AI & Data Culture Problems – Ep. 299

AI can generate text, code, and even insights—but if the underlying organization doesn’t trust its numbers (or can’t agree on what they mean), AI just scales the confusion. Episode 299 is a practical conversation about data culture: ownership, definitions, and the habits that make analytics sustainable.

News & Announcements

• FabCon Community Conference (Microsoft Fabric Conference) — The team highlights the upcoming conference (March 26–28) plus workshop days (March 24, 25, and 29). Use discount code Carlo100 for $100 off your ticket.

• Power BI Theme Generator (Tips+) — If your team struggles with inconsistent report styling, the theme generator helps you create and iterate on a reusable theme so your reports look consistent without reformatting everything by hand.

Main Discussion

This episode riffs on a key idea from the Forbes article Why AI isn’t going to solve all your data culture problems: culture beats tooling. AI is powerful, but it can’t replace clarity, alignment, and accountability.

• AI is an amplifier: if your metrics are unclear or inconsistent, AI will confidently produce answers that look “right” while reinforcing bad assumptions.

• Start with shared definitions: agree on what key measures mean (and where they come from) before you chase automation or fancy experiences.

• Trust requires transparency: lineage, refresh behavior, and data quality signals matter more as soon as people start acting on insights.

• Ownership beats “somebody should”: assign real humans to stewarding definitions, sources, and semantic models—otherwise everything drifts.

• Don’t skip the semantic layer: strong modeling and governance (Power BI semantic models, Fabric patterns, and documented measures) makes both humans and AI more reliable.

• Use AI for assistive workflows, not authority: summaries, documentation drafts, and exploration are great—final decisions still need human context.

• Incentives shape behavior: reward teams for fixing root-cause data issues (and for using trusted datasets) instead of celebrating one-off hero reports.

Looking Forward

Pick one high-visibility metric in your organization, document its definition and owner, and then use that as the “golden” starting point for any AI-assisted reporting or analysis.

Episode Transcript

0:31 out welcome back to the explicit measures podcast with Tommy Seth and Mike good morning everyone good morning sorry I had to rearrange the images there everyone’s everyone’s shifting around happy Tuesday there it is a that’s rough on the ears I I would it would have been mean I I would it would have been shorter but mik kept interrupting it so I couldn’t like that’s just not allowed stop halfway through yeah start start start upop start up main topic for for today is we’re going to be

1:01 for for today is we’re going to be talking about the AI and data culture problems there’s an article that has been written by Forbes the the title of the article is why AI isn’t going to solve all your data culture problems which seems right yeah agree but AI is helping out with things so where is the right place to apply it so that’ll be our main topic for today but before we do that there’s a couple let’s call it news articles Let’s do let’s do the news Tommy anything else that came out on the Microsoft blog or fabric blog that is worthy talking about oh maybe

1:31 that is worthy talking about oh maybe gamechanging well I don’t know but big actually update to for fabric and updating and uploading your CSV and Excel files there’s a new data connector that actually allows you to do direct import of your Excel and CSV files directly from one drive so you have a direct connection star schema detection and relationship detection which is very interesting especially when it’s being loaded into your report it looks for any existing relationships between what’s an Excel and fee and applies that behind in your powerbi

2:03 applies that behind in your powerbi report I’m going to have to play with this so interesting that they’re having some smart schema detection and also looking for interesting they already have relationship detecting for powerb desktop like when you bring in a new table it’ll try and find names of columns that match and turn off relationships there so I usually don’t let that happen inside desktop but I I see that they’re mirroring the same features from

2:33 same features from desktop where where’s the where’s the feature for don’t let users put in bad data into columns that are not the right column like where’s that so I like Excel files I like pulling data and I think a lot of the newer users to parbi starts with flat file dumps and starts loading data in through files but I very quickly feel like we get into a problem where the Excel files change the the powerbi report breaks and we have all this inconsistency and a lot of extra rework

3:04 inconsistency and a lot of extra rework around trying to make and Harden the process on how Excel gives data back to powerbi I don’t see anything in this article talking about like the hardening aspect of oh by the way we will check your Excel file and make sure none of the columns change from when you loaded it last so it doesn’t break your entire process so that’s the feature I would like to see in this spit disruptive too go ahead that’s quite a feature ask when realistically all we’re doing is enhancing the report creation

3:34 doing is enhancing the report creation experience in the service isn’t that what isn’t that what this is it’s basically just enabling things you can do in the desktop now in the service you mean one of our most popular articles from harbi tips tips blog is how to load files from SharePoint it is like literally number one for like years since since it was written by since it was written it has been since day one because you you work with an Excel file or a flat file on your computer and you’re like well now

4:05 your computer and you’re like well now what where do I put it now and where I need to go pick it up and there’s this whole concept of every time I do a training session it’s like okay today our example is we’re pulling data from the local machine what we should really be doing is putting these in SharePoint or a place where they’re accessible to powerbi. com and then you can use them and everyone’s like what I’m like okay so it’s a bit destructive too because this is all great for powerbi but then Fabric’s coming out with articles and they’re pushing the one Lake file explorer to push your files

4:37 Lake file explorer to push your files into one Lake right so and these are not at all this this is where this is where guess what there are multiple teams working on this yeah you a single product you give me your file right we already uploaded the file these are just options guys you can do it over here or you can do it over there true except it’s it’s not the same route though because this doesn’t get added to one lake at lake at all this is just to a single just gets

5:07 all this is just to a single just gets pulled into a power you have to do that to there isn’t you have to have some capability to do I would like it to have all three options like out of the gate right it should say one drive SharePoint and then it should be say they’re saying one drive everything’s on one Lake already isn’t it not this well and they should have another op just get it from one leg right I that would that would make sense it’s just the third option there anyways I’ll have to play around this one a bit interesting build here

