Helping Leaders Speak Data – Ep.541

A two-topic episode. First, Microsoft shipped official skills for authoring Power BI reports with an agent, which changes who can realistically build one. Then the main event: the persistent gap between what executives say they want and what a BI team can actually build, and the specific moves that close it.

News & Announcements

• Power BI report authoring skills — The report-authoring skills in Microsoft’s skills-for-fabric repo. As the hosts describe them, there are four: report design, report management, report builder for scoping, and report authoring itself. They let an agent write PBIR files, trigger a Desktop reload, capture screenshots, and iteratively refine report pages — the agent never talks to Desktop directly, it works on the saved project files.

• AI-Powered Power BI reporting: from design to deployment — Microsoft’s announcement post for the above. Mike’s read is that this is the piece that was missing: before these skills, an agent pointed at PBIP files could make edits but lacked the context to work competently on the report side.

Main Discussion

Topic: Getting executives and BI teams to speak the same language

Tommy sets the stage from both hosts’ analyst years: the job is to enable decisions through data, but leaders often run on instinct and long experience, and there’s no shared vocabulary for what’s possible versus what’s being asked for. The conversation is about building that vocabulary deliberately.

• “Increase revenue” is not a requirement. It’s the kind of goal that sounds actionable and gives you nothing to build against. The work is teasing out which actions the executive believes will get them there, because those are what your reporting should track.

• Ask what keeps them up at night. The useful conversation is not telling an executive how to do their job — it’s understanding what they see as pivotal, where they think the organization is deficient, and what they’ve decided to say no to.

• Data culture and data literacy are different problems. Literacy is speaking the same language. Culture is knowing where to find the right data, who owns what, and what to rely on. You need both, and they’re addressed differently.

• Tie every project to a stated goal. Tommy’s data culture matured at one company when quarterly and annual goals were defined and only aligned projects got worked. It converts “an executive asked for a report” into “how does this support the goal?” — which is a much easier conversation to have.

• This is OKRs, whether or not anyone calls them that. Objectives and key results at the leadership level, cascading into the metrics your team already produces. Executives are usually compensated against them, which makes their motivation legible.

• Executive sponsorship is the number one predictor of success. Mike relays this from Microsoft directly: their largest customers who skip the Fabric adoption roadmap and do their own thing are measurably less successful. An executive sponsor who genuinely understands the work is the single biggest factor.

• Communicate proactively and consistently. Meeting a C-level once achieves nothing. Sending progress every two weeks against what you discussed builds the buy-in that turns into budget. Consistency is what makes this stick — one person cannot do it in a vacuum.

Give semantic models an AI folder

The technical ask of the episode, addressed to Christian Wade by name: put a .AI folder in semantic models. Mike’s argument is that data agents are an abstraction bolted on top of a model, when the instructions, descriptions, and business context should live in the model as lightweight markdown. He also names the tension honestly — metadata stored with the model can be used anywhere, while a data agent requires Microsoft compute, so the two goals conflict.

• Skills should lean on scripts, not prose. Mike’s recurring efficiency point: every extra markdown file is tokens in context. A deterministic script that builds a visual beats a long explanation of the visual spec.

• This opens self-service to non-Power BI people. If the agent has the model context a junior developer would need, an analyst with business acumen and the right skills can build reports without traditional Power BI experience. Mike also won’t hire an engineer without some agentic knowledge now.

Looking Forward

Before your next stakeholder meeting, write down the one goal the executive is measured against — then check whether any report you own actually tracks the actions that move it.

Episode Transcript

0:01 be high. Tell me and Mike lighting up the sky. Dance to the data laughs in the mix. Fabric and AI get your fix. your fix. Explicit measures, drop the beat now. Hop kids, dance, feel the crowd. Explicit measures Drop it loud. Hello everyone and welcome back to the Explicit Measures podcast with Tommy and Mike. Mike. [gasps] [gasps] Hello everybody. How’s it going? Oh, Mike, it’s good to see your face.

0:31 Oh, Mike, it’s good to see your face. We’re back again for another week, Tommy. We’re jump jumping back in. It’s now summer time, so traveling has started to happen here every so often, but but we are good to go. It’s summer time. We got another episode going on. A lot of news and announcements. We actually didn’t talk about one of our articles around Power BI Desktop came out for the month of June, so we had to come back here and catch up on a couple things there. there. But other than that, the main topic for today, we’re actually going to do a mailbag today talking through that. And in our mailbag, we’re going to talk about helping leaders speak the

1:02 talk about helping leaders speak the language of data. So, what does that really mean? What should leaders understand and know about the data engineering space, the data engineering space inside Fabric? what what knowledge do we want to share with them so they could have those those really useful conversations around leveraging data for their team. A thousand percent. I think this is going to be a good one. I feel like that we’ve touched on these types of topics before, but the way the person I think his name is Jed C, which is the

1:34 think his name is Jed C, which is the probably a username, Jed C. Thank you very much for the article. Yeah, agreed. All right, so before we get into our main topic around helping leaders speak data, data, let’s get into some news items, Tommy. What did you find here in the Power BI update blog? So, the first one we have is an awesome one and this is from the Power BI blog. It is AI-powered reporting from design to deployment with agent skills. And guess what? This is actually from Microsoft. This is not just us

2:06 from Microsoft. This is not just us trying to figure out what Kurt did or what you’ve done, Mike, around agentic development through design. And what it really is releasing are specific skills in the skills for Fabric repo that can author PBI R files, reload Power BI Desktop, capture screenshots, screenshots, and iteratively refine report pages. There are four main skills. There’s a report design, there’s a report management, there’s a report builder, which is like help give the scope out. Sure. Sure. and report authoring as well. So,

2:37 and report authoring as well. So, these are Mike, I’ve used these already. And I’ve been pretty astounded on how well they work. That I don’t I don’t know they’re not perfect, but hey, hey, not 2 months ago, you and I were talking about all these skills are great, but there’s no way to actually design your report. Like that’s still the frontier. frontier. And this really is the for me what I’m looking at here, this is the gap where we have the build a report with Copilot didn’t work well. Mhm. Mhm. This is really leading to that next

3:07 This is really leading to that next stage where we can actually really build and help,, in a sense interface the report using pure agentic tooling. tooling. Tommy, this is one of the probably most requested features or items or settings when we were in Fab Con in Atlanta. So, when when we were there teaching around so Matthias Tierbach and I were teaching a full pre-con day around using agents. We’re primarily focusing on modeling MCP servers. Right. Right. So, how do you open up a semantic model?

3:39 So, how do you open up a semantic model? How do you connect to something? How do you edit the model? What things can you do with that editing? What what mentally do I understand about agents? So, there’s a whole new world of what is an agent? How do I use it? How do I use it efficiently? What does that look like? So, all all those things were being taught. But, after we did the the morning session, all afternoon everyone’s like, “Can we do the editing of the report? Can we edit the report side, too, with agents?” And at the time, this was in March, there really wasn’t a good response directly from Microsoft about yes, you can or cannot do it. You could point at

4:10 can or cannot do it. You could point at the PBI P files and do some edits there, and it would be okay. You could get some use out of it. But, there’s a lot of things the agent didn’t understand about how to work with the semantic model. So, there it was a lot of missing context that wasn’t given there. there. Skills for Fabric Was that? Was that? Or you said you said semantic model, you mean report. The on the report side, that’s what I meant. That’s what I meant. Sorry. On the on the report side, there wasn’t a lot of knowledge the agent had about how to work with the report side. Right. Right. Since Skills for Fabric have come out,

4:40 Since Skills for Fabric have come out, which has just recently have been announced and is out now, you can look at those on GitHub. The description for that is in the link of the description below. If you want to go check that out directly, I will also put in the report authoring skill. Mhm. Mhm. That If you want to go check that out here in the chat window, I’ll put that in the chat window as well. So, you can also go to this link as well. this is actually the skills they’re talking about. So, these are skills that you can go use with your agents. It now works with Desktop. And one of the things I just want to point out here, in this Desktop or this blog post Tommy around AI-powered report

