if it’s actually going to get adopted by

19:54 People. Because we talked about the politics of an organization, the culture, the people, all these things. When you think about the definition we just said about what ontology is, and this is what fabric IQ is built on, you are defining for the organization, not just their terms around data, but their terms around dare I say everything. What is the process? What’s a member? , what are our categories? What are the sales regions? And a lot of these change rapidly. They are ever evolving. It’s not necessarily like I’ve defined

20:28 What water is and that’s not going to change. So there are some definitions that people hold on to and there’s a power struggle there and there are other ones that are just very fluid. We’re well northeast region used to include Connecticut but now it’s nor it’s now we’re in our New England like things like that that we are going to go into this idea. And for us Mike we have to introduce not just the technology to the organization. We’re not going to just say we’re doing ontology for you. We’re going to make things easier. But there has to be this introduction of

21:01 Workshops. There there has to be these discussions that have to happen. One to get the buy in of this is what we’re going to do. And I think that’s going to be a friction point. And then two then if we do have buyin getting everyone to have a shared and agreed upon definitions of concepts and the constraints in what we work on. And then the third friction I see is ownership. So I I want to I like what you’re saying here on these things and I was taking some notes here as I was talking

21:33 About these ontologies. So let’s talk about what fabric IQ is doing. I think really the fabric IQ piece of this hangs on top of this this concept of the ontology. Ontology is basically it if I had if I had to use an analogy or a mental model here it feels like a graph database. These are things that we’re building. There are reports, there are semantic models and the examples that they give in this documentation is a lot about like an airport, right? This airport is JFK a JFK airport and then from there you have okay what all flights are

22:10 Going in to that airport. Okay well on a flight there are people what people are on that in that plane that are going to the airport what pilots are in the plane where could those pilots be moved to a second plane? So there’s the ontology basically is this idea of like a graph that shows relationships between different items that you have. So I think that’s that’s one of the key concepts here where I think this going back to our main topic here Tommy I think this is what you’re talking about is where does the friction come into play. I think when we talk about simple things like an airline and then we can

22:43 Easily align on airports, planes, pilots like these are things that are very wellnown. I think most people have common definitions of those things. But there’s other items that we’re talking about here that aren’t going to be as clearly defined, right? When we talk about the business logic of how do we compare sum of sales, what is a customer, how do we do advertising, there’s a lot of terms and definitions of things that I think a particular business unit uses, but not necessarily all business units use. And to your point, Tommy, I I want to hang on this idea that

23:17 Fabric IQ is now requiring or businesses now have the opportunity, let me say it that way. Businesses now have the opportunity to write down what it means. What do things mean to you? And so the tricky part here becomes, well, what happens when we disagree? How do we handle the disagreements? What does that look like? Are we allowed to have the same definition or similar def defined terms for customer between supply chain and marketing? What does that mean? The there there’s two

23:51 Different kinds of potential definitions for customer. And so I think this is where I see the potential friction coming into place here is we need to think through as administrators are we talking about our whole business as the ontology or we talking about department level ontologies and do our department level ontologies need to roll up to the broader business who makes those decisions should every business run the same way or should we have different versions of businesses running however they feel like they need to run so that let me just pause right there your reaction Tommy

24:22 I’m really intrigued that you use the word admins making some of these decisions initially or not necessarily all everything down but that you brought admins into this already to me because I wouldn’t I’m looking at this Mike and I’m realizing here too aren’t making the decision to be clear trying to say that but turning it on or whatever the case may be well an admin will turn on why don’t you elaborate more on the why you brought admins into this now I said the word admin but I’m thinking maybe more of the idea of central BI let me let me rephrase Let me pull back the

24:55 Term admin. Let me call central some to your point Tommy ownership of this I think is important. Who owns the definition of the ontology? Do different departments own parts of the ontology and others do not? So one example that I have that I’ve used a lot over time was when we were in marketing and we were defining products to sell for businesses. The color and the dimensions of those products was determined by engineering because they built the product. Then the pricing of the item came from finance because they

25:27 Were saying how much it cost to make the part and sell the part. Like that was their responsibility. And then there was a whole another arm of like marketing team that was like, “Okay, well, how do we define the size? Like based on the dimensions of this, here’s how we’re going to categorize this item. like is it is it does it fit a certain catery category and and that category defines what that product fits in right so there’s all these different teams that own parts of data and how do we hold them responsible and when like the central BI team