5:39 this one a bit interesting build here I’m not sure how I understand the star schema detection smart schema yeah I I that is interesting to me so I’m I wonder if that is if it looks at if you had Dimension tables and factual things inside your Excel file which I’m not sure anyone’s really that anyway are you are you reading that like it’s in the model already I I think that reads to me just like a valid data validator it says our our connector

6:10 validator it says our our connector automatically detects the schema of your the schema of your Excel and CSV files including data types and structure to me that’s just a file CSV file oh so it’s not actually star scheme I me it’s not saying hey I found a dimensioning your report that looks like this I don’t think so that description seems a little misleading then the more the more I’m going through this I’m realizing the Persona here they’re going after is not you or you or myself yes this would this would be the data scientist and

6:41 data scientist and person that was a joke Tommy that was a joke okay I was about smart schema detection just reads like the error detection that we already have yeah like showing you the distribution of like things that don’t align to your data type which is valuable right up front I I would agree it just got to be clean got the title in my mind makes different sense to what they describe there question do you guys turn on or off the autod detect relationships in

7:11 autod detect relationships in powerbi autod detect relationships always off always off yeah and I all my training too so but I think that’s because our personas if you’re a consumer or self-service you don’t know what you’re do and you’re going to go great and so this is where I think this is more consumer facing I disagree I think there’s two features that always should be off and again Microsoft I think is trying to make it easier for new users in reality what really happens is it actually complicates things more and people get confused as to why did that thing automatically show up and I and again from a from a sales perspective it

7:42 again from a from a sales perspective it looks great like hey it’s going to autod detect a relationship it knows there’s two columns it sees the data inside the columns it says oh I’m going to make a best guess that these should be related okay seems like a very valid assumption when models get more complex though other than like three tables or more I feel like that’s count is somewhat falls apart it falls apart instantly when you import three different data sets and all of them have an ID column but that’s not what you want to join them on it’s just ID right and I

8:12 them on it’s just ID right and I think valid like valid UPF front maybe to show folks how things relate and what you have to build it quickly quickly evolves into you making changes to a report and it automatically changes something behind the scenes that you didn’t expect it to cuz that happens a lot a lot yeah I always that’s that that causes more wasted time than anything because you’re

8:42 wasted time than anything because you’re why is this not working especially when it’s INX calculations or things like that oh gosh here’s your million dollar here’s your million dollar idea on this one Seth just basing your idea there they should have instead of it autod detecting the relationships it should walk you through a we found found three columns that may be related here here’s the relationship view we we pre-selected these columns yeah here’s the here’s what I want to change yes or no yes here is the things that I’m selecting for you

9:13 is the things that I’m selecting for you we recommend this as the thing dis does this make sense to you yes or no approve like I’m I’m the data load we’re gon to change your relationships do you approve schema green red do you want to change yes yes no no and then you can guess what then you can track how useful that is then people are like yes or no yes or no right it’s data back to the AI to train itself there you go there you go more things for co-pilot to

9:44 more things for co-pilot to consume just just show me show me what you want to change before we’ve actually haven’t been doing this podcast for the last 100 episodes this is actually a all AI generated it’s all this is actually co-pilot we said act like you’re an Italian from from New York that moved to Florida and loves powerbi mooved to Chic yeah move Chicago

10:04 powerbi mooved to Chic yeah move Chicago he has no idea where the identity yeah no I love I love powerbi guys comment here autod detect relationships is the slightly less evil twin of Auto date time that’s that’s the other one I was going to go after so that was my next gripe was the always turn off auto date time and yes I understand it does help does it but it does help for new users I definitely think it definitely helps make it start of month for sure very quickly very quickly do you outgrow

10:34 very quickly very quickly do you outgrow that date Table and there are other questions that come out that are like well we should probably have an index column for year month and day and we should have to your Tommy to your point start of month columns and all these other things again going back to like wouldn’t it be nice if there was like an auto setup screen for this hey we’re going to give you an auto date time calendar on these date tables or these date columns which ones do you choose like let me pick which ones I want hey by the way we already know how to do all these other things related like dates are very wellknown

11:04 are very wellknown hey how about you just tell us do you want holidays included do you want like this thing included do you want start of month like you should just literally have a list of things that like tick tick tick tick tick you just check them off and boom uploads the the one table on the one date thing that you need or multiples right there it is now it’s all consistent I agree and when you’re having that conversation in every one of those questions right beforehand emphasize it with hey hey do you want hey hey I think that just reinforces the point yeah like and that’s a Mike De

11:35 point yeah like and that’s a Mike De Carlo ISM hey is for horses I think he said of the Clipper we need na’vi going back to a little Zelda conversation oh my God love that EMP power we we need clippy for powerbi which has been renamed co-pilot the fairy from Zelda oh my gosh the one that would prompt you every time you tried to walk yeah gotcha yeah sorry am I being too nerdy out here no the newer ones the newer ones sorry we’re a little bit we’re on the on the new tears of the Kingdom Z breath of the wild

12:07 the Kingdom Z breath of the wild here breath of the wild yeah I need to catch up na’vi well I don’t think so now you have the this never mind we all flying we’re GNA di that when Tommy when your kids are old enough to get a Nintendo switch you too will learn about all the things so two games you’re going to love or I think I think your family will enjoy will be like the Minecraft and will be the Zelda the Zelda franchise so those are two really solid good ones we had a just as a as a

