Low Code Solutions – Ep. 212

If you’ve ever heard “Power BI is no-code” and immediately thought “so… what about DAX?”, this episode is for you. The team digs into low-code vs. no-code as a spectrum: UI-driven experiences that generate code for you, and the practical reality that customization still means expressing logic somewhere.

In Ep. 212, Mike, Tommy, and Seth use familiar tools (Power BI, Power Query, Power Apps, and Power Automate) to unpack what actually makes low-code valuable, why “no-code” is often marketing shorthand, and how better guided experiences (wizards, pattern catalogs, and AI-assisted modeling) could help more people build correctly on day one.

News & Announcements

• Why Your Organization Needs a Code AND No-Code Option — The jump-off point for the discussion: why orgs need both “pro dev” and “citizen dev” paths, plus the governance/enablement decisions that keep speed from turning into sprawl.

• Microsoft Build — The team flags Build as a “pay attention” moment for analytics and AI announcements. Even if you can’t attend live, it’s usually worth bookmarking the sessions and recap posts afterward.

• What is Power Apps? — A clean overview of Power Apps and why it’s a textbook low-code example: design the UI visually, then add lightweight expressions for behavior.

• Official Microsoft Power Automate documentation — If you’re thinking “workflow first,” this is the best place to learn triggers, connectors, and the patterns that keep flows maintainable.

• Power Query documentation — The episode calls out Power Query as one of the strongest low-code experiences in the stack: click-driven steps that generate M you can inspect, learn from, and eventually edit when the UI isn’t enough.

Main Discussion

Low-code and no-code aren’t a binary switch; they’re a tradeoff curve between speed and control. The team’s framing is simple: the UI can do a lot (especially for repeatable patterns), but the moment you want something specific—modeling choices, transformation edge cases, or calculations—you need fundamentals. The goal is to get users to value quickly and nudge them toward the right defaults.

Takeaways to apply right away:

• Treat low-code tools as accelerators, not replacements: they help you move faster when the “default path” is good and repeatable.
• Expect “no-code” to still have a logic surface somewhere (formula bars, expressions, configuration)—it’s often just code hidden behind UI.
• Power Apps / Power Automate fit the “build by design” model: drag/drop UI + simple expressions can produce real apps and workflows quickly.
• Power BI can feel no-code in the report canvas, but serious work still depends on data modeling and DAX, which are inherently skill-based.
• Power Query is a great learning bridge: UI steps generate M, and inspecting that output teaches you how transformations are really composed.
• “I didn’t type anything” doesn’t mean “I can ignore fundamentals.” Relationship design, implicit vs. explicit measures, and model shape still drive outcomes.
• Better wizards would help: guided relationship setup, measure scaffolding, and sensible default formatting/foldering could reduce beginner mistakes.
• Pattern catalogs (think “Quick Measures,” but extensible/community-driven) could make good DAX more accessible while still teaching by example.

Looking Forward

The most useful AI/low-code experiences will be the ones that generate inspectable artifacts—so teams can start fast, then refine and standardize instead of treating the tool as a black box.

Episode Transcript

0:29 good morning and welcome back to the explicit measures podcast with Tommy Seth and Mike hello everyone welcome back hello Mike how are you today hello good morning how are things with your world they are fantastic the word of the day is Lovelace by the way come on Kevin Greg come on come on Kevin you can’t do that that that’s not fair Okay Kevin in chat already is with a leading the witness happy Tuesday gentlemen happy Tuesday we

0:59 happy Tuesday gentlemen happy Tuesday we did have to say Greg common I I posted about the our episode today so for GR for Greg baldini the word of the day is love lace yeah yeah so we’re even in front of it we have word of the day

1:29 we should give away rewards for people who use these secret words in various ways ways honestly just share the podcast just share the podcast who don’t care it’s not hard we were just talking about you not hard we were just talking about how much the know how much the telling us like the podcast is so invaluable it’s adding so much value to people like debatable [Laughter] by value a few Chuckles here and there my value I think people laugh at us there’s a lot of volume

2:01 by the way we have a tendency right in the beginning because we’re so happy to see each other to do the one thing that our listeners keep asking us not to do which is talk over each other all the time all the time no how often all the time all the time all the time all the time it’s really good when you have a podcast and you can’t hear anyone talking right yeah what was that idea what welcome welcome to another special episode

2:37 solid solid start again back to our earlier comment quality content here reiterate my point a couple major things coming up this this month there’s a couple big items here May 24th at Microsoft build is coming the the Microsoft power bi log log has made a little digital call out here saying that the the event hosts are Adam Saxon and Patrick LeBlanc from the guy in a cube channel it’s a live

3:08 guy in a cube channel it’s a live simulcast I’ve seen that word written a while yeah it’s it’s a chock full of great speakers and the keynote is analytics and the age of AI which I think will be awesome I’m sorry Microsoft done anything like this before like the build conference or oh yeah the build the build conference no brand new brand new new things oh yeah yeah that’s right do you mean like pushing so hard yeah the power bi side of things into build