5:10 blog post Tommy around AI-powered report building, building, this is not telling an agent to go manipulate Power BI Desktop. So, I had some conceptions around what this meant, and I I did a demo of this on Friday. So, the Friday video link of me doing the demo with Matthias is here in the chat window as well. So, if you want to go see that, that is you can watch us go through and set it up, do the demo, build the report all together. There are two main features for Desktop. There There’s this thing called the

5:40 There There’s this thing called the Power BI Desktop Gateway. The Desktop Gateway is allowing a command line interface to talk to desktop. That’s what’s That’s what’s really happening here. That new thing that was added to Power BI Desktop. The command line interface does two very simple things. It lets you have screenshots, and then it lets you have refreshes of the report. So, the command line can tell Power BI Desktop, “Hey, something has changed on the report layer. You’ve added a visual, you’ve changed the PBIP format. Something there

6:11 changed the PBIP format. Something there has changed. You need to refresh yourself without closing and restarting Desktop.” Desktop.” Right. Yeah. That’s basically all the CLI does. Everything else you’re doing is all skills from the repo and explaining to the agent how to build a PBIP project. So, that’s where I think a lot of the effort has gone here. So, in our our mental models here is these skills are not being used to directly talk to Desktop. Instead, Desktop is saving down

6:41 Desktop. Instead, Desktop is saving down all the files, PBIP format, Mhm. Mhm. and from there, you’re able to manipulate things. It had awareness of the the theme file, it has awareness of bookmarks, it understands what pages are, it understands the different visual types and how to add them and move them and move them around, rearranging things. So, you can do a lot faster, a lot more edits on the page by just talking to an agent and having it do it for you, which I think is substantially more fun than using

7:12 more fun than using Power BI Desktop and clicking all the buttons. buttons., , it is. And so, my workflow before this was and when the first skills came out, was having the PBR open because you needed a live port. Because yes, you could have all the skills just connect directly to the files, but it wouldn’t be able to verify the data, right? If you would actually wanted to run its own query. It can only edit, create measures. So, my workflow has been have the desktop open and not P by

7:44 been have the desktop open and not P by X but the P by R. Yep. Yep. And when I was creating DAX measures or testing things out, I was like, okay, let’s do some validation. Hey, this goes right into our previous one on the Claude for analytics, the self-service that we talked about last week. Mhm. Mhm. Good thing I was already following that where it’s like, hey, validate, there should be this many rows. Okay, it does the validation. When I was ready for the design, I let Claude know or I let my ID know, I’m going to close desktop. Here’s where the local files are. My workflow here when I wanted

8:15 My workflow here when I wanted to design something because I I found it to be very sporadic if I just said, hey, the measures we create create a report design. You can’t do that. This goes back to the context and also the prompting. I would say I sent the screenshots to hey, look at the report page X, Y, and Z. I want to replicate this in some way to some,, in the design using this with the harness, using this like and look at my previous meeting with the stakeholder, the transcript. What did they talk about? Let’s use those metrics and just

8:46 about? Let’s use those metrics and just plug and play there. And when it had the proper context and instructions, it does a nice job, man. It does a nice job even doing the conditional formatting for you, which I don’t know why to this day, man, the manual conditional formatting is so tedious. It’s such a great thing, but you’re like, every time I’m building it, you’re like, there’s got to be a better way if I have to do this. Because I don’t do money designs without conditional formatting. So, the fact that you can have these agentic things building that, saying, I already have conditional formatting, if it’s above

9:17 conditional formatting, if it’s above this, make it green, X, Y, and Z. So, that’s been my workflow, but Mike, for you, how have you seen the change your own workflow in terms of how you were applying Claude or applying,, your normal AI in integration with Power BI and what has this changed for you? Mhm. Mhm., it’s definitely shifting what I think I want to work in, right? Do I want to spend a lot of time in Power BI Desktop? Not as much anymore., I want to spend more of my time in other places

9:47 to spend more of my time in other places, working with agents. I want to figure out how to get the agents use but a lot of this is,, Tommy, I’m I’m getting to the place where I’m trying to what they’re calling like token maxing, right? I want to get the most amount of value [snorts] out of the agent without sending it a bunch of tokens. And the more I’m reading about things, the more I am firmly convinced that building tools creating creating systems systems Mhm. Mhm. that can run scripts and understand the script is we actually want to send as little little Right.

10:17 Right. text to the agent as possible., and so I’m finding myself doing a lot more of, whenever I’m evaluating skills, I’m really looking for skills that seem to be most effective for me and have produced the most results have some sort produced the most results have some like scripting or, processing in of like scripting or, processing in them them Yeah. Yeah. that like pre processes some of my data and then sends the results or only what needs to be sent directly to the agent. So, distilling that down into like real real real terms here.

10:51 I’m finding using different agents for different things is very helpful. So, all this to say, how’s this changing my workflow? Back to your question, Tommy. I think this is, opening my eyes and giving me more confidence that Microsoft has a proper agentic development story coming. I don’t think we’re there yet. It still feels a bit rough around the edges. It’s not quite as easy or simple as I would like, but I feel like this is a lot of the infrastructure, the framework that we need in order to build out what we want inside reports.

11:21 want inside reports. When you combine this stuff with what’s happening with Fabric apps Mhm. Mhm. that those are two, like,, think about what we can do now. I can

11:30 think about what we can do now. I can talk to an agent with the MCP modeling server, I can do changes to the model. I can talk to an agent around the PBIP and now I can make changes to the report. I can talk to an agent and have it build an entire application data in and out and out with Fabric apps., this is just those three things together really revolutionize what you can build or how you build things with agents. And I think that’s Yeah. Yeah. seeing that shift go from zero apps, almost zero agent ability to now almost

12:02 almost zero agent ability to now almost everything I touch before can now be handled with an agent is just very hopeful for me. I think this is very bright and it’s going to need it’s going to require us as users to figure out There’s a Let me say this way. There’s a whole new challenge evolving. The new challenge is, Tommy, we have a team of people. How do we distribute parts of this, all of this, some of this to the team? What agent do we use? What harness are we going to be using? A lot of the stuff

12:33 we going to be using? A lot of the stuff that you’re doing, Tommy, you’re using Claude code, you have a subscription, you have a certain amount of tokens. Do we want to give everyone on the team a $100 Claude code subscription? Do we How do we distribute this? What does this look like? so, where do we run it? Is it okay to use Deep Seek? Is it okay to use other providers? Still Still Which provider is giving us the best So, there’s there’s a whole new unknown world here that’s happening. And I see a lot of people on LinkedIn just doing these really elaborate 3D visual things. I’m like, dude, these are

13:03 visual things. I’m like, dude, these are like, okay, interesting, but no one’s showing you how to build any of this stuff. How long it took them to build these really impressive apps and things. So, I don’t know. We’ll We’ll see where it goes, Tommy. I think it’s It’s not changing my workflow directly yet. yet. Well, I think this comes to the age-old question that we’ve had in I would love to put together, by the way, our episodes we’ve done on skills because I think we definitely have a series there. and again, talking about how do you actually share skills? I think the management of skills to Claude’s point, and we’ve been saying

13:34 to Claude’s point, and we’ve been saying this for a while, too, is a first-class citizen. It is so essential and something that I take a lot of time in because it is what you rely on. But, let me ask you this. Does Does this new agentic skills for report design change your opinion about junior developers? This is something that we’ve asked our ourselves many times on can you hire someone off the off the street without any Power BI experience, but let’s say they have,, good business acumen, and could you give them

14:05 business acumen, and could you give them all the tooling in the world? Again, saying in the ideal world, you give them the Claude subscription, they have all the tokens, and they have all the skills, skills, could you theoretically have them doing just Power BI all the time without any Power BI experience? I want to take this question maybe two directions, Tommy. Oh, boy. Oh, boy. The first direction is to your question, to answer your direction, how about junior developers? Okay.