12:37 solid good ones we had a just as a as a side family note we had a whole family like well except my wife because she doesn’t play Minecraft but everyone else loved playing Minecraft and we were all everyone had their own screen we were all playing Minecraft together and it was like this is great we’re all sharing diamonds back and forth and it it was wonderful it was a good time yeah we did get a switch so I’ll have to introduce that quick shout out to my to my two daughters and then we we’ll keep going they saved their money on their own they said how much is this switch and we’re like well these for Christmas way more than you have yeah exactly but like hey if we put

13:08 have yeah exactly but like hey if we put our money together from Christmas could we get one and I was like oh and I like okay but that means no this or this or this they said that’s fine we’ll we’ll put our money together so shout out to planning ahead which I was definitely not doing at 7 years old teach them well Tommy teach them that’s it teach them like you got to make decisions you got to choose one or the other pick what you want that’s really good our our kids did a very similar purchase they saved up for something and I’m like okay well you saved up for it I told you what the number was you did it you did the work

13:40 number was you did it you did the work all right we go get you the thing excellent that’s really good news wa to go parenting win for Tommy Ching all right let’s move on to our main topic today because everyone’s loving to talk about our personal lives all the time so we’ll jump into our main topic today so today the article is why AI is not going to solve our data culture problems I’ll go snag the link here for everyone to go follow along as well if you’d like to check out this link and read along with us as well go grab the link here from Forbes

14:12 well go grab the link here from Forbes and I’ll put in the window Tommy give us a little overview of this article and where are we starting for the conversation for so for survey it’s actually Brent dkes who does effective data storytelling that book article Bunch articles from and it starts with this data executive culture ser surve that was conducted and from 2023 to 2024 the number of companies that reported they had a datadriven organization Rose from 23% in 23 to 48% in 24 and the number from data analysis

14:44 in 24 and the number from data analysis culture from 20% 23 to 42% so large shifts just in the last 12 months on what they’re seeing on organizations reporting data culture the PE the people who conducted the survey said well what could this be and they said it’s the confidence in the generative AI now Brent Dykes is actually saying is that the actual answer here that was a speculation on what was the cause of a a giant spike in these numbers and basically Brent Dykes is going through

15:14 basically Brent Dykes is going through thinking you’re a data driven organization what AI can do to help you think that but actually being a data driven organization and really what are the effects and the outcomes of that there’s a lot of things going on here yeah and this is and I’ve said it before and and maybe this is part of what I feel like it’s going to be happening right with the Advent of AI you have more need to know where

15:44 AI you have more need to know where truth is what is that truth thing because the AI is going to be able to produce whatever it thinks it needs to produce to answer your question it’s going to be able to say things that sound really really real but unless you can verify or there’s the AI that you’re using for example I I’ve been using copilot a lot more now I really do like the fact that when I use it it provides references here’s the some of the pages here’s some of the web pages I used to go find the answer for this thing that was helpful for me and then I can read what it said and then go to

16:14 can read what it said and then go to those pages and verify myself do I trust that information but to your point Tommy in in the internals of a company where is that AI generating its data from where is it it needs to be whatever you’re doing we have to have AIS that are like reference the content and saying oh I found in this report with these filters this insight oh interesting let me go look that up okay well that’s a a nuanced answer not sure that’s right like you there’s G to have to be some more like people in the middle here to

16:44 people in the middle here to figure out how do we know this is right and I think a real quick note too here when we’re talking about AI I think especially from an organization point of view we’re not just talking about chat GPT or even the independent or individual co- pilot a lot of organizations right now I’ve been able to talk to they’re working on their own co-pilot or their own basically AI for the organization some of them haven’t released it yet so we’re not just seeing AI as in chat GPT or AI as in something

17:14 AI as in chat GPT or AI as in something on your local machine or just you and I Mike with our individual companies with a co-pilot these are AI That’s integrated with the organization and I think it’s important to think about that way of AI as well when we’re talking about this conversation yeah so one of the CH I I do want to start by talking about that graph right and the reason why is like I I don’t have enough details like why what what question are they asking a bunch of questions to the

17:44 they asking a bunch of questions to the survey that are like causing that Spike or is it directly like Hey how do you rate your data culture and it’s just a direct percentage because obviously they’re coming out of that with a the report authors basically in the article says we can think of no other reason for these momentous improvements in culture and and confidence both right then the Advent inv visibility of generative AI so like the the challenge around like looking at

18:15 the the challenge around like looking at a graph like this is that is it because they’re looking forward to what AI can do or is it a representation of this is my data culture right now or I have a data an data an organization like what and and the reason why this is an interesting point to me related to like and the counter arguments made in the article is there’s regardless of reality I think there’s certainly a perception

18:46 there’s certainly a perception shift of where organizations think they’re at and I think Brent does a really good job of breaking that down but what do you guys if this is a false perception if we’re going from 20s to 40s in a year and all of the sudden organizations believe that because they’ve turned on AI in certain capacities that all of the sudden they just skip to the front of the line in terms of data culture and

19:18 terms of data culture and initiatives what is that going to do for current state objectives and and I think he brings that forward but like what are what are your thoughts around that I think it’s important to focus question I’m G to focus on one of the words you said and I think it’s the perception of organizations and people perceiving that they’re more data driven and rather than the actual whether it’s the fact that organizations are being more data driven those are two very separate separate

19:49 those are two very separate separate conversations yeah but are but are they because and I think I think what’s fantastic about this article is Brent is pointing pointing Parts out of this out but it’s comprised for this dissonance between a a cautionary tale right

20:08 between a a cautionary tale right because the lead of this article is and I just want to like knock through is the when talking to Chief data officers like he’s using some of the output of an AWS AWS survey around like ad CD CDO agenda in in 2024 and the main call out points that he has are 70% of cdos say difficulty and changing organizational behaviors and attitudes was the most frequently