26:00 They’re necess they’re not necessarily in charge of the ownership of the data but they’re the ones that need to be bringing these teams together and saying okay someone needs to take ownership of these properties these data pieces You need to have responsibility to this. You’re going to define it. And when do we stop other work to make sure that those things are defined correctly? Who’s who’s governing that? Who who ensures that there’s quality in that data? Right? This is where I think a lot of the friction I see coming from Fabric comes from is someone needs to define it and someone

26:33 Needs to own and manage it. What is that going to look like? I don’t know yet. I agree with everything you said, but I’m going to challenge you that I think you’re already to me three steps ahead of where we need to be when we’re talking about friction possibly. Oh, absolutely. So, here’s the thing with to your point about the definitions, marketing needed to know what they need to do, engineering, etc. Well, they knew if we didn’t do this properly, we’re not getting better reporting or accurate reporting. But ontology and fabric IQ is not for just better reporting, right? So ility,

27:06 It’s finding things. It’s it’s going to be for AI. Yeah. But it’s not. Yeah. Right. But it’s not for like my PowerBI reports. This doesn’t have a great marginal enhancement to my PowerBI reports. This is meant for getting we talk about more the soft data too, but for optimizing AI. That’s why Microsoft said to this intelligent, , we’re we’re calling fabric a intelligence platform. Now, that’s not by mistake. So, for me, if I I’m starting at the point, Mike, the first

27:38 Level of friction and the I will agree we’re definitely going to have the friction around the definitions. Everything you said is 100% accurate. But the first one is well I and I I foresee is well, why are we going to do this? Right? Just because this is a new feature in fabric, does this help our reporting? Especially Especially Mike too if the what is now known as the BI team is saying this we’re going to start this entire groundup conceptual definitions across all the organizations someone’s going to ask for reporting

28:10 What’s going to do for reporting it’s like oh no this is for AI this is for the intelligence side people that’s to me the first level of friction is buyin buyin is going to be ridiculous here and unfortunately too the concept of what the BI team is is not set up right now for this in terms of what organizations consider what BI does. They think business intelligence if they even think that at all and they think reporting and data. We’re now introducing this whole thing where you need leadership or people trusting that leadership in that space to understand

28:43 The intelligence side that we’re they all are also the ones to lead you to the promised land for AI because we’re not doing fabric IQ for PowerBI. Heck, even fabric in a sense, , in for my notebooks. That’s not the point here. And to me, this is going to be like that shift too where we’re like, hey, we’re going to do PowerBI, not Excel. Why? Excel works. , what’s this PowerBI thing? Are you a PowerBI person? What? I don’t even know what BI stands for. And to me, I think this is going to be the first step we’re going to be in.

29:16 I I’m curious to see if you do agree with that or you think that’s going to be easier for organizations than I’m saying. Am I am I casting doom and gloom here or Okay, so I think I think that there’s a there’s a push here to move things forward in some direct in some directions away from not away from Excel entirely, but the idea is like what can Excel do really well? Excel does really good job of like taking data in it does a lot of data transformations. You can get things done in there efficiently. Where do we look at this in lie of what does fabric bring to the table? Right?

29:48 So other things when we look at the documentation here we’re talking about like ontology model management. We’re talking about connecting connect live data to the enterprise data. So this is linking reports together and say okay here’s a handful of reports that relate to these flights. Again going back to the analogy that Microsoft is using here flights and airlines and things right so now we can have real reporting attached to directly into the ontology that describes we’re talking about flights. Here’s our flight data. Here’s flight information. , another thing that they’re trying to talk attach into here is , actionbased

30:22 Elements. So, they’re talking about , automating actions on top of these ontologies. And that’s another part of this that I think is really interesting here as well is when you start looking at, okay, when every 5 minutes I want you to go check for flights that are delayed more than 20 minutes or something along that line is I think the analogy they’re giving here. And so they’re actually showing you an alert, Jeff Gif, where you’re saying, “Okay, based on this information that’s coming into the system, we want to be able to every so often check this data, look for this information, and when that occurs, the action is go send a a a meeting, an

30:56 Invite, or go send a team’s message out to this channel, right? It’s that stuff. So now when we’re starting to talk like all the things that fabric offers, these are things that Excel is not doing, right? So that’s the push towards these more modern systems which is Excel is good at loading that one-time chunk of data. I’m not building an alert system inside Excel. I’m not sure building ontologies or related data pieces. One Excel document should not be a database. there’s there’s a lot more data. We could be talking

31:28 Hundreds of millions or gigabytes or terabytes of data that need to be supported by this whole data platform. And us as users of the system, we need to be identifying, okay, when do we need to pay attention, right? And I think that’s a part of this that is instead of to your point, Tommy, I I don’t know if businesses are saying, let’s drop Excel and let’s only go over to fabric. I think more of the idea is what is fabric offering feature set wise that we can use leverage that Excel or other, , what we’ve been doing in the business hasn’t been able to leverage so far. How is this going to extend our capabilities?