14:35 Okay. What does a what does a junior developer do here? do here? A junior developer doesn’t have the business context. It It’s someone you’re bringing I I would argue you’re bringing someone new into the team, you’re trying to figure out like where they fit and how they how they align with the team, right? So, on one hand, I’m looking at the junior developer going, “Okay, somewhat interesting, but not really helpful. Like, I have to spend a lot of time with them to give them the business context.” context.” Then, there’s the sweet spot. Then, we have the the team of people that have been trained on Power BI, they understand our business context, they

15:05 understand our business context, they know how to build They’ve been building things for Power BI for a long time. That’s That’s what we’ve hired, right? Those are the people that are already in the seats of building Power BI reports and and Power BI semantic models. Then, there’s the other end of the spectrum. spectrum. It’s a senior. It’s the PMs. It’s leadership in your organization. And I’m not going to assume that they always know how to do the data pieces. I’m not going to I’m not going to assume that they’re always understand how to under what’s the semantic model do. What

15:35 under what’s the semantic model do. What does the report build look like? How to do all those things. So, what I think the agentic space gives us, us, it takes a lot of technical skills that we would have had at the at the core engineering Power BI report development layer. It’s actually expanded that audience to more people. So, on one hand, Tommy, you have like the,, if I had to quantify the new green engineer, that’s someone who’s doesn’t understand our business. We’re getting more people.

16:05 our business. We’re getting more people. We’re hiring that new engineer into our system to to build things. One, I would expect them to start at least knowing or understanding. I wouldn’t want to hire an engineer now who doesn’t have some agentic knowledge anymore. I think this is becoming very quickly a needed skill. Organizations know they need to have it. You’re going to need to work with it and use it. That’s a given to me at this point. point. On the other side of this, this opens up, I believe, more capability on the individuals who have not traditionally

16:35 individuals who have not traditionally been really in the weeds on semantic models and report belt development, but they can now, with minimal effort, get started. And so, they have the business knowledge. They know what they want to go build. They know what reports will work. So, does this really I think we’re collapsing skill gap between we have to always go to Tommy to build reports on semantic models to we can now give middle management and even maybe some leader level support. These are how we build things. I still

17:05 These are how we build things. I still want to make sure we test them. I still want to make sure that’s right. I still want to make sure it’s doing the value that we want. But I really do feel confident that we could open up that door on the other end of the spectrum. So, maybe Tommy, you what I’m I think Yeah. I think what I’m seeing or maybe that Again, I was reading a little post on on LinkedIn that was talking about the new business context context is how do you How do you get all the knowledge of your business team people into a searchable, usable knowledge base where anyone can show up and get access to that knowledge. Agents or people.

17:37 to that knowledge. Agents or people. So, I I put on my glasses as I was going to do it for dramatic effect, but you stole a little of my thunder here. here. Okay, sorry. Go ahead. Because you can No, but you you were answered what I was going to say, too. say, too. The point what I was going to say is because again, the glasses. Anytime you take your glasses off, it means you’re saying something serious. Bom bom bom. Yeah, like So, hold on. So, let’s say I can do it. All right. So, not only do skills allow users, and you can clean your glasses at the same time. I think that’s what you’re supposed to do. But, not only do skills actually

18:07 I think provide a path for the engineers and the developers, but I think this is now a gateway for self-service. That I think we’ve never seen before. If I want to start allowing people to design their reports who are not anything to do with Power BI, but they are on the self-service side, they’re an analyst, they have the business acumen, and they’re more on that operations Oh, let me give them a few trainings with skills. Let me set them up with the right skills. Let me

18:37 them up with the right skills. Let me set them up with the harness. You have your own repo because we’ve set up managed self-service. I think we’ve opened up a whole path on that avenue as well. The concern I have I don’t care about them what I’m about to say. I’m not as concerned. The concern I have with what you said about giving someone who did not have prior knowledge of semantic models to get started is are you ever going to learn semantic models? If you are on the engineering developer data side, if you’re more leaning on the

19:07 data side, if you’re more leaning on the IT side of things, yeah, I I set you up with this, and you can start running, but are you ever going to learn it? Are you ever going to master it? Again, I think I’ve had success with agent skills because I understand what I’m asking and I’m what I’m requesting and what I’m prompting prompting because I’ve done it for 10 years at this point, 10 plus years. But if I give a junior developer who has limited knowledge of semantic models, when are they going to start learning semantic models? How can they test what they

19:37 models? How can they test what they don’t know? And I think that’s always where my concern is on the engineering side of things. The self-service, go crazy, we’ll validate it. You can get started pretty quickly. But on the other side, what’s their learning path then? This is an area that I think is really interesting, Tommy, because I think the agent agent is the builder and the educator all in one. one. I don’t Okay. I’m not sure if I agree, but

20:07 I’m not sure if I agree, but Well, if if you and I are doing a good enough job or a better job, good job, better job, I’m not sure how you how you want to phrase it, but if Tommy, you and I are spending a lot of time making sure that the agent has the information it needs, needs, the same way that we would give it to a junior developer, the same way we’d give it to a senior developer, right? The knowledge we understand about the relationships between tables, the measures that we’re creating, the descriptions, we have some business context to why those things need to exist. Right. Right. If we’re doing our job right, we should

20:37 If we’re doing our job right, we should be able to transport most of some of part of that knowledge directly into the semantic model either with instructions or skills or something like that so that the agents are understanding exactly why it’s there. What does it do? Why Why does this stuff exist? I think that’s incredibly important, honestly. honestly. And And if we’re doing that, if we’re dumping all this business context knowledge into some central place into the model, whatever you want to call it, now when we come back to the business user

21:07 user or the leader and say, “Here you go. Do self-service now.” I think our new task, Tommy, should be training those users on how to use those agents as the education vehicle for what does the semantic model do? How does this relationship work? What does a measure need to look like? I’m writing less and less DAX every day. I’m just telling agents to go build it now. now. Yeah. Yeah. And it’s and it’s pretty dang good. Does the tooling exist right now to have

21:37 Does the tooling exist right now to have agents agents though, and I think that’s the biggest thing. You can direct the orchestra because you played in the orchestra. I don’t know if that’s a true statement, but do what I’m I I I’m still figuring that answer a question out. But conceptually, though, like if when we build it like if looking at the model, if you just step back and say, “What measures? They’re named a certain way. Why are they there? What tables and what relationships to the dimensions?” That’s what we’re building in the semantics layer. Sure. Yeah. Because we’re building it in a in a in a space that the agent can see it,

22:09 space that the agent can see it, our new business users are going to also need the same information and same knowledge. So, the agent is both the creator or the like the the support person, the technical support person, right? It’s the developer. Right. Right. The agent is a developer and the agent knows every single thing about that model. So, if the business user doesn’t understand it, you can actually ask the business user to say, teach them to ask questions. Hey, what does this measure mean? How does it work? What dimensions can be used with this measure? The agent should be able to understand that information and be able to produce

22:39 that information and be able to produce that information for you. I thought of a great skill, actually, off of this, which I’m I may work on is if you have the semantic model, can you make it human readable and push through your center of excellence? Not just the strictness of columns, but there’s a skill there to translate that to hey, here’s what you need to know before you go, thing, like,,

23:01 you go, thing, like,, for for for for before they get to a semantic model or report. report. Where does that data live? And that’s right now where I think I’m having a struggle point with semantic models and reports in Power BI today. Currently, there really is no structure structure inside a semantic model to have a whole bunch of AI generated things at this point. point. I would slightly disagree. There nothing literally in the model, no. There’s nothing in the TIM doll. There’s nothing

23:31 nothing in the TIM doll. There’s nothing in the definition that’s going to be defined there. But that’s what I’m That’s kind But that’s what I’m That’s what I’m Like, I understand your of what I’m Like, I understand your point, Tommy. Like, but that’s if we’re going to take business context and give AI more information about the models, it should not live in an ontology in a separate thing that costs me more money to run. Yeah. Yeah. Honestly, the instructions for the model should live inside the model. I think that’s the ideal approach. I but I think there’s still a practical approach today. But data agents is like the same thing. Data agents is But But this is This is

24:03 Data agents is But But this is This is my point though. Data agents is like an Data agents is an abstraction away from the model to stick an agent on top of the model. This is a good point. Junk, right? And so, instead of making a data agent separate from the model, why doesn’t the instructions Why doesn’t more of the Why isn’t there a whole Why is Why does a semantic model not have an AI folder? We’re talking about like markdown files, simple things. This is small weight stuff. It’s not expensive. or skills for a semantic model that you can actually have in there.