20:39 and attitudes was the most frequently mentioned mentioned challenge 59% said their largest problem is absence of data driven culture or datadriven decisionmaking 50% say the lack of data literacy or understanding and and ultimately the outcome is saying that survey the survey and interviews leave little doubt that it is necessary for cdos and or that it is necessary for cdos and their organizations to flourish

21:12 that doesn’t paint a fantastic picture of like we’re in a great spot with data culture and our our people knowing and making datadriven decisions but we just spiked in the largest we ever have since 2019 is it all false flag or or are we seeing the results of a lot of work that’s being done across organizations and we’re starting to make the the the right steps this this feels like so initial

21:44 steps this this feels like so initial thought here is two thoughts around this one so one question you asked earlier said that we didn’t really get an answer to which I think was a very relevant observation here is what is the because of the way this is handling or the because of the way this is being reacted to what are the potential downfalls of thinking you are in a place of data culture when in reality you are not and I think later on in the article it talks a little bit around like there’s a question that is given that is more difficult for people to answer the

22:15 more difficult for people to answer the harder question was when data says one thing and your manager says another how do you address the difference using like data to support your reasoning right so that’s a great opportunity to say you know a datadriven culture I think and again it also points this out in the article which I really do agree with a data culture encourages exploration and collaboration and discussion on difficult things and I I really do think looking at the data should conform your your understanding of how the business

22:45 your understanding of how the business is running and we always walk into these situations with a preconceived Assumption of what the data is going to tell me typically we have to be open-minded and let the data tell us or explore the data to get to a place where we feel like we can produce an answer that will help us drive for actions that will then provide the desired outcome so all that to say to answer your question earlier what will be the outcome here I think we’re there’s a potential here for having many different teams walk down various

23:15 different teams walk down various different paths thinking they’re doing the right thing but in reality they’re getting further away from their ultimate goals or missions so I think potentially What’s Happening Here is if you feel like you have a data culture that’s more mature mature when you start trying to align to key objectives or key results for your organization you’re trying to drive sales you potentially could be spending a lot of effort in the wrong direction where it’s maybe adding some value but not the maximum amount of value and because you’re relying on the data and

23:45 because you’re relying on the data and analytics provided to you that isn’t supporting an alignment to your main you supporting an alignment to your main objectives of the organizations so know objectives of the organizations so that would be my my answer to that piece yeah to me starts with the the definition of if you asked all the people that they surveyed what your definition of data culture and how does it actually manifest in an organization I think there’s going to get a lot of different answers I think Mike to your point I think a lot of people when they say our data culture

24:15 people when they say our data culture has gotten better in our ability to work with data is because well I I’m more familiar to access the data now or more comfortable to go to a tool like powerbi and look at my reports it’s not as foreign to anymore I can ask maybe if AI is a big part of that well I have the ability and I can ask questions I have better access and ability to in a sense interact and touch and mold my data from a consumer point of view too that doesn’t mean data culture though and from an organizational point of view

24:47 and from an organizational point of view I and I think that’s where the misperception here is I can ask questions with my data and I have a better ability to work with my data in different capacities than just Excel therefore we have a better data culture but I think that’s very different from an organization actually having a healthy data culture yeah but data culture in and of itself is it it is represented by people making databased databased decisions right it’s not necessarily

25:18 decisions right it’s not necessarily that they know how to build all the tools or build all the reports or do all the analysis themselves it I I think it it drives it drives deeper into how are people basing their decisions right and he drives into those specific areas which we can towards the end of the article but I because of that right like we want to raate we want people to say the data is telling me this this is why I conveyed the information regardless of a situation we

25:48 information regardless of a situation we can dive into those later but when we talk about data culture and we just did an entire series on Tuesdays about adoption about all all of the things powerbi and data culture and adopting new technologies business intelligence tools and the benefits of that of the three pillars though people process process technology what are we constantly talking about all the time oh it’s definitely the technology oh for sure right no it’s more the it’s it’s the

26:20 right no it’s more the it’s it’s the people people right so what what what is striking out of this is if all the cdos ree as well because there’s a glurb in here that I love Brent’s spot on here I think he says each year respondents have consistently emphasized how culture people process issues people and process typically in the range of 80 than than 90% represent a far greater challenge to becoming datadriven than technical

26:51 becoming datadriven than technical limitations which is 10 to 20% so if I take that information which resonates with my head as well as all of the findings of Forbes what drastic change in how we interacted with people in process in the last year lit up a wildfire across everything the answer is I’m not aware of one yep so we got fabric so that was a good thing we flip the script

27:21 that was a good thing we flip the script and all of the sudden have a techn like and this is this is what’s striking right history repeats itself technology when put in front of people can’t solve all the problems that’s why you need people in process but amazingly enough technology becomes buzzwords to people who don’t understand it and they start saying that they have things that they don’t because they don’t get it and that puts everybody in the organization I think especially cdos or people that

27:52 I think especially cdos or people that are datadriven in a really precarious place or just not comfortable one because automatically all of the things that you’ve been working hard and diligently to build for the organization somebody thinks they just skipped to the front of the line and they didn’t the one the only thing I would possibly argue with is no there’s been no Catalyst or major change of organizations is I think every tooling

28:22 organizations is I think every tooling every platform now has or has had some type of co-pilot AI to make it whether it’s Google analytics or sale through or like confluent or any tool you can think of has some type of AI feature in it now and has had one and I think that’s including a lot of these analytical tools copilot just came out for powerbi obviously recently but everything else that people work in there’s some type of co-pilot and I think that’s to a lot of organizations to a lot of people making

28:53 organizations to a lot of people making them feel more accessible and making them feel like they can talk with their data so I would say if there’s one AI Catalyst there that would be it and I’m not suggesting that AI can’t get us there no but how do you like the project and and this is where it’s like the chart are you are you confident that you’re going to get there or are you there and I think that’s the big distinction because even in our conversations with about co-pilot about emerging AI Technologies Tommy in your experience like if do you think data