32:00 Right? And and that’s part of my argument was I’m talking 10 years ago, Mike, when we had to make everyone’s mind shift to why PowerBI was valuable, more valuable than Excel. Question off of that off of that you you’re making me think, Mike, my first thought is I feel like we’re going to need here with Fabric IQ a different way rather than this being a project in a sense of just something we’re going to start working on. It’s an initiative for the organization. So to me, in order for this to work, I feel like we need a separate executive

32:32 Sponsor around the organization. It could be the same person as what’s already happening in fabric and PowerBI, but there has to be someone in the sea level, someone close to that to get this done correctly. Because to your point, you already said when we’re in the weeds, well, if someone has the same definition like sales, customer, order, project, who wins the definition, right? So there there’s there’s governance here. There’s a lot of things once you get into the weeds. So I’m thinking of the buildup here. So we don’t one we can

33:04 Actually start running with this. So I see Mike first off that I need a statement on what we’re going to do with Fabric IQ, why it’s going to be valuable to the business. And then two, I need a sponsor. I need someone to sponsor that project on a high level because this to it’s a change in platform. And I think maybe that’s what Microsoft’s trying to do here. And because all the data now too, we’re also not looking just at Mike and this is important. We’re not just looking at an onto a a new entity in

33:36 Power in X fabric. We’re talking too about everything we do around our data is built towards the intelligence of our organization. And this is a complete shift to reporting and dashboards and KPIs. So to me, this is really how do we build our team up? We have to rethink everything. Okay. So two reactions I have to your comments there. One is I don’t really want to get more executive sponsors in place here. I I think it’s just the one

34:08 Because I think as soon as you start having multiple executive sponsors, we start getting fighting in amongst executive sponsorship level like what’s really important? Who’s doing the who’s doing the value here? So again that so that’s one one point. I I don’t really agree like I think it’s definitely this is the responsibility of the executive sponsor to understand what this capability is right does it add value for us I think there’s multiple ways I’m looking at this ontology piece to start adding value to this which is as a as an organization we can start writing down on paper

34:41 What matters to us right I think this is the first place that I’m kind this is probably done in other places it might be in a bunch of word documents it might be on shareepoint someone’s got the idea of what the business is doing. What this is doing, I think this is bridging a gap between like that mental model of how your business runs and actual real data that supports all of this that supports all these other data pieces that we’re giving. So you think about it, think about it this way. , , our business is going to run on a series of like real-time data,

35:14 Incremental loading data, a lot of transactional systems, and then we have products we sell, things that we do, services we offer. Like all these things are like how businesses run, right? True. Going back to what Microsoft is really good at. Microsoft is really good at standing back and saying, let’s just observe how businesses work and build products that serve those businesses at a at a holistic level. So when I look at this going, oh this is interesting. When we look at the whole business, I think what ontology is doing is it’s saying look we have potentially silos of business data real time batch loading

35:51 All these tables of information or semistructured or unstructured data all exist somewhere. What’s happening now is okay now that we have all these things in our environment how do we start linking the things together? And I think to your point, Tommy, this is a a preparation step for AI, right? The AI can’t read all of our tables. The AI can’t understand how all the things relate to our business. So, this is a a way of the business stepping into this and saying, “Here are the reports that relate to flights. Here’s the reports that relate to the airport services. Here’s the the information that relates to pilots

36:24 Information.” All those things can now be brought together. Previously, they were disassociated. They were not connected. that now we’re adding relationships in a way the computer can understand and this is a stepping stone to giving AI more context around our business if the lemon is worth the squeeze and I think the biggest thing I agree with what you’re saying from the technology side but to me and we’ll move off the buying side because I think we’re we’re we’re doing well with that but I I’m just going to go back and say people are only going to be willing to dedicate

36:56 That time and resources if they see the value on doing that and unless you again you can say top right and this is a big part here all the stuff yeah absolutely with the technology building these these ontology models but if you’re just saying this is what we’re doing now and again you’re not just bringing in BI into it has to be everyone in the organization dedicated to this and the effort well again those people have to see the value the the future value that I don’t even know think we know what it is so I I think we we’ve touched on that quite a