24:33 can actually have in there. Again, same it’s the same thing as VS Code, right? There’s a dot VS Code, there’s a dot cloud, there’s a dot something folder that’s inside the semantic model. Dot AI, whatever, right? That needs to live there, and that’s where your skills go. Hey, when you’re building this data, here’s Here’s some skills that attach directly to the model. All right. All right. I think I I think I But this is the point is My point is Microsoft’s not building the skills or the AI folder directly into the semantic model. Now, I do want to point out here, there is a feature an open source spec

25:06 there is a feature an open source spec called the open semantic interchange. Now, some people I Chris Wagner was kind Now, some people I Chris Wagner was being a little bit verbose of being a little bit verbose here the other day around like well that’s not a good idea. That’s just someone who doesn’t have a semantic model. Like whatever. You debate me if you want. you want. But inside the open semantic interchange, there is the ability for you have a whole section around AI. And you can add AI specific columns, data fields, information that is related to an agent at the column, the table, and the semantic model level. That makes sense to me because we want to add that

25:38 sense to me because we want to add that metadata back into the semantic model. And do I want to have that data and information separate from the semantic model model in a separate storage area or should it actually live inside the semantic model, which actually I agree with. I think it actually needs to be inside the semantic model somewhere. was percent and Mike I’ve actually worked on my presets here. So, Arun, Arun, a dot AI folder semantic models. Put it in the model. Yeah, pretty simple. So, if you’re listening now, I I and if you the thing is not a terrible

26:08 is not a terrible Arun may want to know about this, but really we want to talk to Christian Wade. So, do it again Tommy. Christian Wade. Wade. Christian, hi. How are you doing? Put a dot AI folder in semantic models. Thanks a lot, Tommy. For those of you who are not online and watching this online, Tommy continually has found the zoom in button on his camera. camera. button so I can So, now he can with emphasis bring him Tommy in close. Come in close. Come in close. close. Come on. Come on. Camera two. Camera camera two. Camera two.

26:38 No, and I think this is not a heaven and earth thing. Just add that ability. Yeah, I I think Mike I think we’re going to see this. I don’t see why we wouldn’t see this. I this is absolutely going to be part of the future. I almost want to make a bet with you on when that’s going to happen because I think it’s going to happen in 2026. So, So, you don’t think it’s going to happen? No, I do. I would put the under on December 30th, 2026. Hmm. That it’s going to get a full it’s going to get a full So,

27:08 to get a full So, There will be a folder in the PBIR and then the report definition in the semantic model definition called dot AI or some type of folder. Some place to put AI things inside the model. model. I can put skills, and I can add it like a dot,, a agents. md with a skills folder and other resources. Per semantic model. Well, maybe. I’m not sure. I’m not Do you want Do you want to bet a steak

27:38 Do you want Do you want to bet a steak on it? We’ll make it a little fun. Tommy, you’re betting on something that I feel like the Microsoft team is very opinionated about the semantic model layer and they’re very hesitant. Like they very rarely add things to that as well. well. So, I I I really I think it makes sense, but but Because? Because? but let me hmm I’m going to put on my I’m going to put on my skepticism hat here for a minute. Okay. Okay. If I So, not not going to talk about the bet, but I want to talk about like when I look at Microsoft and what they

28:08 when I look at Microsoft and what they do, do, one of the things that I feel like is happening is they’re always looking for where can we add value, but get people to get more consumption. So, there’s like two conflicting goals here. Hmm. Hmm. If you put AI directly in the semantic model, model, that means you’re storing the metadata with the semantic model, but that can be used anywhere. Sure. Sure. I don’t have to use Microsoft compute to do that. do that. This is true. When I give you a data agent, the data agent now requires you to use the

28:39 agent now requires you to use the compute of the fabric to run the data agent. So, I feel like all we have a while we while we while I feel like we have a good idea Tommy about keeping the data in the semantic model makes sense skills should be there and I don’t disagree that’s just storage of information and skills and such so that’s actually not a way that’s not a way for Microsoft to like get more see you usage out of it via ontologies or get more see you usage

29:09 via ontologies or get more see you usage out of data agents directly so this is where I think I don’t know if they would push it into the semantic model I think they might try to leave the semantic model just pure and that’s why I bring up the OSI I think that’s I think that’s really why I like the idea of the OSI I I and I I agree with you but your point is more open source you can use it with anything it doesn’t have to be Microsoft that that’s an interesting point now the one hope I have or

29:39 point now the one hope I have or the one positive is they did incorporate skills they’re building skills which are also Universal in a sense language now yes you got to get with the program got to you got to start somewhere so I feel like this is a starting point do their skills do do everything that I want want probably not do their skills have heavy token usage based on what I look at the skills and their the fact that they’re all markdown and they’re very large markdown files at that if you go look at their skills

30:10 their skills good skills but there’s a lot of like their skills there’s sub skills there’s a lot of extra it’s got to read all that markdown and that’s all tokens that are going to the context of the agent to have to help you build things so my my understanding is if I I feel like I want to have more scripts deterministic things hey here’s a script to build a visual here’s how you run this

30:41 here’s how you run this you don’t have to know how to write the exact JSON of this. You can give it some just general in pro input parameters. Here, here, here, and here. And then out pops from that all the things that needs to be corrected and verified from that function. I feel like we need more things like that. I think I need more scripting that’s very tightly coupled to the skills and the agents that makes it very regular to produce the the output. I I would agree with you and I’m curious what language you would be in for the report design. Would it be in JavaScript

31:12 report design. Would it be in JavaScript because of DS3? Would it be in, you because of DS3? Would it be in,, know, No, no, no, I’m not I’m not talking about I’m literally talking about making the PBIR assets. Right. Right. Yeah, but I’m saying yeah, I’m I’m with you. you. the scripting can my preference is my preference is put it on Python. It’s just a Python function you call. So, a Python function you call, you say, “Hey, agent, when you’re trying to build a visual, here’s the things that we need. This is the function you call and when you run this function, you pass in this. You pass in the size of it. You

31:42 this. You pass in the size of it. You pass in the XY location. You pass in this.” So, that way the agent can do a lot less management on It’s just like us, Tommy. Like We don’t We don’t care to have all the details of the visual. We just want something to show up on the canvas. canvas. 100%. 100%. And and and that’s where we should be spending our effort on. We should spend our effort on like what does the page look like? Right. Right. the KPIs go? What’s the general layout of the page? that should be all known to the agent and it should just say, “Okay, I want to just spot,, here’s a couple visuals. Bing, bing, bing, bing.” Just put them

32:12 Bing, bing, bing, bing.” Just put them across the page. Yeah. Yeah. And all of that should just be able to describe what you want and where you want things and the agent should be able to pass that into functions that automatically understand how to use all those inputs. That reduces the token size substantially. And it doesn’t have to understand the spec of building a visual on a page. Right. Well, I love that and we’re already halfway through so we should Maybe we should get to our main topic. All right, let’s go to the main topic. two-topic episode. Well, I would also argue, Tommy, what

32:42 Well, I would also argue, Tommy, what we’re talking about here with agents and agentic things is starting to help with this next topic, which is Ooh, I don’t know if I agree with that at all, but let’s let’s Okay. Well, remember remember I you were talking asking me about like the junior developer and then what’s an actual developer versus like those those leader developers, right? Oh, yeah, I’m hearing you. So, I think,, what we’re going to emphasize now is and we’ll read the question, we’ll see where it fits, but I think what we’re talking about now is where do agents fit in this new world? And as we’re talking to leadership around communicating about data, the

33:13 around communicating about data, the importance of data, it becomes an asset for us. for us. How does that become easier for us with the addition of agents? Can we give those leaders more usefulness? So, let’s go back to the question. buckle up. Let’s see if you agree with Mike here. So, here we are or you’re going to be you’re you’re better or you’re going to kick you off the podcast. All right, this is from Jed C and this is the mailbag. Really like your Power BI Analyst 2. 0 episode.