29:25 experience like if do you think data culture and data like people leveraged AI tools enough to see this representation and I think the answer is no I agree with that because every time we’re talking about our interactions with AI tools and data and things that it’s doing for us there’s severe limitations on it it’s if you don’t have a data catalog if you don’t have the data sources if you don’t have all the documentation if you don’t have all the things that are going to feed the AI what is the AI doing nothing it’s just

29:56 what is the AI doing nothing it’s just regurgitating things that aren’t relevant to your business and yes but I think it’s still making a little easier for people or the perception I’m I’m not suggesting that it’s not super powerful and it’s it’s

30:10 it’s not super powerful and it’s it’s gamechanging what I’m suggesting is it’s I haven’t seen nor experienced something in my line of business or my area and and our conversation that would lead me to believe believe that yep 100% I believe that we’re almost at 50% across all organizations in culture trends that everybody just instab instab boosted I I think potential is there but

30:41 boosted I I think potential is there but I don’t think we have and it like it’s still the same base Source not base the same Source things that need to be there for any data culture initiative I need data I need it Consolidated I need it clearly documented I need it like my people or the people in your organization to understand where to go how to access what tools to use it for and if AI shortcuts the tools part and they could just ask questions and they get the data and the answers that they need fantastic that would be a huge timesaver right but that also requires

31:14 timesaver right but that also requires and we talk about in the future individuals to understand how to use data and AI doesn’t do that 100 per. because I think to your point data culture relies on the dialogue and interaction between people it’s not just enough for my ability to talk with data my ability to access or manipulate the data it’s my ability in my boss my team to change our viewpoints to go over what our numbers are to agree on it and we

31:44 our numbers are to agree on it and we talked about this in a data literacy episode it’s talking with it reading it and arguing with data is a healthy data culture and by itself just having that tooling is not enough do you have that interaction with your boss do you have that interaction with your team on if the numbers were to change or to Mike you always say this and I I think it’s so a Paramount here if the numbers don’t align with my expectations or my viewpoints am I willing to change my own Viewpoint or do I disagree with the

32:15 Viewpoint or do I disagree with the data and I I think a lot of this to me where I’m leaning a lot of this point right here is I think a lot of this culture data culture piece starts with the leaders of your team it starts at the top what does that leader tolerate where are the objectives coming from one one note here that Brett makes is there’s a atalon at alons atalon the state of data culture maturity research report was produced you can the link is on the website I’ll put the link in here as well but if you

32:46 put the link in here as well but if you want to go download that you can but inside the report it it communicates a little bit around okay well because of that data culture 29% of respondents felt very confident that their leadership has a the leadership sees a link between investing in data and analytics and staying ahead of the competition I would hold hardly agreed we’re competing on analytics like that’s that is the one of the things early on when I was thinking about building programs and software like anyone can build a program anyone can build an app now the the app the

33:16 can build an app now the the app the barrier to building apps is becoming so low cost anyone can get one of these things done where does your differentiator come from as a company well it comes in how you collect your data what do you do with that data what comes back from that what are you doing to leverage that information in order to be competitive against other people and I think that’s that’s that is your your company’s strategic asset no one else can create your data period no one else can can can make that so in

33:47 no one else can can can make that so in a place where you can get more of that internal data made around your customers your customer base and figuring out what your Market looks like that’s very valuable to companies and I think companies will continue have to and this is where I see maybe potentially fabric playing a big role here is the data and analytics competition space is heating up five years ago powerb did not have any fabric experience to it and we were like wow look at this powerbi is great so

34:17 wow look at this powerbi is great so you’ve made the point I think in the past like it we can join all these data sources this is this is awesome there’s nothing going to compete with this and all of a sudden now we have lake houses and all this other new stuff and so we’re seeing the data and AI piece of things become very Central to our organization it’s becoming more and more important to organize what we’re doing datawise because that is our strategic asset moving forward and I agree like so one I’m all in I agree with I agree with Brent in the in the first parts right when we’re

34:48 the in the first parts right when we’re talking about perception of where where people think they are right versus where they are I think the there there’s a a clear line in here all I haven’t I haven’t thought about like the articulate question right but he he he has three different questions are you data driven everybody says yes right and then we go into harder in the past month have you changed your opinion Based on data well like how how involved and how engaged are you etc etc and then the harder one is like when when data

35:18 harder one is like when when data combats or is at odds with what your man mention earlier yeah yeah like are you are you pushing that forward and I think regardless of your systems the technology Etc that that is something that a a company could Foster right out of the gate which is listen if we have access to information within our systems and the outputs are telling us something we want you to tell us what you’re

35:50 we want you to tell us what you’re finding right like there should not be a Culture of Fear around like univ idual or raising up a a potential problem because you’re you’re hiding something and and ultimately ultimately that problem will very rarely just go away it just becomes a bigger and bigger and bigger problem to the point where it becomes a a much larger challenge for

36:22 becomes a a much larger challenge for the organization to solve as opposed to what whatever the euphemism are like just hey just get it out of the way rip the Band-Aid off we’ll figure out how to how to Pivot and we’ll move forward versus you just cut my leg off how do you expect me to run the race right and that that’s what happens with some of these data problems where you’re just hiding behind like somebody’s not going to like the message and I’m glad you said that Seth because I was going to do a hot take but you basically said in so many words hot hot

36:54 basically said in so many words hot hot take but honestly I think regardless of the tooling like with the people process technology for successful data culture technology is could be the least dependent or least crucial aspect of those two foster one because it doesn’t matter if you have powerbi or Fabric and you’ve have the greatest data analysts or Engineers Seth if you have that situation where I can’t show these numbers or we’re worried if someone gets fired or that’s not what my boss said I fired or that’s not what my boss said that immediately kills whatever