37:28 Bit Mike I want to introduce to you I three places for me I see friction being a very where the most tender parts of the organization are going to be around where we’re going to see friction and feel free to add or modify but I think it’s going to be the cultural side of an organization the political side and it’s going to be the the roles the responsibility side I I’ll start with when we’re talking about the cultural side here well we’re talking about people’s already current habitual ways of dealing with data and technology in the organization whether they’re doing

38:00 Intelligence or not. I either love my PowerBI, I may be doing self-service and I have my definitions already intact. We have our process, we have our teams, we have our skill, we have our already center of excellence. Now we have to change that the political side. I part of this is better left unsaid than said but we all know what this is. People love owning their certain things. There’s certain power to it. Their role don’t go p don’t go over me. marketing owns this and you dare not step on marketing shoes in some organizations. But that’s but to

38:33 Your point there though there Tommy around the political side there is there is some political headwinds I think you have to fight here as well you have to be aware of certain departments are going to be more this is our data you don’t touch it we’re going to own this part right we don’t share this out yes I that makes total sense but if we’re talking about like this ontology or the scoping of the whole business piece of this this is where I think the executive leadership has to step in has to and that’s and that’s where we’re having like okay we need to bring these two teams together and again I’m thinking more of this to your point around culturally, right? Departments

39:05 Work in a vacuum. They do what they got to do to get their job done. They just deliver based on what their tasks are to them. Right. Right. So the political side of things are becomes very fiery I guess when right I’m trying to get things from you where I’m stepping on your toes or I’m taking over as something that you traditionally owned or I’m I’m defining things differently than you are because my reward is based on how I define it and your reward is based on how you define it. this is one of the one of the key pieces I think is really relevant here is when you start talking

39:37 About like bonuses for like sales teams and things. that’s that is a very important process to define people and because that’s how you pay your people and so that’s how you keep your talent. If you’re going to have a good reward system you need to clearly define it and you need to have everyone understand like what what is the expectation because if I’m marching towards a goal or I want a certain amount of bonuses or revenue or whatever those things are everyone in the whole organization is going to align to that particular item. And I think this is where I start seeing to your point Tommy that the political

40:08 Side where friction starts in place becoming in place here is when two different teams have different visions of their goals or their objectives right one team may be looking to reduce the amount of scrap while the other team is looking to increase customer satisfa customer satisfaction. Well, if our process is not in control and we’re producing more scrap, but we can get things out the door faster to make the customer happy, like now we’ve got to figure out how to make those teams work together so that we’re we’re able to like satisfy both things. I can’t tell you the number of times,

40:38 Oh yeah, sales teams would promise all these wonderful things to the business to to customers and then come back and say, “Okay, we promise all these things.” And and the the internal engineering or ops team or distribution teams are like, “Dude, you can’t promise that. There’s no way this is going to happen.” like that you’re making promises that you didn’t talk to us about first. You should at least review it with us first so we can at least have a conversation. So to your point, Tommy, there’s a lot of conversations in that political space that need to be happening. And I think that’s where executive leadership needs to step in and say, “Look, you guys need to get in a room. You need to hash us out. We both

41:11 We need to be aligned on the the direction of where we’re going so that way everyone can march the same direction.” No, I see this is where I’m I’m going to expand on what you said at the end because the idea the phrase of that get in a room, figure it out sounds nice. That never works in in a in a normal culture unfortunately because people are holding on to I I this is where I’m going to back up a little. I Mike, this is where I think the introduction of I said this on a previous podcast, but data governance has never been more important than it is

41:43 Today because of fabric IQ. Unless if you’re just bringing people in a room and the sponsor says figure it out. I’m not saying that’s all you’re saying. But it’s saying because if you have people saying they have sales finance operations have the same, , definition or they have a different definition of customer and they have to figure out what that means. They’re just going to figure out a room. It can’t just be the leaders of those rooms because they’re going to fight for their people. They’re going to fight for their careers. They’re going to fight for their property. And unless there’s honestly Mike and something established

42:16 In the organization around governance here, , you have we’re dealing with people. We’re not dealing with technology at this point. And people are are going to be selfish at times. This is just the runs of the political side where they have to have the concept of what we’re trying to do. They have to have the concept of the ownership here in what governance is before we can get to that point. That would be my argument. I I see your point there, Tommy, but I I