33:44 BI Analyst 2. 0 episode. I agreed in a lot of things you talked about, but I’ve been thinking a little different about the topic. It is about time we start setting the demand on the CEO, CFO, etc. They need to be educated in what they want and what it should look like. They want A and we need to find B because that is what they need. And their design wish is usually a copy-paste from Google, not a connection to the data.

34:15 Google, not a connection to the data. Would it not be nice that would it not would it wouldn’t be nice that we could talk the same language with them? So, I think a big part here, just kind So, I think a big part here, just going off the bat, is Mike, let’s go of going off the bat, is Mike, let’s go back in time. Let’s go back to our data analyst years when we talked

34:32 analyst years when we talked Oh, I said back in time. I thought that was a trigger. Oh, sorry. Let’s go Yeah. Back to episode analyst 2. 0. I was going to say back to when you were just a lowly data analyst. Back when you were just building reports for clients. I was in diapers. Back when you were Yeah, back when you were Power BI diapers thing. We were just learning how to walk. Mike, a lot of my beginning of my career in Power BI was,, bending to the knee of the C-level. A constant thing we said on our

35:03 C-level. A constant thing we said on our team was, “We’re not doing everything, but if comes from someone that their title starts with the C, you just say yes and you do what they ask.” And a lot of times there were great executives I worked with, but there were a few that their ask had nothing to do with providing business value. It was, I I had a one point Mike, we were just starting with Power BI and they asked one of the junior developers to build You’re like, “Actually, I want this in SSRS.” And the person didn’t know SQL. And they spent three nights in a row at the office building this That’s

35:34 a row at the office building this That’s what the executive wanted. And I This is nothing that’s necessarily changed where executives do not seem to have a good sense or handle on their own data. Mike, just yesterday I was talking to a partner of mine and they’re working with this company to They’re merging companies together. They said, “Hey, you should probably have a dashboard that sees,, what your utilization is on,, the number of staff.” What a great idea. They’re like, How do they How do you not think about that as an executive? So, we’re on this always this battle or

36:05 So, we’re on this always this battle or this this back and forth with our role is to provide insights through data to help people make decisions through data. And a lot of executives still rely on hunches,, or their own 20 years of experience. And there becomes a lack of language, a lack of understanding for what they want to what’s possible to what we should be doing. Mhm. Mhm. So, that’s listen that’s setting the stage and I’m curious to hear from your own past, your experience working with

36:35 own past, your experience working with the higher leadership of companies on their own literacy around data. Yeah, this is some leaders get it in the in different leaders dependent on where they’re at in the business. I think understand differently how data is designed and engineered. This I think this skill set varies widely across different people in the organization. also I think leadership if they take on this

37:05 if they take on this perspective of we Data literacy is a skill. It can be learned. learned. It’s something that the organization needs to adopt. We’ve talked in the past about this a little bit, Tommy, which is when we talk and build reports, we’re actually communicating in the language of data. We’re not saying words, we’re giving you a bar chart. Does the team understand how to read that bar chart? And I think that data literacy element should go all the way through the

37:35 should go all the way through the organization. If leadership doesn’t understand or isn’t talking that same way with the rest of the team, we’re having miscommunications. We need to over over communicate there. Right. Right. I also think this varies widely between organizations. Some leaderships are extremely data savvy. Extremely. Others, maybe not as much as we’d like them to be. be. And I think there’s, this is where I’m going to go back to my training way back in the day, Tommy. I did this one for

38:06 did this one for black belt. I did a a black belt training around analytics. And and our leader at the time was very much, , we’re doing a lot of category management. And so, we went to a Delta Associates, had them come in for like a month and train us. Like it was an intense like working through Excel, building calculations. What does category management mean? How do we calculate these things? The entire team team got on the same page with analytics.

38:37 got on the same page with analytics. Things like that are really unifying. So, when we say, how does your category man- How does your category in whatever market you’re selling, did you grow in market share? Did you lose market share? What does What does that look like? Now everyone is at the same level and we we understand the same fundamentals. And so, I think that really brought the the entire team up a level into a place where we could really understand and talk together as a team holistically around what analytics looks like.

39:07 like. So, I think there are leaders out there that have great acumen for this and others that do not. And so, you you have to be able to get that all to happen together. how do you get leaders who are not in that space to really jump in and lean into this? I’m not sure, Tommy. I think some of this is this is Yeah. Yeah. you have to gently like guide and give good dashboards and like work with them, but I think that I think this is a it’s a skill that everyone can learn. I do feel though someone in the

39:39 learn. I do feel though someone in the organization needs to spearhead analytics is a thing. We have to realize not everyone has the same level of knowledge and analytics and we’ve got to work hard as a company to bring everyone’s analytical knowledge up better, think through it more. I would even back up from there and I and I I love what you said here because again, doing an actual training with your leadership on,, the basic fundamental of the analytics is huge, but I think I would even back

40:09 is huge, but I think I would even back up to what you’re saying because there are four main areas I think you need to look at to get really that language to be on the same page. Mhm. Mhm. There is data culture, which is simply the defined set of expectations, the defined set of where to go to find data,, speaking the same language is going to be literacy, but data culture is,, where do we find the right things, who do we get to,, who’s responsible for what, how do we rely on data literacy

40:39 how do we rely on data literacy is the ability to read and know what to ask for. I know and can define my metrics. So, I can therefore communicate that with my team or with a BI team across the board. We know leads, we know marketing leads, we know sales leads. We have a full understanding of how not only what the definition is, but how it impacts. So, data culture, data literacy, literacy, and I’m going to add another one here, you mentioned here, the cognitive load. We’ve talked about this a long time ago. Yes, we have. If I introduce a new visual to a company

41:09 If I introduce a new visual to a company or a new metric to a team, their cognitive load is going to be high because they have to process that new visual. If I added, like say, a running total, total, and or if it’s a new metric, I have to introduce them,, what that metric represents, what the impact is. So, my ability for them to their ability to cognitive load on my report and my ability to teach them is huge. And the last thing is buy-in, that we actually believe in the power of

41:39 that we actually believe in the power of data and we’re going to use it. Those four together are what’s going to help elevate your executive team. Unfortunately, I know Power BI can become from the bottom up, but I think real success comes from the top down. So, So, yeah, I’ll pause there. Yeah, I think you’re right. I’ve used a lot a lot of times I communicate an immense amount around the Power BI and the Fabric Adoption Roadmap. I think that’s a really strong tool. I think Microsoft recognizes

42:09 tool. I think Microsoft recognizes that’s very useful. Most organizations that I believe adopt that, talk about it, use it, leverage it, find a lot of value from using that tool. as as a a system system on how to measure and monitor what’s going on in the organization analytically. I think that’s a great measuring stick. Because a lot of times you can walk into an organization and say, “I know what analytics are. I know how to handle this stuff. There’s no source control. There’s no there’s no regular communication of

42:40 there’s no regular communication of things. We don’t have a maturity around certifying data.” People are just So, having a standard to look against, to measure your company against, that’s what the fabric adoption roadmap does. It tells you all the governance things. So, I like to walk in with that conversation and be like, “Look, I’m not making this stuff up. I can tell you till I’m blue in the face on how to do your job or how to run things in your company, but at the end of the day, you need to have you need to look at like baseline yourself against every other company that’s out there.