37:26 mean that immediately kills whatever you’ve done to get to that number in the first place and the technology at that point is irrelevant so from a data culture point of view be it AI or powerbi and you have your Enterprise if people are not willing to accept that above you and not willing to change that perception above you that immediately is the deal breaker towards any fostering data culture technology can’t do anything anything there I I like your your comments here Tommy and I’m thinking here I keep trying to write down a little list of

37:56 trying to write down a little list of what is the measure of good data culture right and I I’m keep from this article and what you guys have been saying it it starts with a good leader I I feel like there’s there’s a part of this is we want to foster a culture of Explorations and Discovery and and really leaning into the information I think you said something really relevant set there which was we don’t want to have a Culture of Fear so I don’t want to continue being in a place where I’m looking at data and I’m fearful what the results are because the results are

38:26 results are because the results are contrary to what my leadership expects and therefore I work hard to shape and conform the data into something that it’s not really there and Tommy you’ve been very big on this proponent as well you’ve been in an have experienced a number of instances where the business is so fearful for what they’re trying to do instead of pushing into the analytics and getting the real numbers people keep trying to filter out some things or do these other things and so there’s there’s a there’s

38:56 so there’s there’s a there’s there’s a fear there that is being displayed in how their actions are so don’t do a Culture of Fear one note that I have around that one is sometimes I hear teams bantering or relegating and this is typically comes from the it space where they say well that team’s just Shadow it or we don’t accept Shadow it here and to me that feels like a way of shaming some team into it really does It feels like a way of shaming a

39:26 mean it feels like a way of shaming a team into saying you you can’t do it we as the central bi or Central it team has the has the authority to make all the reports and if we can’t produce it everything you’re doing is irrelevant and not helpful that is a that is a great killer to data culture like that should not be said and I get annoyed now when I hear the word oh we don’t we don’t accept like when I hear Shadow it yeah I just feel like that is the wrong impression of how to think about our data culture that is not right it’s

39:58 data culture that is not right it’s completely opposite you say that and and what I think yeah like those are your best friends like these are the people you want to BR closer yes because they’re so close from like like they’re

40:13 they’re so close from like like they’re already doing tons of stuff yeah right they’re on their own great they’re on their own they’re doing stuff how do you guys make the best like you want to talk Big B business impacts like get up next to each other figure out where the handoffs or how you accelerate like the what they’re doing into a wider audience because it’s obviously already working for them yep right Andor like hey you find out that there’s some inefficiencies or there’s some data quality issues what make them more efficient why not and and to your point

40:46 efficient why not and and to your point so your point leads into my next point which was there needs to be better boundaries and providing trust and responsibility across teams if we have a central bi team that’s going to make models and we’re going to say we will trust this other team there has to be a mutual trust and we have to say look you’re not Shadow it I will build this centralized model here you go team you are now able to build reports off of it and you have to trust what they’re doing there if you don’t have that trust across teams again that’s another

41:17 across teams again that’s another indication of poor data culture in your organization sorry Tommy I didn’t mean to cut you off I just want to make that last point because I was following along that line of thought I think that’s a great point because with to your point with without trust what do you have there’s conflict I’ve seen this as being a data analyst where now every time you have a meeting with that department or with those group of users it’s almost a competition or a conflict where we’re gonna we have every time every time fight oh yeah and because again all of a sudden you’re opening up things that now they’re not owning even though it took them hours and hours or whatever their

41:48 them hours and hours or whatever their old process was and we’re now opening up this world of well here’s your numbers here’s the data that we’re actually getting and if it doesn’t J or align with again their past years and not willing to accept where we’re moving or where we’re heading in terms of how we’re going to access our data and then more importantly who’s going to have visibility to this because before everything was controlled that old way of I’m gonna put stuff in the PowerPoint we’ll have a meeting with my Executives I’ll put the numbers there done well all

42:18 I’ll put the numbers there done well all of a sudden now with all the tooling and AI anyone can ask a question and get a number that was previously unavailable to them to them or or again they can just access the report so we’re now if you want to have the lack of or if you have lack of trust you have conflict between the people who are accessing the data and the people who are creating the data data it’s I it I almost little I don’t want to say laughed out loud but it like your

42:49 to say laughed out loud but it like your tone of how you were saying like talking to the that business area I think potentially could be part of the problem sometimes because the subject matter experts in their business area likely have better datatime depending on like we’re talking well formed Etc as opposed to like the central bi team coming in saying like this is where data is and like the the problem with that perception is the central teams need

43:19 perception is the central teams need need to get up next to those Shadow it teams with the mindset that they’re wrong like Central it is wrong right I need to understand what that business is doing and and and how and I think as long as the tech Stacks are are unified like it’s really just a data conversation I do think that Mike while you were talking like I think there are going to be some challenges there potentially especially if you have like

43:49 potentially especially if you have like business units that solve their own problems by solving it with their own technology choices and the organization is hey gu what we’re doing this now that like that is that is going to be conflict right you’re going to have challenges that you’re going to have to solve because you have resources that either have to ramp up into new technologies or you need to find a way to adopt the new solution right like because that’s

44:19 new solution right like because that’s an organizational change that’s not a one Department telling another department that they can’t be doing something and that that I think is one of those conflict points where you do have organizations that have just chosen to be the wild west of solutions and Technologies is you have people that specialize in them and especially when you have sub subject matter experts data experts in business areas that are very proficient at different tool sets that