42:48 Don’t see like I can’t I can’t justify to myself to say like we can’t figure things out. I I think this is I to say to say like there’s political headwinds against defining what certain terms or things are like I think at the end of the day when we step back and look at the business at a broader perspective like is this is our activities contributing to the bottom line satisfaction of our customers and making us more money right that’s that’s the business objectives right businesses wouldn’t be alive if they weren’t making increasing the revenue or being able to

43:21 Sell a product that people like and use so without that right if we if we take that away you’re right there’s we’re just going to have infighting against data governance who owns what who doesn’t own what and so I agree but at some point you’re going to have to come to alignment now it may not be like a single conversation and putting one room and just say figure it out it’s going to have to be multiple conversations it’s going to have to be putting responsibility on different teams and also and to this point as well when we’re talking master data quality that’s that’s something that has to be defined

43:54 Centrally by someone and other teams have to admit that we’re going to absorb and just use this. This is what we’re going to do. , I if I go back to the airport analogy, I can’t define airport names however I want to based on what I feel at the time, right? It has to be someone’s defining these things and that’s what they are. So, I I think at some level like yes, you’re I understand what you’re saying, Tommy. There’s going to be friction there, but I’m also of the opinion that it can be worked out. we can figure out how to govern these things, how to find common ground across these different

44:25 Departments. And I think this is the reason why OKRs exist. OKRs exist. So we have this defining value of the company and that trickles down to our teams. And so you’re right, but that’s why we have hierarchy and leadership levels inside organizations because someone has to step into this meeting and say, “Look, we have these two departments that are bickering about data points or who owns what things. We need to figure out how to work together. And so we’ve got to play nice. So I see this a lot in times with central IT teams and the business intelligence teams. There’s a

44:59 Lot of friction there. Central IT is running the systems that make the business make money and then the business intelligence teams come in and say we need access to those real tables and central IT says no, you can’t have access to our production environment. That’s we’re not going to give you that. So there has to be negotiation there that says okay the business is going to have to bend on some of their requirements. The IT is going to have to make some new requirements to get the data in the hands of the business. So we can’t assume like we have to be able to negotiate here at some point. So that’s really what I’m pushing for.

45:32 No. And I I would I would I agree with that and I do. But Mike, I know I’ve been in the rooms and I know you have too. I’ve been in rooms where product and marketing were trying to figure out the category information and and after the meeting it was worse than when we started without because of low maturity and low governance at the organization. You’re dealing with the people whose in a sense life and their team was on the line. So I agree with you in theory we got to negotiate that but I think we have to be conscious too. We’re dealing with culture here and you have to keep morale up and if you just get people in

46:05 A room who are headstrong and again I know I’m I’m bringing in things that are not necessarily measured here but if you work you work in a company there are people that are easy to work with and harder to work with and some of them are leadership and it doesn’t work. So actually I’m going to take what you said and I’m going to ask you a question because I think this is a it’s a good question what you’re bringing up. Does Fabric IQ require some pre prerequisites around the maturity the data maturity of a company? I would say so I’m going to do they need to be at a certain level

46:38 You need to be at a certain level. No I don’t think you need to be at a certain level. I think what the ontology will expose is the level you’re at. Right. , , so if you put in on better or worse, for better or worse, I think you I think you put in the ontology or you start, , defining what your business looks like, right? If you show up to this and there’s no fighting, you have a pretty healthy culture around defining what things mean what stuff. And and again, I’m going back to this idea of like there’s a there’s already a mental model of how your business works. Let’s let’s

47:10 Take like I was working with a company in in the past years, Uline. Uline? Have you have you heard of [clears throat] the the business Uline? It’s like it’s like the Amazon for office materials. Okay. Like paper, printer stuff, ink, , you need a trash bin. You need like take a look at that desks or things. Anything that you need to buy, you could go to Uline. There’s a catalog that shows up and there’s a whole bunch of things there, right? So if I think about their business, right? We can clearly say we have warehouses, we have items in the warehouse. We can all agree upon those

47:43 Things. Like you could have a healthy culture around identifying what those pieces are. But if we really if we’re if we’re going to negotiate and debate around who’s the customer, who’s not the customer, if there’s a lot of infighting around what Uline does as a business, then we don’t have a healthy culture in the business that can understand what we’re trying to accomplish. And so if we if we think about like a business a shipping business or a distribution business in in that way I think the the this is where I think why the