43:11 other company that’s out there. That’s what that system does. I would I Mike, I would almost back up a step. I think that’s step one, but I think there’s a step zero here. Does a company even have in place OKRs or some semblance? And OKRs are going to objective key key results and yeah, yeah, What measure measure success? What measure success? And I think honestly, a lot of the times I have found personally, companies who do not buy into into [snorts] [snorts] analytics have not even bought into expectations around what they’re

43:41 expectations around what they’re actually trying to measure. And it’s not just oh, we have a KPI. We’re looking at sales. We want sales to go up. That’s not an OKR. That’s not a good way of looking at your data. Honestly, because that’s just again, doesn’t really make an impact downstream. A good company, a healthy company, even again, before you start with the infrastructure, a healthy company around data understands the metrics important to them and understands the metrics important to their team to do their job, whatever the goals of the company are, and what supports that. Just start there. That

44:12 supports that. Just start there. That could be an SSRS. It That part doesn’t matter on what technology you’re using or what analytics or visual. Do I know for my for myself what am I measured against and what do I need to do for whatever our goal is? And do I know what supports my team for them to do their job successfully? That that’s what we’re not only just being measured against, but I know I will be successful if I do X, Y, and Z. I know this sounds like,, your business 101 course, but when you get into the world

44:42 course, but when you get into the world real world, my friend, a lot of companies are failing at that. And it’s hard for me to say the Fabric Roadmap and all those things around analytics if that’s not in place. And I’m reading this mailbag over and over and this question, and I I find really this company does not have a leadership who are even asking those questions. So, how in the world are we going to introduce to them governance? And how we’re going to introduce all those things because they obviously have no idea idea what impacts their team and why

45:14 what impacts their team and why I how to make my team more successful. That’s what we’re trying to do here. I’ll go back to like the standard, right? I like looking at I like looking at the standard of the Fabric Adoption Roadmap going through every single section in that whole documentation article describes it has a whole bunch of sections as okay, if you’re level 100, here’s what you’re doing. Here’s the actions and here’s how your business is operating at a level 100. If you’re level 200, here’s what you’re doing. Here’s a level 300, what you’re doing. And so it’s as you

45:44 what you’re doing. And so it’s as you evaluate where you’re at on those scales of things, you’re able to go through and understand where does your company fit. And I think to your point, Tommy, Tommy, having alignment on measuring the right things things fundamental. Right. Right. Very very But that’s that teases out

46:02 Very very But that’s that teases out having those conversations teases gets teased out when you go through the fabric adoption roadmap. And I’ve heard it from Microsoft directly. This is these are in the largest meetings with Microsoft’s largest customers customers if customers are not going through the fabric adoption roadmap and they’re just doing their own thing, they’re not nearly as successful. The the number one area of Power BI success in an organization is having an executive sponsor sponsor that understands it and it puts

46:32 that understands it and it puts resources and allocation behind it to make it a strategic asset. Right. Right. That that is exactly that is the number one thing one thing that helps companies win. So So It’s yeah, go ahead and keep going. Since I’ve heard since I’ve heard it directly from Microsoft and the customers they work with that are much bigger Some of my customers are big. The customers they work with are bigger, right? So if they’re saying that stuff works and these are patterns that they’re seeing evolve is the executive sponsor of the Power BI system needs to

47:02 sponsor of the Power BI system needs to understand and communicate clearly around data and that needs to be trained and educated to the broader part of the team team then we’re adding value. And I think I think that’s a good thing to lean on. And I So let’s go to the language idea here and because I think part of obviously huge part of this is the roadmap, but to talk the same language. This is what Jed C initially said. How do we make sure that with our executive or leaders will will say the leaders as a category here not necessarily just executives. How do we make sure we’re speaking the same language with them?

47:34 speaking the same language with them? I would challenge you, Mike. If you looked at all the clients you’ve worked with or even again back in the day, do they have a set defined of goals for the company for that year? usually and it’s more than again not that’s not necessarily a metric. Like, we want to have a bigger you , we want to have a bigger, print in a market share. We’re know, print in a market share. We’re introducing a few products. We want to make sure they’re successful. Whatever the case may be. Mike, when everything changed for me when it came to the first company I was working at where their data culture

48:04 working at where their data culture really started becoming healthy and I matured was when we had to define goals for the quarter or the year for the company. company. Yep. Yep. And the only project we worked on that helped us were aligned with one of those goals. goals. Correct. Correct. So, rather than we’re just going to ask on whatever an executive says, well, how does this request, this report, this model align or support the goal that you’re trying to do or the company’s trying to do? That’s a very easy way to

48:34 trying to do? That’s a very easy way to start beginning to talk the same language. Because here’s the thing, you may have someone in marketing who’s like, oh, we need to increase emails. I need to see all the emails and the subjects. Okay, we can do that, but now we’re spending all our time on something that’s pretty tedious. And if that’s not really supporting the goal of marketing from an executive, what are we doing? And I think you have to start with what is the Again, this is the empathy part. What keeps someone up at night? And I always go back to this. I will talk to an executive. I will ask them, what keeps you up at night?

49:05 them, what keeps you up at night? Now we can talk the same language because I know what bothers you and I know what you’re trying to do. The funny thing is when you do that, they may ask for a report that has nothing to do with that, which I’ve always found confusing like start jarring. We’re like, the things that cause you agita and problems, but you’re asking for 18 things just for shortcuts. Mhm. Mhm. There’s a disconnect here. So, Mike, when you think about the language side of this, when you think about how to get leaders to start getting getting in the cadence of asking BI

49:37 getting in the cadence of asking BI teams teams for the right things, what’s your approach? Like I mentioned mine, but do you have a certain approach on that introduction to let’s just talk about the right measurements here. Yeah, I’m going through this a lot right now, Tommy. It’s more about,, for me it’s a lot about listening what they say, right?, it a lot of this is having that empathy at that first level of tell me what runs your business. What are your objective goals? Let’s talk about short-term goals. Let’s talk about long-term goals. So, I think this goes into while the language may

50:08 this goes into while the language may not be used in those conversations, Right. Right. this all alludes itself to the OKRs. Objectives and key results, right? That’s That’s what I’m hearing a lot of you saying right now is Yeah. Yeah. Does the company have OKRs? At the leadership level, what do we care about? Is it amount of production? Is it amount of revenue? Is it amount of customers? Is it growth in a certain area? Is it a hiring head count thing? You hiring head count thing?, if and if we do our job right, know, if and if we do our job right, there’s likely a certain level of metrics in our company that says if

50:39 metrics in our company that says if these metrics are going up or if these metrics are going down, we can then measure the company as being a success. So, I think just having those conversations, Tommy, and talking through through how do we measure I don’t know if you did this, Tommy, when you were in companies. Now that I have my own company, I don’t do this anymore cuz I’m the boss. So, It’s good to be the king. Every year Every year you would have to sit down and say, “Okay, what are your measurable goals for this year?” Right. Right. Right? And so, you had to go pick out,

51:09 Right? And so, you had to go pick out, “Okay, this year I’m going to do these three or four or five things to determine whether or not I’m being successful at work or I’m not successful at work.”, things like you at work.”, things like, I have know, I have at the company level, we’re we’re building a new program, right? Well, my objective may be we need to get through X number of design cycles or we need to get through X number of releases, right? So, tying tying your goals of things that you work through, Right. Right. it’s highly

51:42 possible that people get distracted with so many shiny things. And so, And so, I feel like this is a book somewhere that I don’t know exactly who does it, Tommy, but it’s this this idea of saying no. no. It’s better like the art of saying no or say no more or there’s got there’s some concepts here there’s a book out there with those titles. titles. I don’t maybe there is. But but they I think the concept here is you could say yes to a lot of things. Mhm. Mhm. Being successful is more about what you say no to. We’re not going to do that.