3:39 build I would probably say I have not seen as hard of a push as they have done this year I don’t know if they if they’ve had made very specific call-outs because I want to Google the blog a little bit more yeah more yeah but I don’t think they’ve made a very clear clear delineation or as such a bold announcement on the Microsoft blog power bi blog that is a round build yeah you bi blog that is a round build yeah what’s unique about this one too so know what’s unique about this one too so for for listeners that might not be tied into all of the Microsoft folks like we are true there are a lot of leaders and

4:10 are true there are a lot of leaders and a lot of push around big exciting announcements and all this coming up plus doing a simulcast on the first day yeah my guess is that there’s you’d want to tune in and you can on the 24th for free at 9 00 a. m yep so the link for the Microsoft build conference is in the in the chat window and it’ll also be on the the notes so the description of this podcast as well so if you want go head out there you can

4:40 so if you want go head out there you can there’s a link at the very beginning the article where you can register for the Microsoft build conference we’d recommend you do it check out some of the sessions I did find it quite interesting or you I did find it quite interesting or they have know they have it it’s interesting to see what Microsoft promotes because there’s on their little thumbnail for the video that they put on there there’s Patrick Adam and Arun so the director or VP of something is this a VP of data analytics well he’s the owner of

5:12 analytics well he’s the owner of most of the like the Power Platform stuff isn’t he yeah but he’s like a very high up Microsoft person oh yeah yeah yeah and then there’s there’s Adam and Patrick not that Adam Patrick aren’t high up there but it’s like if you look at like the the pay scales and like so people recognize Adam and Patrick like they know like that is power bi so they put them there front and center which is I think very funny or interesting to me as well after this oh look it’s an it’s an announcement oh and Patrick and Adam are there surprise called marketing assets marketing yes you’re right that’s exact value add value add value so I

5:44 exact value add value add value so I loved it in the audience it’s going to be a good session they’ve Adam’s and and Patrick do a great job of producing all the guy in the Cube videos very professional very well done they’ve been doing this for quite a long time too very early on if I I think most of our listeners have had the opportunity but if you haven’t had an opportunity you’ve seen actually Adam and Patrick present as well like they do a great job on the podcast or I’m sorry on their show show but it there’s the Dynamics between them in a in a presentation is always fun to watch yeah I agreed so that’s one

6:17 fun to watch yeah I agreed so that’s one big announcement that’s coming up and another announcement that’s probably equally as impactful sure probably definitely yeah you definitely want to pay attention to this one as well so well so we are doing a special episode so Tommy a little background here Tommy got really bent out of shape when we did a special s episode with Charles Webb around just I don’t know data Marts and so Tommy Tommy got really bent out of shape because what happened is we did one on a Wednesday and then the episode

6:49 one on a Wednesday and then the episode numbers shifted a day and so now again I don’t know which one’s which so the the even numbers were now on Tuesday versus Thursday and this is just rocked Tommy’s world he still hasn’t let it go every time in one of these pre-calls Tony says what are we gonna get the numbers back to normal Tommy wants the numbers the even numbers are on Thursday the odd numbers are on Tuesday that’s how the podcast started it’s hot should always be somebody’s writing the agenda and putting these on podcasts and every time I put the numbers wrong please yep so and by the way just for the record this

7:21 and by the way just for the record this is not anything to do with Charles Webb Charles Webb well it was great oh yeah there’s just a special episode but we should have done we should have just not in label we’re trying to keep it part of the podcast so we had we wanted to label it or have a number with it so that’s why it it threw off the numbers well we’re we’re fixing the issue so we’re gonna we’re gonna resolve this problem so tomorrow tomorrow I think it’s around noon I have to double check I think it’s around noon we’re gonna do a special episode and we’re gonna do a special episode with Emily Lisa another senior product

7:52 Emily Lisa another senior product manager from Microsoft and we’re going to talk about editing your data model in the power bi service this is a new feature that’s been recently released it’s currently in preview I think it’s very exciting and it’s it’s very remiss of like our very first episode so if you haven’t listened to our very first episode I went back and watched it oh boy is it rough the audio is the audio is different it was just it was still good I think we I still enjoyed it and I laughed and thought it’s interesting to

8:23 thought it’s interesting to listen to but definitely we’ve matured a little bit since the episode one of the podcast this is you episode one of the podcast this is this is the same thing though that know this is the same thing though that we we talked to anybody who’s doing anything whether it’s a Blog that’s true whether it’s speaking you have to start somewhere yes everybody just assumes that you can pop into this space where you’re just amazingly good at all the things and it’s never that way so it’s hilarious that we we put ourselves on video that you can go back and be like well we’re no longer talking

8:54 well we’re no longer talking ourselves so but that’s what I I know we’re just one at a time like the flow is so good that’s I actually enjoyed that with really long podcasts that I haven’t heard about before because I you’re you go along the journey with folks too where it’s like ooh I really liked you and now you’re getting better and better well I don’t know if it’s getting better and better the quality maybe maybe maybe the production it was rough I’m just trying to whatever [Laughter]