48:16 Topic is the way we are today right fabric IQ will expose to you your business how mature you are in defining what makes your business run what makes your business valuable and I think if you have a lot of infighting about these definitions and that just tells me there’s not enough culture around really aligning and getting on the same page on things. I’m going to sit on this and I’m I’m going to let this bake for another episode too because my initial reaction, Mike, is if I were to go into an

48:48 Organization today and everyone was building their own reports and it was all that whole hectic wild west. Yep. I’m not going to recommend in the slightest that we start fabric IQ or start these models. We’re a mess. if we can’t even get our reporting right. Maybe not for the whole business, but maybe for a department. Sure. Maybe maybe for that, but automatically there’s the red flag. So, I’m I’m going to sit with what you said. My initial reaction is still there is some predetermined or maturity minimum requirements that you want to see your

49:22 Organization before you start the process. And I I think we’ve hit a lot of the friction, Mike. Mike, people like to say that we’re doom and gloom that we’re all very concerned and negative, but I think if you if you have any more things, I think it would be good too to talk about what do you think we can do to avoid or minimize the friction even though and I this is all being said, this is all new to all of us. Yeah. All right. So there’s no magic recipe here, but I think it would be good too to talk about we’re already identifying or really trying to bring to light where

49:57 You’re probably going to experience this. But I think yeah, I think it would be healthy if we went through what are some of the things that based on our own experience introducing technology and information where we can minimize or alleviate that. So I think I like this point. So let’s talk let’s stop talking about like where we think friction will occur, right? Talk about yeah all the bad things. Let’s talk let’s not talk about anymore like cultural political challenges that we see. So these are things that we think you’re going to hear when you start doing these things. So, I’m I’m going to maybe pull back

50:30 Here a little bit and say regardless of this item. Let’s let’s , report semantic model, an ontology. In all these situations, if you put junk into it, you’re going to get junk out of it. So, I’m going to go back to semantic models because that’s what we know. We’ve been building these for years. If we have a semantic model and we put bad data or non-consistent or inconsistent data into the semantic model, the reporting that comes out of it is going to be okay at best. , I have a lot of relationships between tables that

51:03 Have a lot of nulls between them. Well, then whenever I aggregate something across a dimension to a fact table, I get a whole bunch of blank values. Well, that doesn’t really tell me anything about the data. There’s not a lot of value being added there. We’re missing things. So, what happens in those situations? Well, you go back up to the upstream tables. to go figure out what was wrong with them. , does this value make sense? Should it have this many nulls in it or are we missing something? Do we need to fix something in our process to enrich that data so the reporting becomes valuable? So this whole concept of junk in will produce a junk out result. The AI is not going to

51:35 Get better at your organization. If you put a bunch of junk relationships and junk data into the ontology, it won’t actually be able to find the relationships you need and you’re going to find minimal to no use or value from the agents you’re you’re attaching to this. So I think at the end of the day here like you need to find an appropriate use case and and this could be like experiment inside a department, experiment inside a business unit. I don’t think you have to go.

52:08 Let me just say it this way. When I read this article, the article seems to feel like we’re talking like the entire organization must jump into this ontology thing like like Fabric IQ is for like the whole business. I get that. But even in the marketing department, you can align on some things that are happening in there and provide some wins and some successes. And how do you use this tool to help you provide discoverability? And this is a very unknown space I think for BI developers in general, Tommy. And

52:40 So I think a lot of this is just we need to find appropriate use cases where to apply this and then the community at large needs to say here I’ve used this tool. Here’s how we’re signing finding some value. And that’s one of the reasons why in this article you see like there’s multiple callouts here from customers. NMAX power is excited to leverage fabric ontology to unify transmissions and distribution of grid data and there’s a video for it. Right? So even inside Microsoft’s examples here they’re they are very

53:14 Clearly calling out customer wins and how they’re trying to leverage ontology to help them manage and organize their business better. So I think that’s where this is going right. I think this is a stepping stone into getting a better handle on what your business is doing, what data you have across your organization and and becoming more structured with that load and and ingestion of data. I love that. I I absolutely love that. And I think it’s that important aspect here, Mike, of like what are we trying

53:47 To do? And to me, that’s where I’m looking at this where let’s take a step back. I think everyone does because why are we doing fabric IQ to just do it? So we have ontology models and we’re not even does everyone remember who’s listening what the definition of ontology is? Can they [laughter] who’s listening to the podcast? Yeah, there’s your pop quiz. But truly though, Mike, I think we need to take a step back and write down and begin that initiative with your business on what are we trying to do here? And to me, I’m