52:12 say no to. We’re not going to do that. We’re going to focus our attention here. I could go a thousand things, but I need to hone in and lock in on the feature, the app, the software, the, whatever that thing is when you and when you lock on lock in on doing that that one thing. I’ve seen a lot of business leaders do that. Like, okay, ignoring all other metrics, what is the one metric that we care about the most or two or metrics that we care about the most? Let’s lock in on those. Let’s dial in on those. And then after those get achieved, you can

52:42 then after those get achieved, you can step back and look at other things. So, I think having those conversations of what does success look like? like? Mhm. Mhm. Then the harder conversation comes around, well, how do you measure that success? success? So, I I I love what you said here because I am intrigued to hear your thoughts thoughts on how do you do that to an executive? Because here’s the problem with what you said. This works great when someone’s asking me what my goals are if I’m, you asking me what my goals are if I’m,, a director or, in my own know, a director or, in my own company because obviously I need to do

53:12 company because obviously I need to do that. Someone I have someone overseeing me. But if I’m an executive, I’m not going into a meeting with my CFO and going, “What are your goals? Can you tell me,, what what are you going to measure yourself against this year?” Right? No, but what I’m trying to say here what I’m trying to say here is you can but you can ask the question of like You don’t ask it that way, I would say. I know, I know. I yeah., they’ve already If they haven’t told you, one that’s a huge miss in in the to beginning, right? So, that’s a major miss in one. If I’m a CFO and I’m not directly, I can assume some

53:44 not directly, I can assume some things, right? Obviously, our goal is to increase revenue, reduce costs. Right. Right. Right? Right? I could walk into any organization, and none of them will argue with me about any of these these scenarios. What you do ask though is, how do we get to that goal? goal? Yes. Yes. Right? Right? Yeah, yeah, yeah, yeah. What,, What,,, “Hey team.” Or hey,, “In your Give me your vision for.” Right? If you have some Again, they’re all being The C-level is all being measured on something. Someone’s getting a bonus

54:15 something. Someone’s getting a bonus somewhere on something. Right, right. Right? So, they’re highly motivated to get the company moving in a certain direction based solely on what they’re saying. So, the I think the trick here is teasing out Yeah, yeah. what they think are the correct actions to get to that goal. They own it. that’s where we focus our reporting. you focus your attention on like So, So, How do we focus our reporting, our attention on making like aligning to those actions that they just said is the

54:45 those actions that they just said is the way that we win. So, like if their objective is we’ve got to increase revenue, revenue, okay, that’s too fuzzy of a term. Exactly, yeah. What does that look like? Does that mean getting more money out of the largest customers? Do we double down on existing customers? Or are you thinking that we need to have a portion of that revenue coming from brand new customers? Right. Right. What Where do we spend our time and effort? And so, that’s that’s the kind effort? And so, that’s that’s the conversation that we want to tease of conversation that we want to tease out is directionally what actions does that leadership want to hear, and how do we move that towards

55:16 to hear, and how do we move that towards getting to the goals? This is actually a great opportunity here if this is not something that’s been done in your in your company, where you can actually talk to your boss. And what I would recommend here, this is a great opportunity what you’re saying, to get everyone on the same page., and this is going to be a scenario that I want to hear your thoughts on where if I get my boss aligned on this, that we’re going to start some new initiatives here. I’ll meet with the executives. I’ll meet with leadership and say, “Hey, I want to set this time with you. We’re starting some

55:46 with you. We’re starting some initiatives for the new year to get our reporting better value. Get aligned with the goals of the company.” So, that being said,, just want to go over the goals that you have and what teams you oversee. What are you trying to do for those teams? Because we want our reporting to align with that. and I think for an executive, you’re coming into that then not saying,, “We’re just building new reports.” Or “What’s your goals?” Right? Where you’re actually going, “Hey, you say you want more revenue, but you’re overseeing the marketing team. What are What are they

56:16 marketing team. What are What are they trying to achieve then to do that?” They’re They should have some insight. They do have responsibilities, obviously. And you’re going to say like, “Every single one of our reports is going to provide value towards that. We want to make sure that everything that you can see is aligned to with what you just said.” All of a sudden now Now Now only have you, I think, begun to build that bridge between the language, because they know the reports they are going to see is going to help the you going to see is going to help the, the new products that they’re know, the new products that they’re creating. Or it’s going to help creating new accounts in,, and overseas.

56:48 new accounts in,, and overseas. Right? All of a sudden, those two goals, that’s what all of our reports are based on. on. Yes. Yes. So, it has really some It has some sentimental value to them, which is important. The other thing it does that I wanted to introduce, it provides some authority to BI team. Right? You’re just not,, we’re not just building apps for you or setting up the network for you. We’re a strategic part of you doing your job. And I think those are two huge things where you want to speak the same language. Because I can’t go in and just tell an executive no. Right?

57:18 and just tell an executive no. Right? You can’t tell a leadership usually no. But But We’re going to I’ll pause there. I’ll stop there. You’re You’re not saying no, but I think you’re just asking for clarification. I think it’s I think there’s a non-threatening way to come into this, which is I need to clarify. Like, if you you understand what direction they want

57:32 you understand what direction they want you to run, then that’s not leadership. 100% 100% And so, I think it’s it should be and I would be very hesitant to work in a company that’s not at least allowing their employees to ask for clarification. Hey, I heard you said this is what you said in the past, can you help me understand clarification on these specific things? And then I make Greg makes a really great comment here in the comments, which is, our job is to leverage data points, restack our resources, align people in what they work on. Once we have

58:03 what they work on. Once we have alignment of our teams, our data into what’s going on, our job is now designed to be building visuals that identify what executives are trying to work through. And that’s exactly right. So, I think the conversation here is less around, telling an executive how to do things. It’s more about help me understand what do you see as being pivotal in where our work our workloads needs to shift to be that like are we deficient in an area? Are we don’t Do we not have

58:33 in an area? Are we don’t Do we not have enough staffing in a certain part of the the team that you feel is missing? Are we not getting enough accomplished? And then everything you should be focused on should be, okay, are we leveraging a AI do that job or does people need to do that job? that job? Right. Right. And one Yeah. One caveat here is you probably if you’re an analyst or you probably if you’re an analyst or, know, manager of BI, you probably can’t do this on your own. I would not recommend just setting up these meetings and going in and just going, “Hey, CFO, I’m I’m

59:03 in and just going, “Hey, CFO, I’m I’m roguely meeting with you today.” I think there’s a huge buy-in part of here that is so essential here where, consultants can do the job, happy to talk, but really more importantly talk to your boss and if you have their bosses’ boss to say, “And this also sets you up as a leader too in your company.” And all of a sudden that you’re proactive where I would really recommend the buy-in side before you start this. If you really desire that same language to be spoken, talk to your boss. If they have a boss, have them talk to say, “We want to start

59:33 have them talk to say, “We want to start something for the new year or for the next two quarters. We really think we can transform reporting for the executives.” Because I guarantee you if you’re if someone’s if you have an analyst like Jed C who’s saying it would be nice if we talked the same language, guarantee you executives have pain points with the reporting right now, too. too. So, this is a win-win. Get the buy-in, and then you can just begin to set up these meetings where you’re introducing new habits,, for the for the company around