44:49 proficient at different tool sets that that could be part of the conflict in terms of like Shadow it and they do their own thing Etc you just hit the nail on the head you just hit the nail on the head and I would love to give a shout out here but the the biggest Catalyst I’ve ever seen at organizations or departments for the change in the shift of behavior and process is the leadership saying moving forward we’re going to use this platform or we’re going to be using this report this is where you’re going to get your number on moving to Quarter Two or whatever it was

45:19 moving to Quarter Two or whatever it was and if you see if you see an issue if you see a number please contact X Y and Z but that came from leadership those te shifts on behavior of departments came from the people with influence and that’s aren’t left up to the Departments to fight over right yeah so so what are two proven ways that Brent points out to build a data culture I think we just talked about the first first one well major leadership right yeah

45:49 one well major leadership right yeah leadership yes I definitely agree with that one and I’m going to I read another article recently by I think it’s Inc inc. com I think it’s like Incorporated and the article was talking a little bit about something that Jeff Bezos has banned PowerPoint from Amazon I wish all companies would do this I think this would be a good ide this would be a good suggestion so I agree with you Jeff Bezos and I think you are a smart man for doing this but one of the comments in this other article which was and I think this is one of the I’m trust me I’m getting back to what we’re

46:19 me I’m getting back to what we’re talking about here are your question Seth you in the middle of the article it says you you don’t want persuasion you want truth and so what Jeff basos has done in his meetings or his executive meetings or some of them anyways he says look people are very busy they’re running from meeting to meeting all day long and so he starts his meetings now with silence people come in and if there’s a meeting about a topic someone writes up a page a paragraph It’s a written dissertation or written thing of

46:50 written dissertation or written thing of here’s the problem here’s what I’ve analyzed and here’s how I think we’re going to have to solve this it’s literally it’s it’s like a written document and so the idea is the goal here is we want everyone to read the memo the the piece of the literature basically and think spend 15 minutes just thinking about what you’re doing what’s going on give the space to be able to do that and so it’s it’s PowerPoint tends to feel more like a sales tool you’re you’re putting

47:20 a sales tool you’re you’re putting pretty pictures and you’re distracting from some things and so if you take if you strip all that away and focus on what is the truth we’re trying to get after what does the data tell us and I think that to me again I would say Amazon is a pretty good Beacon for data and analytics they’ve literally built an entire business on like a Distribution Company turned data analytics company and they’re huge now for all that stuff so to me it feels like a lot of what they have done has worked for them and so there should

47:50 has worked for them and so there should be some lessons learned there where we focus more on the culture should focus on more what is the truth we’re trying to seek and giving people ample time to digest that and think through that with the data and reports that we provide to people I’ll pause there no this probably worked too because Jeff has probably invested his time not just we’re gonna do this and going back to the the article like from a leadership point of view it’s not enough just to say sure

48:20 view it’s not enough just to say sure we’ll do this and we’ll go this direction it’s the investment of leadership with their time not just their money on any initiative from a data culture point of view and I bet Jeff is probably in those calls going no 10 more minutes five more minutes and he’s there at the call too not just sending his own email then moving forward with whatever he wanted to do it’s that example of leadership and that example of taking that initiative and time spent towards we’re gonna actually do these data governance issues we’re

48:50 do these data governance issues we’re going to work on these data literacy workshops and I will be there like to to me that’s that’s the example because it’s more than just saying we’re going to do this I’ll send out an email good luck on your own no that I need to see that example from my leadership and from that that sponsor with their time I always like to disagree with Mike and there’s a CounterPoint in the in the chat well that does speak that does speak to just this and it’s not an argument right is the chat that that

49:22 argument right is the chat that that works with Jeff Bezos right powerbi guy points out that not everybody consumes information the same way and like in in general I’m summarizing his point yeah that’s fine so so sometimes like in the same way that we talk about like do people consume things by video or reading or listening right better and I I I agree with Mike from the standpoint that especially in

49:53 the standpoint that especially in even where where I’m like now spending my time you spend an inordinate amount of time on building Powerpoints and could I spend a fraction of that in a well summarized synopsis of why we’re talking and and what it is we’re meeting on that would take me a fraction of the time and be better than a PowerPoint yep it it would be so are you

50:16 PowerPoint yep it it would be so are you like there’s a Time Saver there there’s an efficiency there’s a getting to the point showing some data like rather than presenting something and convincing people right I I do like the idea of the format whether it works for everybody I don’t know and again the format I think is actually meaningless honestly I think that I think the main points here are people need time to think about a particular challenge you want to bring smart people into a solution and say what should we be thinking about that to

50:48 what should we be thinking about that to me that is very relevant and then the other half of this is driving for the understanding of Truth these are two qualities I think that are very valuable in data culture can you do you have have you given yourself space and time to think and are you understanding truth and does your leadership allow for that if your leadership doesn’t allow for those things to be happening I would question what are we really doing and are we are we just trying to produce reports and produce data to make someone happy or are we actually going after Where Do We Go From Here what is what is

51:20 Where Do We Go From Here what is what is the right decisions to be making and I think I think that’s going to be helpful yeah I I I’m I I I’m actually somebody who would say I 100% agree I’d love all my meetings to be that way and maybe I will change them if I run them like because I I do doesn’t hurt to try if nothing else try it out for a little bit and start with something that’s more reflective maybe in the nature and maybe that’ll work again I don’t think it has to be exactly that pattern that is what is what they have found that works for them the way their culture is but again

51:52 way their culture is but again the the principle should be no matter where you are it should be allowed allow time for thinking and allow finding for truth yeah and I think so Brent Brent like finishes up the article with the other aspect of tried and true ways in which to build a data culture is ample resources and I think I I commented on this a lot through the adoption Tuesdays right where it’s like man we’re making an assumption here like we got a team now all of a sudden really and we’re engaging oh we’re engaging with the business then we have deliverables we