9:25 [Laughter] the content is still mediocre like content has not changed the production maybe it’s gone up just a tick so anyways yeah the thing I love is this was not not only remiss of our first episode this was our first episode was Data modeling in the cloud which was obviously not a thing in 2021. by the way good timing on everything because in just two weeks we will be at it officially two years of doing this there we go so I love that we’re talking about

9:55 I had my opinions about this two years ago and I still have my opinions about this two years from actually today so I’m going to be very intrigued to where the questions go and where the answers go with Emily so I think this is also a good opportunity Tommy for your point right when we when I re-listened to this one one of your main sticking points was desktop for powerapps disappeared there’s there was a desktop version this one here look I just can’t I can’t see

10:26 here look I just can’t I can’t see myself removing myself from Power bi and

10:28 myself removing myself from Power bi and having the desktop application disappear so that might be one of the good questions we ask to Emily and see where where some of the Microsoft thought is around desktop where does it fit and how does that feel long term so I think it’ll be a good good conversation regardless so stay tuned you’ll see something come through on YouTube and all the social media channels we’ll get that scheduled today just want to confirm all the things we’re still happening and everyone’s okay with the content and we’ll send out an invite it’ll be probably on LinkedIn and YouTube and stuff like that

10:58 LinkedIn and YouTube and stuff like that as well so stay tuned we’ll have another episode tomorrow just in case you didn’t if two times a week wasn’t enough for your your list your ears if you if you if you just had a Hank Green on a Wednesday and potentially you really just disagreed with the number the numbering system on how we how we the the episodes of the podcast we can we’ll fix that tomorrow or the actual value in that we have a senior product manager from Microsoft talking to us like exactly that’s not bad yeah

11:28 to us like exactly that’s not bad yeah exactly one of the things that we we will be taking questions right so if you join us live and we have a question about this stuff that’s your opportunity to like get into the mix and and have us ask Emily those questions and one thing we’ve always promised on this podcast is if we’re gonna have an interview it would be an R format so the questions are not coming from Microsoft just so you guys know there’s not a sponsored event so to speak we really just want to ask some questions that we deal with from our day-to-day lives dealing with

11:59 from our day-to-day lives dealing with power bi so come join ask questions you’ll hear the questions that we ask each other and hopefully we’ll see how the answers go exactly so anyways moving on that so that’s our major two announcements coming up coming up now we’re going to be able to move on to our main topic of the day so the main topic for today comes from PhD pH data I guess would be a PhD Ada or pH data I think it’s however you want to speak about it so Luke Sankey stanky

12:30 speak about it so Luke Sankey stanky stanky it’s better than I was gonna say so okay so Luke writes us a great article around why are we interested in low code or no code is this user friendly so let me throw the the article link here in the chat window and for anyone else wants to read along or jump in with the article this is going to be our topic for today low code or no code Solutions and why or why they are valuable to an organization I I really like this article

13:00 I I really like this article I I I guess we should probably step back here for a second for people who aren’t familiar familiar with this concept of low code or no code Tommy want to kick us an example here of where we would find a low code or no code solution yeah I’m first off I’m not a huge fan of the term but that being said there’s a lot of uses for it I think that one of the biggest ones that at least Microsoft has pushed is the power automate but especially the power app space when

13:32 especially the power app space when you think about coding usually think about application development and with powerapps especially it allows there is obviously a language involved but is much closer to excel than it is to c-sharp or HTML and it allows a user to use the drag and drop a UI functionality with also using some of the language features so if I were to build a full functioning application in powerapps I may have to say like what happens when people select a button but I can just say hey here’s a button let me drag it let me put on the page let me change the

14:03 let me put on the page let me change the color the entire application or interface for the end result is not all coding half of it is going to be a user a UI a GUI and the other half is going to be maybe some language but a very straightforward language so when we talk about so you’re talking about power apps in this first yeah perspective so really any solution is going to have that combination of the UI and some language involved so if I really step further back

14:33 so if I really step further back here here a lot of these low code Solutions are low code or no code Solutions are already tools you already use every day right you you could again I’m just riffing on here an idea here that you were talking through Tommy like Excel word PowerPoint all of those Solutions are already low code or no code Solutions right you you could write code right so if you think about how do you build a document and all these documents are based on top of a code base right there’s a

15:04 of a code base right there’s a definition of how the word document gets stitched together so you’re able to kind stitched together so you’re able to come in write your ideas down use the of come in write your ideas down use the tool for a specific purpose and then out pops an a piece of material that’s valuable a slide deck right a Word document so you can print and use or you can move or do things with it right Excel I think would be one of the exceptions of this one because there’s a lot of code I would say probably more code than most right to write functions and things inside Excel however though the math behind

15:35 the math behind all of the functions that you’re using inside Excel there’s actually a lot more code that’s being written to be able to make those functions work inside Excel so so there’s we’re already very familiar with low code no code Solutions I think what I see things changing here is we’re seeing these low code or no code Solutions move more toward application development and that’s Tommy I think where you were is right hey we can go to a power app so I can drag and drop visual elements on a page or a canvas and that builds me an application that is then written by