60:03 , for the for the company around BI. You’re introducing that initiative, so it’s not just me randomly meeting with you. This is something we’re doing company-wide. Now, there’s a habit that’s one There’s a common cause. There’s common,, in a sense, what to look for. That can really begin. And I think buy-in is such a huge part here because you cannot do this on your own. You cannot just one person wrongly I get everyone on the same page. Yeah, I think I think I again it goes back to asking a lot of more of these clarified questions. Right. Right. How do we get there? What is your main

60:33 How do we get there? What is your main objectives? What are we trying to get to? So, while I’m not pushing the executive to say, “You’re wrong, and I’m going to,, I’m right, you’re wrong.” That’s not the That’s not That’s not It helps no one in the situation. Right. Right. Right. It It It’s about about like, “Okay, well, help me understand to get to that objective. Where do Where do we need to land so that it works better for us and the and the company, the team, this month, the quarter, whatever the thing is, let’s really align to that direction.” And I think that’s where I really want to spend my effort and time. And I think having a

61:03 effort and time. And I think having a couple wins or successes like that is good. One thing I maybe would caution here is here is one of those meetings may go well, but making this part of your culture, Mhm. Mhm. routine, the process that you use month after month, quarter after quarter, there’s something to say for just being consistent. Yes, that’s such Mhm. Mhm. I I think a lot of times we we focus on a particular goal or direction or North

61:34 a particular goal or direction or North Star for a period of time, and then something happens, leadership changes, I get a new person in, the North Star direction shifts, or something like that. So, I really do think that leadership level needs to be like really focused on a specific direction. And and really honing in on this is what we this is what we know to be valuable and what will be good for our team and how we get stuff done. Does that make sense, Tommy? Like I think, that Yeah, not only does it make sense, it’s obvious. It’s honestly

62:06 sense, it’s obvious. It’s honestly No, Mike, I think I am so happy you said this because you cannot just do this in a vacuum. I I think noth- it will not succeed succeed Consistency. You got to have consistency with it. with it. Yeah. Yeah. I’m like, “Hey, I meet with you. By the way, I want to send you every 2 weeks the progress on what we talked about.” Exac- especially if it’s a C-level. And that becomes the goals of the BI team, but not only that, that you communicate proactively with the people you talk to. Because guess what? You get them to buy in, all of a sudden there’s a little more budget, there’s a little more, you

62:37 more budget, there’s a little more, you more budget, there’s a little more,, know, leverage that you have to get things done the way they need to get done. So, you either have that leverage or the the things fail and you’re now swimming upstream. So, that’s, that’s a huge part. But, Mike, this has been an awesome conversation. I will leave you with a question here. as we wrap up here, is there a certain habit that you think when if you’re How do you gauge success here? How do that teams executive leadership is now

63:10 that teams executive leadership is now speaking that same language that has that data literacy? Is there a habit? Is there a certain way they’re talking, you there a certain way they’re talking,, or asking questions that you can know, or asking questions that you can begin to gauge that, hey, we’ve actually met what we’ve been trying to do here., I think , I think I think there’s maybe two habits that stick out in my mind around this one. One is that there’s quarterly or regular meetings happening on a calendar to align direction. People are going to drift, goals are going to drift, it’s going to happen. It’s just It’s just inevitable.

63:40 inevitable. nothing that I have ever observed gets more organized over time. Everything goes to entropy. Everything goes to chaos. Everything goes to like the least amount of effort. Yes, my kids’ bedroom, the stuff in my house, like Great. Yeah. most everything was like was like So, when you’re in these situations and when you’re pushing for a direction, realize you’re adding energy into things and you’re trying to make things better.

64:10 and you’re trying to make things better. And so, there’s just the same way like to your point, Tommy, around like our kids’ rooms or whatever. It’s just an atrocious mess all the time. Right. Right. But, you have to put in effort and say, “Okay, I’m not just going to put this shirt down or these clothes down or this toy down wherever. I need to go put it in its proper place. Do I have a place to put it?” So, there’s extra effort I’m always exerting to keep things aligned and straight and clean. So, I would translate of that back over to business and say, is there a cadence, a culture, a repeatable pattern

64:41 cadence, a culture, a repeatable pattern that’s being used over and over again that the business can lean on to say, “This is the pattern of behavior and this is how we’re realigning at minimum quarterly, right? Hopefully once a month. so, that would be one area. The other one would be is if we’re doing things right, right, we’re going to be asking for more reports or more things around the OKRs and the direction of the company. So, if our KPIs are measuring what we want,

65:11 if our KPIs are measuring what we want, successfully. I would hope when we reallocate resources to make the numbers go up or down, whatever whatever you’re designing for, right? If you can if we can distill if we can distill our objectives into key things, right? We want customer satisfaction to go up. We want server delays and and other things to go down. What are we measuring? And if our measurement is now aligning if the direction of the trendline is aligning with the direction

65:41 trendline is aligning with the direction of what we want, that’s another I think indicator of success. And I think organizations that start doing those things and start having like really clear metrics and steps and aligning the direction there, I think that really helps. Kim Manis just put out a really interesting blog post on LinkedIn, Tommy, where she talked about how she has agents summarizing her emails, customer feedback, and doing a lot of like aggregation every morning, where she would spend hours in the morning reading through emails, materializing data. She now sends her agent at it and

66:12 data. She now sends her agent at it and the agent goes through and absorbs a lot of that information and provides her a much more condensed summary of what’s going on. going on. That is a a feature of an executive that’s always retooling, realigning the direction of what they think they need to work towards, towards their customers, towards customer towards satisfaction. Are we getting more usage? Are we meeting customer needs? Cuz I think at the end of the day, if we make products and things that people love to use, they’ll consume more of it. If we

66:42 use, they’ll consume more of it. If we make it more economical to use those products, they’ll consume more of it. That seems to be a trend. So, anyways, those are the things I’m going to be looking at to see if we’re heading in the right direction. Dude, I love it and I’ll just add here, I’ve said this, I think on one of our early episodes, easy way, if you’re walking to get your coffee and you see a meeting happening with leadership and they’re looking at a Power BI dashboard, you’re on the right track. Yeah, well, that’s part of the meeting. But, hey, Mike, this has been a great episode. Love the question. Thank you,

67:12 episode. Love the question. Thank you, Jed C. Got I love these types of conversations with you. you. Yeah, thanks for the question. It was a really good question and I think this is a wonderful one to unpack. A lot less technical, a lot more like,, people and conversation-wise and like, how do we handle this inside our business context? All right, that being said, thank you all for listening to the podcast. We really appreciate your time and your ears today. We know you could be doing a million other things like actually doing real work and really working on your KPIs or doing other stuff. So, for those of you who have

67:42 stuff. So, for those of you who have spent time away from your normal day to to listen to us and talk to us, thank you so much for your ears. Do us a favor. If you like this conversation, please give us a thumbs up on the video. It definitely helps us get this out to more people. We’d really appreciate that. And if you’re not a subscriber, either just subscribe to the channel, that’s free. You don’t have to do anything there. Right. Right. Or become a member and help support us. That’d also be great, too. So, either one of those things are really useful things. So, make sure you subscribe down below. That’s free. It helps us also get the word out there about these conversations as well. Thank you all so much and Tommy, where else can you find

68:12 much and Tommy, where else can you find the podcast? find us on Apple, Spotify, wherever you get your podcast. Make sure to subscribe and leave a rating. Helps us out a ton. You have a question, idea, or topic that you want us to talk about in future episode, head over to powerbi. tips/podcast. Leave your name and a great question. And finally, join us live every Tuesday and Thursday a. m. Central and join the conversation on all powerbi. tips social media channels. Thank you all so much and we’ll see you next time. Explicit Measures pump it up be hot. Tommy and Mike lighting up the sky.

68:43 hot. Tommy and Mike lighting up the sky. Dance to the data, laughs in the mix. Fabric and AI, get your fix. Explicit Measures, Measures, drop the beat now. Podcast Kings, feel the Explicit Measures Drop it Drop it loud.

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.