54:20 Going to go back to the implementation planning a bit because I would want I would love to see this with an initiative a tactic a tactical a strategic point of view where we’re going to do in the long term this idea and you don’t even call it fabric IQ. I don’t care if you’re saying your business in an effort to alleviate and to bring better information to the business better automation and bring this introduction of AI to help teams you be empowered whatever you want to say not just with their data but their

54:52 Information we’re going to begin this initiative this project and it’s a long-term project unlike a notebook and I think this is a concept everyone else needs to understand is unlike a notebook where I can I can optimize my data today my by myself like I get off the call before lunch have a notebook doesn’t and no one needs to know okay that does not exist with fabric IQ what you’re if what we’re trying to do so I think we need to start with the strategic side where we have to have with the business a contract or the buyin with what is fabric IQ but more

55:27 Importantly and really start with what are we going to solve where are our problems around information and yeah I think part of that is like beta, so to speak. Like, we’re going to start introducing some new things to everyone. But in order to get there, this is a long-term project to me, Mike. I I again, I it it fails me to figure out how long a company could implement this today, like three months, maybe if you already have the definitions. And again, you can have a working ontology model, but whether it’s used efficiently, this

56:02 Is all new to us. So I would start with this Mike with the strategic side getting the buy in and making sure that people understand what the efforts are going to be and then yeah if anyone wants to be part of a beta I think this is an earlier stage than a pilot to your I think you said just use case where we’re going to test things out great that’s what we need but the goal is not an ontology model at the end of the day and I think that’s going to be an important concept for people my goal that’s a that’s a that’s a means to an end. That’s not the end.

56:34 It’s it’s part of the tooling that I think you’re being producing, but it’s a means to an end. But I think this is also a stepping stone. Again, the AI space is changing so rapidly and quickly here. I think this is also a stepping stone. So, let let me go back to like the article again. I’m going to reference out a couple key points here is like, , where do we start? How do we start handling these conversations? I think you start small. I think you start figuring out where does this thing actually add value to your business? Do you have areas of your business that are confusing between customers, relationships, tables, datas, facts, measures, dimensions, like all these things that we’re talking about here?

57:07 There is also to some degree here. They they have this idea of reuse your semantic model investments. So the one of some of the key features here are new we you don’t need to generate your ontology from scratch. Use an existing PowerBI model to build the ontology. I have that for another episode. [laughter] Yeah. Well, I’m just saying. So there’s already you’re already so let me say it this way. You’re already building ontologies. They already exist, but they’re being represented to you as a semantic model that’s already defining

57:40 Relationships between customers, facts, calculations, metrics. These things are already occurring inside your organization. Whether you like it or not, you’re already doing it. Someone’s organizing information that can be used. That’s what we’re doing here. Fair statement. All the all the ontology is doing is tapping into maybe a broader scale. And this is one of the things that we talked about in a prior episode Tommy which was we had this idea of your let’s say you have some organizations have lots of these hundreds if not thousands of semantic models across your organization.

58:12 Do we have any single place where we can look at to say where are all the data tables coming from? Mhm. Like when we’re pulling customer data, how many sources of data are required to get to customer? What does that look like? Yeah. And so this is the concept of you go back to like standard data warehousing techniques. This is what we’re doing. We’re going back and saying here’s all this data that exists for the organization for us to make smart decisions. We need to pile it or pull it into the same space. We need to figure out where it’s coming from and have

58:43 Definitions for it. And so all this is doing again my my really my main point here is the ontology is just accelerating the need for us to have discussions around what makes our business tick because in order for us to get it out of our mental heads into something that an AI can use. Yeah. Now we can have an alignment between what the AI understands and what we know. And then we can start saying okay I can then talk to it in it with the same understanding that I have. That’s what this is doing. I think

59:16 I think in that one statement alone, you have three more episodes for us to thank [laughter] you for that. This is good. No, this is really good. And I really do, if you’re listening, please go to our mailbag because I’ve never I’ve probably said this before, so I apologize, but I feel I’ve never been more interested in what people think about this tool, if they’re excited, what their questions are, than this new feature. I think in the last at least in the last five years like even when fabric came out I I think I was dealing with my own questions but I really want