16:06 that is then written by low code Solutions and you still have the right code there’s still 30 things that you write in there this button does this when I click this button I’m with this this thing to interact so there may be some additional elements there that you’re wiring together but you’re you’re removing a lot more of the code writing capabilities so I think I would design I feel like in the no code Solutions there’s not very many of those I feel like most solutions that I see today are

16:36 like most solutions that I see today are low code and I would think I would throw power bi into a low code Type solution because you still got to write Dax and data model things yeah ultimately all these tools and and maybe they’re getting maybe the the phrasing low code no code Solutions is is bigger over the last five plus years because they’re getting more advanced but like your point right like if there’s a user interface that is allowing a business user to interact and like develop code by clicking buttons or

17:10 like develop code by clicking buttons or putting a flow together or specifying what they want to have happen like that’s what these tools shine at right and I think that’s where where things really take off with things like power power automate yes but that almost fits into some of the tooling that was there before the powerapps is is uniquely different I think from the standpoint that there’s actually an interface where business users don’t have to go directly into code right you’re You’re Building these artifacts that are become pages right that become

17:40 that are become pages right that become interactive experiences for end users and it’s it’s much more of a code by Design and or I should say Page by Design as opposed to developing straight from a coding perspective it’s development by Design and I think the real definition is the substitution of what previously was code so I don’t know if I’ll put like Microsoft Word in there because there was there wasn’t there’s not a substitute there of what just a pure coding development typewriter

18:11 pure coding development typewriter would you call that coding I don’t know okay okay there’s no substitution like with with report development you had a no SSRS right you had a that was really from the ground up you had to really write all of that that let’s be honest here though let’s be honest let’s ask the SSRS person of the team here team here in SSRS we’re rebuilding a lot of visuals in SSRS or was SSRS really a table generation tool table is a matrix yeah I’ve seen this both though I I’m

18:42 yeah I’ve seen this both though I I’m not saying the vat like yeah sure of course you’ve seen both yes because that’s on the demo examples these are examples of things there might there might have been a bar chart maybe yeah good yes but the the the vast majority of yes reporting Services reports are just tables tables okay and so and that’s that’s my so if I think about in an article here what Luke talks about here halfway through he goes there are there are no code to no code tools such as Tableau Sigma Computing and power bi and

19:14 Tableau Sigma Computing and power bi and so he’s picking on the visualization side of things where there is low code or no code Solutions around data visualization right if you think about again let’s go back to word excel and now data visualization where does the most code live in any of those scenarios to your point Tommy there’s probably not a lot of code that needs to be done to get a Word document done plain and simple done right so that so that’s almost no there’s almost like nothing there right so let’s ignore word let’s go over to excel in Excel you can

19:45 let’s go over to excel in Excel you can build lots of data points but then the visuals the visual portion of that that’s again a lot less code and then you go to Power bi you have this concept of like okay I need to I need to have this data model interact dynamically I need to select something and something needs to change and then if I think about the the amount of code that takes to write or build a visual that the amount of code required to do that is actually a lot higher and that’s I think also to the point of where powerapps comes in powerapps to build an application is not easy and also it

20:17 application is not easy and also it requires a lot of code to be able to build a well-designed or good application so I think what we’re seeing here is we’re seeing a trend in the industry that is we’re taking things that traditional really would take a lot of code to build a bar chart yeah like a lot you have to go like d3. js you can build all these libraries like you’re using so the the effort to take to build that visual was actually much harder than just saying I’m just going to make a table of information right cut it back to the

20:47 information right cut it back to the SSRS request right so SSRS was using to build tables but now when you build these interactive visuals that communicate to each other and they they

20:55 communicate to each other and they they pass filter context back and forth that’s an immensely more difficult to build from a pure code development standpoint and this is one of the the reasons why when I walk into organizations that are looking to build applications I talked to him a lot about like how long would it take your developers to build two visuals that do this and I demo them like they’re like oh that’s that’s really hard like that’s that’s challenging because you have to filter one and then it’s to pass the context the other and have to re-rerun the query for the other one so I’m like yeah that’s the point the point is a

21:26 yeah that’s the point the point is a large majority of this code stuff is being handled for you the the desktop application I think that’s a huge point that even I forget even in conversation right you have a framework and yeah you’re bound by certain constraints but at the same time think think about that experience as as a bi developer right drag an object on the page it just works drag an object on like a completely different object on the page it just works set properties and filters on each one of these visuals and like they still

21:57 one of these visuals and like they still interact in the way that they should like there’s so much you’re getting in these tools out of the box would would you classify power bi as no code though I would not no I definitely would not what was interesting is where like in the article it’s no code tools such as Tableau Sigma reporting power bi provide a user friendly interface that allows non-technical users to access and analyze data without coding knowledge I don’t like and I agree with you guys it I don’t think it’s no code if you’re

22:28 it I don’t think it’s no code if you’re developing but where through my mind was technically technically you’re creating an experience that allows for a lot of analysis yeah in the reporting page which is no code right it allows the business user to get a lot of value in picking and choosing slicers filters changing context with bookmarks so you can create experiences that are no code and I wonder if like that’s where they were headed because even if I look at I