52:23 business then we have deliverables we have actions we can provide back to them all of which require the organization to put their money where their mouth is right if you want a data culture and you believe that’s the most important thing that shouldn’t be a one-off hack job in every business unit right you can’t build a data Culture by picking one or two things to go attack it’s great that you are it might improve the area of that business but it’s not doing things holistically so if you want to say you have a Chie a data officer like a chief

52:54 have a Chie a data officer like a chief data officer and a data culture you have to take in mind the entirety of the data organization all of the things required and how you’re going to support the people and process part of that because the technology implementation is the easiest we can say we have powerbi but that doesn’t mean people are using it and I think ultimately like that that is going to be one of the biggest challenges anywhere is if if all of us collectively want to be have a data

53:24 collectively want to be have a data culture want want people to be doing datadriven decision-making you need leadership you need the guidance you need to have those things in place and in part bringing up all of these other people who don’t know how to use data teaching them and then amplifying the people that do know how to use data and giving them better and more skills and Technologies to create more efficiencies throughout an organization and it I I agree with him I I love the article I

53:55 agree with him I I love the article I think it nailed all the points and I think really shines a light on a potentially not dangerous but like another area that we deal with that we’re going to have to be talking about like it’s not powerbi is taking over the world it’s oh all of a sudden AI just skipped all of the the leg work that still is required for organizations to do I hope that’s where we’re heading too because I think we’re still at a point with a lot of organizations where there

54:27 with a lot of organizations where there is this incredible lack of established data communities or that data literacy leadership and again whether or not you have a CDO and if you don’t there’s still well who is actually this role and I I don’t think that’s really been established yet but I think that’s where we’re going how many organizations have like a community team right who put on those events at the end of the year or put on all the the the cultural team and in that same fashion I think that needs

54:57 in that same fashion I think that needs to happen if we really want to see this effective at organizations for data again you may have your Coe you may have the bi team but that doesn’t mean they have the authority or the resources both from a budget and people point of view to pass things on and Implement changes right we’re we talked about that with the implementation Tuesdays but that doesn’t mean that they’re companywide or that they have the reach that they need

55:29 that they have the reach that they need and I think that’s not established roles I think that’s a huge part where hopefully that’s where we’re going to see in the future where oh wow we actually need to invest in roles for data literacy not just someone’s extra cular activity or they’re they’re volunteering for this after hours at work but really spending and focusing that time on how are we growing data literacy with an established team or person I I’m my final thought here is I’m going to probably lean on the very end of the article which I found

55:59 very end of the article which I found very very relevant how do you Pro how do two proven ways to build data culture and I think he does a good job summarizing those two points it starts with leadership providing people with the this is a direction here and then it ends with the second point which is I’m scrolling forward here providing ample resources training time and so I’ll speak for this personally this is this is Michael’s personal experience my personal experience around

56:29 experience my personal experience around when I found data to be very valuable I have I have I feel like I have mean I have I have I feel like I have decided at some point in my mechanical engineering degree I decided data is so important that I will shift my career to over over into data and so if I’m looking at me from a company perspective right I want individuals in my company who are pushing for data trying to learn the new new tools they’re getting in they’re learning power query they’re taking self initiatives I’m potentially

57:00 taking self initiatives I’m potentially sponsoring them a little bit from a company perspective but I as me personally I was spending so much time outside of work thinking about doing investing in learning and educating myself around data this has pushed me into that career now so I really do feel like if you if you Harbor a good culture of people who are enjoying what they’re doing and they’re finding the value and the success uccess with the data in their company potentially those people will even spend a little extra time

57:30 will even spend a little extra time above and beyond investing in themselves so another outcome of this is look at your people are they are they hungry for learning more about better data things are they spending extra time above and beyond are they asking for more opportunities to learn about these things that’s another to me another measure of does the data culture support encouraging your team to be more data driven right does do that does that help so I like those things I think ample resources is a really good one there’s not enough out there we need to

58:00 there’s not enough out there we need to talk more about it the tools and the technology are always going to be exciting it’s going to be fun but we need to supplement that with learning and helping our our teams our people right we want people who invest in themselves just because they know it’s valuable that to me sounds like it’s it’s a it’s a good story to tell always right with that I really appreciate this conversation wow this was I feel like this is I don’t know if you if you can quantify that I feel like this was a thick conversation I don’t know if that I don’t know if that’s a right term you could hear but I feel like this was a

58:30 could hear but I feel like this was a thick topic really liked it thought this was really really good with that I hope you enjoyed this conversation as well as we continue talking about the intersection of like AI powerbi data fabric all that culture related things as we keep pushing around what does this mean for us what does this look like for your organizations and pulling from our experiences we worked with a lot of companies we’ve had a lot of experience in this places hopefully this is helping you think about these topics as well we really appreciate your time we thank you so much for listening with us we we

59:00 you so much for listening with us we we enjoy you being here we love the chat chat has been awesome you’ve been asking great questions and adding good commentary here and Seth’s always teasing me in the chat so you can always see that as well super fun and you can always get some good laughs out of it as well with that we only ask please share the podcast with someone else if you found this valuable hopefully somebody else will too Tommy where else can you find the podcast you can find us in apple Spotify or wherever get your podcast make sure sub subcribe and leave a rating helps us out a ton also maybe share with five other people if everyone does that we’ll get five more people do

59:31 does that we’ll get five more people do you have a question idea or topic that you want us to talk about in a future episode head over to powerbi tips podcast leave your name and a great question and finally join us live every Tuesday and Thursday a. m. Central and join the conversation on all of powerbi tips social media channels thank you all very much and we’ll see you next

60:19 [Music] n

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.