59:50 To know what people are are talking about here what the water what the water cooler conversations are with Fabric IQ. So I want to hear from everyone else because we have our a ton of questions already and I think that’s why Mike this is why I’m so curious what other people are thinking is because you and I are looking at this too and it’s it’s a new horizon. We found the new land. the it’s like finding America thing where it’s like, okay, I don’t know what that mountain is. We’re going to call it this. Why should we go that way? Nope. There’s a bear trap or whatever it is. And so it’s really new to us, too. It’s new for

60:23 Everyone. So, I’m curious for how other organizations are are going to tackle this or approach this. But dude, I love this. I think I think this idea of like the ontology piece to me so when I look at this concept right this is an opportunity for business to take leadership in what data you’re bringing in and having real conversations across multiple teams to define what makes your business tick and so I really do think that this feature is very useful and we already have lots of semantic models we have lots of really good

60:54 Information and knowledge appearing inside powerb.com and even fabric this is just a way for us to start reorganizing the information in a way that we can all that we can easily understand this one. I also think this is a really good example for when you bring on new employees or if you grow the team like here’s how our business operates these are the key things that we care about here’s how our OKRs and what makes us money trickles into the data and the decisions that we’re making. I think a lot AI is just love that. Yeah. this if we if we step back and just say look

61:27 Data warehousing business intelligence at the end of the day take away all the fancy terms it’s decisionm all we’re doing is this is data for decision-m if if we if we distill it down to the most fundamental thing and then what we’re looking at here I think for ontology is this is an example of us being able to organize information that helps us make decisions and so this is the evolution that we’re going through right now which is no longer do I feel like I need to build everything inside the Excel sheet. The the capabilities of PowerBI and Fabric,

62:00 We’re getting so many new rich tools in our in our hands at our disposal that we can now think about what we do to make decisions differently now. This this is changing how we make decisions. Dude, it’s great. Mike, I can’t wait to talk about this more. And by the way, remind me for future episodes for me to slouch in my chair. I feel like I’m much more calmer that way and I I think more rather than sitting up like this when I’m slouching back, I’m listening to you. I feel like I’m just jamming with good ideas here. Why do you think I lean back in my chair every single episode? [laughter]

62:32 I’m told I know I look a little back in my chair. I feel I’m like, “Oh, that’s a good topic.” Oh, yeah. That’s a session thing. I’m like in the middle of me talking, I’m thinking that’s great. [laughter] So this is but I think this is the this is the stuff that the community I think needs to grapple with, understand, and unpack. I think I think this is where we’re at right now. I think this is the the the place where we’re doing. It’s good stuff. I like where this feature is coming from. I really do hope that the community and the usage that Microsoft needs comes out of this ontology thing because one of

63:05 The things I’m I would be very disappointed about is if Microsoft gets close to the hitting the mark but not all the way. And to your point, Tommy, right, we don’t want this one to go by the way of metric sets. So, I I’m excited about the feature. I like the idea of what it’s doing. I do think the fact that it has a heavy tie-in with AI and Agentic things, I think that’s going to give it some lasting power because Microsoft right now is heavily pushing everything agentic or AI based or adding AI into your workflows, whatever that may be. And so I’m my mental model of like where AI can be used effectively is

63:40 Just every day it’s changing. We’re seeing new things come out all the time. and so I think the speed of development of AI into applications is just accelerating and I think it’s going to be hard for Microsoft to just keep up with the rate at which they need to continue incorporating AI into things. and so some things are going to get right. Yeah. Some things are going they’re going to dial it in. It’s going to be exactly spoton. other aspects are going to be way too difficult and they’re going to miss the mark. And so I think it’s going to be important for them to add or glob on to the things that work well and

64:14 Accelerate those as best they can. But for things that aren’t working well, they’re going to need to like sunset them quickly as long as we’re along for the ride. So I think so. Great. Yeah. Awesome. That being said, thank you very much for listening to the podcast. We hope you’ve enjoyed this conversation around fabric IQ, how it relates to your business and and just being open to having more candid conversations across your business teams, talking about what is important to your business and how you make decisions. I think this is really what this boils down to. Hopefully this added some value for some things to think about in your organization as well. And that being said, Tommy,

64:47 Where else can you find the podcast? You can find us on Apple, Spotify, or wherever you get your podcast. Make sure to subscribe and leave a rating. It helps us out a ton. And do you have any questions on our episode today or in general or or topic that you want us to talk about in a future episode? Head over to powerbi.tips/mpodcast. Leave your name and a great question. And finally, join us live every Tuesday and Thursday, 7:30 a.m. Central on all of PowerBI tips social media channels. Thank you all so much and we’ll see you next time.

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.