22:59 were headed because even if I look at I don’t know Sigma compute but like tableau tableau we’re talking about code all the time like there’s a delete it’s at least low low code for end users to build their own reports one of my favorite quotes I saw was like someone said my favorite thing about low code no code Solutions is that little box you put your code in and you think about both languages like regardless if you’re in the formula bar yeah the formula bar or the power query and right yeah we’re not necessarily developing power bi and vs code which

23:30 developing power bi and vs code which you could using Power pvi Tools but you don’t have to but that being said there’s still obviously a language behind it but I I think there’s something to be said about the no code part of this or at least at least again again I’ve been doing a lot of dashboard in days lately and I’m always I have to remind myself people can be amazed by some of the simple features that power bi can do we’re just going through simple Transformations building the decks and doing that interactions in my

24:00 decks and doing that interactions in my to your point even in 2023 people were like afterwards you’re like what everyone thinks like we are blown away we could not believe this could be done like this opens up so many doors and we didn’t do any that yeah we wrote a sum and we did a year to date market share nothing it’s Excel so if you consider Excel low code then this would be the same thing but really it’s not that difficult in a sense like at least from the basic point of view from a business user to take up I feel like a better example of low code

24:31 I feel like a better example of low code or no code would be power query so so I feel to me like so this is where RBI yeah yeah so so this this part of power bi I think is actually much stronger when it comes to this low code no code solution because in my mind that’s the space where I can literally click a column and click a button and a menu pops up and it gives me the prompts that I need to write in the the expression pieces that would then populate the code that actually

25:01 populate the code that actually generates the new data table and yes under the hood there’s a lot of code underneath of it but if you read Gill reviews book when he talks about this he goes like 75 maybe even up to 80 percent you can get away with just using the UI right so power query everything that’s for you Alex so using the power query you can get 75 80 of what you need to accomplish by just using the UI and I think while some of us are incredible

25:33 think while some of us are incredible coders right so Alex it doesn’t even use the UI he basically just opens up a notepad and starts writing M verbatim and notepad so Alex doesn’t even use the the the Hitman editors in studio code yeah he just starts straight up just writes it from memory writes everything verbatim and paper exactly yeah not everyone can do that like Alex does however however what I would say though is if we’re talking about how can we bring the largest audience directly to

26:04 bring the largest audience directly to power bi or power query for that matter you need some user interface and honestly when you have tools that do this yeah the auto generated code may not be the most efficient bit of code that it’s writing however I’m learning and I feel like this is what I did when I started working in Excel actually learning Visual Basic which was record a bunch of things in Excel that it’s doing watch the code that came out and then Google the gaps between what I what it generated what I wanted to do and then I was able to write my own code and

26:34 was able to write my own code and actually functions and things inside Excel so I feel like for me as a learning person right to learn a new language it’s very easy for me to use a UI UI get 90 80 to 90 of my stuff done there and then when I get stumped I can then go into the formula bar edit modify those things as well I think people more comfortable looking at smaller sections of code and being able to manipulate them them the raises interesting question where Alex says something about Tabler editor

27:04 Alex says something about Tabler editor but moving forward and I think with also power bi modeling the service and moving that way like Mike you and I could survive or without Tableau editor for a year like it would be difficult but we could still do it and do you think it’d be incredibly frustrated yeah frustrated but we could still do it things would get broken yeah my family would be like why are you so angry Michael I don’t have tabular editor why are you so grumpy I’m not saying your quality of life would go down like ours would definitely would but we could still

27:35 definitely would but we could still accomplish like it’s not necessarily like we have to have these coding Solutions in order to do what we’re doing doing for them for the majority of things and I think as we think about where power bi is heading where they’ve obviously invested in the Enterprise now maybe now there’s this shift to going back to the the end user using power bi do you think that we have enough with low code and power bi that can meet requirements coming up when it comes to easier access to data integration more things that are accessible to the

28:06 more things that are accessible to the end users do we have a good solution out there or are we does the end users have the right in a sense assets for them to accomplish what they need to do well I think in general and I’ll step in here I agree with a lot of what Mike was saying right like we have think about how simple the experience is for the most part around the major connectors out of power bi I just connected something as long as I’m like and Azure ID has made a huge book like that’s super easy if it’s Microsoft tools right really easy to connect to a data source

28:38 really easy to connect to a data source even Salesforce is interesting like just connect to it I see all the objects connect to it I see all the reports like as long as you have the perms you’re good to go the other thing that’s interesting in in that I think dovetails into here is like is a tool going to solve all those problems no but at the same time low code low code tools are designed I think like Mike was saying is if you can get a business user 80 of the way there way there through UI through them figuring out their problems that they want to solve

29:08 their problems that they want to solve because their bent is not from the developer perspective nor is it heavily from the data infrastructure or how to build things correctly perspective it’s I have a business need I have data I want to do this thing can I get there doing like eight eighty percent like power query if I can do 80 of all of my transformations in the UI how what is that coverage like for all of these lay users it’s probably almost all of them and then when they hit a point where

29:38 and then when they hit a point where it’s like oh man I can’t do that they’re still willing to invest the 10 or the next five because it’s only the last 15 that that technical experts are going to be trying to solve right then you’d reach out and you’d help out but like that swath of business problems that you’re solving for those end users is so great that I they ha like those tools allow them to connect to the data transform the data and then are there always modeling issues yes there’s always modeling

30:09 issues yes there’s always modeling issues and how do they get things and then they wind themselves around the excellent Dax more often than not if you get a above one two data sources Etc but so I think there’s I think there’s some more room for improvement once we get into the modeling side or areas to help users actually put together their own reports but from a data access connecting to data transforming it I think power power queries got it on lock yeah yeah my only gripe with power query is more around is

30:40 gripe with power query is more around is less around what it’s doing currently it’s more about the performance of power query so I think my only my only Google right that’s your same point yeah like at your earlier Point whether you’re generating M or some other language it doesn’t really matter what you generate power query generates SQL statements today right you can you can see how it query folds things back up so whether it’s generating M or SQL in my opinion it should not matter and and that’s honestly like how I started learning SQL again

31:10 started learning SQL again what I what I think is happening here and again this is my Take On The World here a bit is the low code or auto code generation tooling is actually producing

31:21 generation tooling is actually producing you a learning method right so if you think of it very abstractly if you say click these buttons do these actions I I know there’s a gap between what I want the data to look like and what the data needs to the shape of that data right I can I can delete these columns I know what I want to do with the information so in concept I’m there what I can’t do very well in these higher code Solutions is what function should I write to select these columns if you go to python select these three

31:51 if you go to python select these three columns from this data frame I maybe like I like I if I play with python more I’ll know you if I play with python more I’ll know right off the top of my head how exactly how to do that however if I click a button that just says select columns that makes more sense and then when I see the code being produced that’s now basically teaching me that this is the type of code that I would run to make this output and eventually you get to again where Alex is where you just jump into an editor you now know enough of the functions you’ve seen the code enough you’ve done enough work with it

32:22 you’ve done enough work with it it’s actually teaching you so I think these low code or no code Solutions are actually what they’re doing is they’re actually teaching tools to teach people how to code no you probably don’t go not every user is going to go all the way in and learn every little Nuance of every piece of code but I think inherently you’re naturally going to learn how to do coding things and so instead of going to school school to get a specific coding scaler be a software developer you can now hone your skills in on a very specific skill set

32:52 skills in on a very specific skill set I’m going to be able to write VBA for Excel I’m going to write Excel functions I’m going to write I’m going to build power bi things and what these tools are doing now is these tools are being very specific they’re giving training instructions basically around a particular tool or a certain thing I I think you’re writing the line which is why Tommy’s probably I’m interested in Tommy’s feedback here but I I think in the same way that you just said Excel VBA all these other like deep dive

33:22 Excel VBA all these other like deep dive parts of excel the capabilities were there it still takes the person to go learn them and go do them by Nature these new tools are not doing something revolutionary yes right in in the fact that it still takes somebody the their own initiative to go learn about the capabilities or be pushed into a situation where they have to solve something something right and power query is the same way one of the things we did we talked about earlier that you said is you’re going to sacrifice efficiencies initially with

33:54 sacrifice efficiencies initially with these tools because people who don’t understand them are going to be using them right so power query for example Performance is always going to be a problem technically right off the bat because people aren’t going to bucket their transforms they’re not going to have the same types of activities in a single trans transformation step they’re going to go like hide column add column remove column transform column do the do something else do it like and then you’re gonna get this mishmash right of transformation steps and they’re not going to know that they’re not going to solve the oh I

34:26 they’re not going to solve the oh I should bucket these all together and things are going to become more efficient until they have a performance problem yeah so I think I think where my mind goes technically is is you get a long way until you have to solve a technical problem and that’s where Mike I think those doors open where an end user can say okay I need to go like figure this out and then there’s a body of like everybody solving this problem and then it gets deeper it’s like oh underneath the underneath here is in this m query and I need to go into

34:56 is in this m query and I need to go into like the advanced editor and oh okay I see what’s actually happening here and if I do this and I can modify this thing and right so there is a progression and I think it’s a natural one because from a business user perspective you have this this the the the low code tool is is giving you a mental framework right of how these things it’s a it’s a way to think about data and these steps being there give me an idea of what is happening behind the scenes that leads me into a

35:27 behind the scenes that leads me into a more technical route and I think that’s the value there as far as like them them providing that framework to people but I I don’t know if they go to like the teaching level I I and I think the reason I I may be giving a face or two is I think part of me wants to take a step back from the definition here because I think everything we’re talking about at least from the power query that’s a low code component that’s not the solution and I think when we’re talking about the

35:57 think when we’re talking about the solution it’s the end result when I think about an end user going through building the model power bi and the end result of the data set is the low code solution solution power query Dax and the modeling the transforming is just a component of that overall solution of low code it’s not it’s not the entire and but no but the reason why I say this is because the output is the most important part whereas if we’re just going to focus on the power query that may be just a table that comes out but the problem I don’t

36:27 that comes out but the problem I don’t say the problem the concern is with low code Solutions think of the output on power bi what’s the output of a data set it’s published the service now there’s this other source of Truth with powerapps if anyone can do it they may the com they all those components make up an overall application that now people are going to be used and I think that to me is where I’m categorizing a component of low code and a solution of no code I didn’t say power

36:57 a solution of no code I didn’t say power bi was a solution no no well technically okay I get your point yeah so power query is a great example of a a good local code solution I I guess what I haven’t seen is how far can you get in the next layer which is modeling and visualization with low code right I can create a relationship I have a diagram right create a relationship between two tables I’m always inherently going to run into the fact that this is going to be

37:27 the fact that this is going to be bi-directional which is going to create problems for for calculations right but for sure for sure so already that that doesn’t bode well but like how much how much are people just doing sums right how much are implicit measures right I can drag a column I can do these okay I’ve got to figure out some areas to to fix things but like what is the percentage of people that are doing that versus instantly where you need to go to the next level and I I need to create a Dax calculation so so this is where I

37:58 Dax calculation so so this is where I think the Lo so I agree with you 100 South on this one and I would also argue that power bi has been focused a lot on the the model the data modeler and the report developer is like the two main areas I focus on there so if you think about so in the in this low code solution right if I’m building if I’m building a thin report on top of a model that already exists a lot of that is already low code or no code solutioning I drag things together stuff appears on the page I snap bookmarks all that’s

38:28 page I snap bookmarks all that’s generating code behind the scenes it just works so the the report experience I think is a lot more optimized for low code solutioning the modeling experience and I think this is what a lot of people are talking about in the chat window as well is how do you get the modeling experience to to match more of that low code experience because it is not low code there’s a lot of things you’re writing there and I think a mist from the Microsoft desktop team is to make the modeling experience more like a low code solution hey we like you

38:58 like a low code solution hey we like you like a low code solution hey we like for example the experience I would know for example the experience I would think here is you load in a handful of tables from Salesforce the the UI basically says welcome okay I see you’ve added these tables help me understand how these tables should be linked together I see that this table looks like a fact table how what data pieces link to this table and you actually prompt the user to walk through the relationships and this could be in an experience that is entirely built through chat GPT because chechi BT could read the tables you could read The Columns it could it could assume some

39:30 Columns it could it could assume some more information there and you could you could literally talk to the data model and be like the relationship between this table and this table is on this column and then the chat GPT could go check the two columns and say well I see one column’s a number and one column’s a not or by the way you if if we make this relationship you have 50 of the joins occurring between the two tables right the the chat GPT should be more interactive on how you like communicate what the data is doing so like you’re removing all really

40:01 so like you’re removing all really development side and if to Alice’s Point everything’s a wizard at this point well maybe but the so let’s keep going forward with that one right another challenge I think that’s there is a lot of people don’t understand the difference between implicit and explicit measures sure I think if you went through the model and had again like another type of wizard or just automatically did it every numerical column here’s all the numerical columns you had in your data model we’d like we’d recommend you do these things do you want to hide the old columns do you

40:31 you want to hide the old columns do you want to make a measure on top of the columns what measure should you make on top of these columns is it a sum is it an average is it a Min or a Max have the user go through again more like a wizard type thing and again if you went through a model and pre-built all the measures hit all the columns you wanted put them in folders I think you’d go a long way to giving a lot of value to those initial custom those initial users and you’re now moving more towards that low code solution and not as much down the heavy coding solution I I think that

41:01 heavy coding solution I I think that works but it’s there still has to be development on the back end so one thing that doesn’t seem to really catch on is trying to Global prints or these Global blueprints that Microsoft or a lot of Technologies try to do what always seems to work much much more effective at organizations is yeah you give the end users they’ll say that low code solution but is still developed by that organization based on culture accessibility and what constraints they want so think about this wizard where

41:31 want so think about this wizard where they’re going to always try to Auto populate everything like Hey we’re you see we’re trying to do these numbers you’re trying to do this model why I’m thinking of like think about the the automated insights right that has not really come to fruition

41:48 that has not really come to fruition and I think one of the biggest requests was people were like we want to be able to model the back end of this and allow the automated insights to be developed by from a developer point of view to give weights to certain elements because Microsoft was trying to give a this low code solution of here’s your automated insights based on the model but we’re going to use the same AI generation for every model without customization and it didn’t work it hasn’t worked I and it didn’t work it hasn’t worked that that was that that was

42:18 mean that that was that that was Microsoft build 2022 or I think it was just last year that was the number one feature or at least two years ago where are we at now with that features exactly not no one’s talking about it right right so go ahead tell me and no I was the last part to say is I think all these things we’re talking about only makes sense when there’s still someone involved at the organization or at the whoever’s part of that culture that can still give weights and customization to it it so so the interesting part of this

42:49 so so the interesting part of this conversation that’s striking me is like I’d really like to understand the challenge it because going from Power query into the model and creating a model is is really important and into Alex’s point a lot of I think a lot of end users may just have one giant table and that’s what they don’t have to deal with this right but I don’t know if that’s ever use case because a lot of the folks that I interact with it’s always at least two three four tables and typically it’s a mishmash of the data in the two three four tables so

43:21 data in the two three four tables so what I don’t understand is what are the challenges around not guiding the end user between power query and the model right like are you is it so much that we’re not going to get to the 80 where you couldn’t have a wizard or something that says hey here’s some basic concepts don’t call them a fact and dimension right call it a unique list of identifier information this is important because your your other table should be us counting a bunch of different things blah blah blah blah and

43:53 different things blah blah blah blah and allow it to help an end user shape their data because if you can identify that this is going to be a bi-directional relationship or there’s multiple unique like there’s no unique uniqueness in the stem table and somebody labels it their their unique dim table then kick them into fixing the data right or identify for them why what they want to have happen isn’t going to happen right because if if I say this table needs to be a unique thing and

44:25 table needs to be a unique thing and you’re like yep that’s the unique thing and then you try to create the model it should be like no it’s not unique because what you have actually is these in these records that makes them solve the problem before they get into a situation where they have no idea how to solve it yeah right and that’s that’s to me I guess one of the interesting things around like why we haven’t been able to at least get to a point where you’re guiding folks or at least allowing them to pick this wizard thing that’s going to help them build the right thing because then then implicit measures

44:56 because then then implicit measures would go a long way or calculations would be really simple because their model was designed the right way but instead we have we go from this very this very good low code power query to have fun yeah yep and I think that is one of the big challenges but the challenge there too is also power queries functional whereas Dax and we’ve said this to we’ve said this too many times Dax is always going to be dependent on your model and yeah we can say AI can solve for this it’s a

45:27 we can say AI can solve for this it’s a lot easier to solve this in power query than it is to solve this in a every custom model where you may not have everything set up just so whereas power query I need to add a column I need to transform this table generally speaking Yeah like columns matter but not nearly as much as what you’re trying to do in the model based on the model we Mike and you both all three of us have said this your Dax is only going to be as simple as simple as your model is

45:58 as your model is yeah and I gotta think Alex appreciate the the sponsorship there we thank you very much for throwing in there in there we’re trying to make a great discussion here so thank you very much everyone I appreciate that and acknowledge it as well so I I agree with your your point here Tommy and I think well again I want to I want to bring up a feature that I’ve seen inside the Microsoft blog as well one of them is creating power bi reports with Jupiter notebooks now granted there is no data modeling really happening here right this this is purely a there’s

46:29 here right this this is purely a there’s a data frame that is appearing basically a single wide table that you may be building in your notebook and then you can then basically bolt on a powerba report and you can then pick Columns of the data frame and it will then Auto generate visuals for you okay so Microsoft already has the technology to do this they already have the technology to say based on what we you send us some information and we’ll try to pick out interesting data points around you and to your point Tommy right right automatic insights super cool it was a

47:00 automatic insights super cool it was a neat feature but did it really in reality provide the value-added data story that we’re trying to produce with reports I would argue the answer is probably no right it was limited in what it could do however I do think the fact that you’re able to do a large majority or or get a lot of information out quickly I think there are some layers it’s going to get better right so I’m just I’m just saying this is this is a progression of things right that we’re going to continue to see keep exploring these generative visuals or

47:31 exploring these generative visuals or generative report elements so that it makes sense so I guess at the end at the end of the day when I bring things back modeling and being able to build your measures and that story of the data that’s a much more user user generated or user based decision decision feature right so it’s going to be much more difficult for people to just show up and be like oh yeah power bi automatically create me a

48:01 power bi automatically create me a report and it just builds this incredible report out of the gate I think it’s just going to be not it won’t be able to get there technology for a number of more years I think it’s going to take more time to get there however I do think we’re already seeing the benefit of a lot of these low code or no code Solutions around generating the visuals interacting between the visuals right there’s a lot more buttons that are doing more of the report authoring because I think the report side is easier to generate where I’d like to see more integration

48:31 where I’d like to see more integration is let’s start lifting off these heavy loads around Azure creation incremental refresh like though those are things that should be more based around AI or or generative or low coding type Solutions these are these are very known let’s call them let’s call them these are very known workflows right right so if you think about what you’re doing in data model you’re doing a lot of workflow based things right you’re building relationships you’re making measures and there’s a workflow that’s associated with that sequence of

49:02 associated with that sequence of designed designed items right so another one that I think would be very useful here is and this is why we built our Dax generator so Dax the patterns inside Dax are very repeatable the tool today doesn’t give you a lot of help around hey I are you trying to build a cumulative total or are you trying to build a star rating based on a number column right there should be more libraries or information that is here’s a bunch of Dax that you can go pick from there’s

49:33 Dax that you can go pick from there’s there’s patterns that are existing and this is why micro Marco Russo and Alberto built the entire Dax patterns library because there’s the over and over again they kept seeing the same type of calculations occurring all the time what we really need is we need a better tool that lets you take from a library of Dax patterns drag them into your data model and and rearrange those patterns so that they work on top of the way your columns are named that I think is is much more impactful and Powerful so Microsoft did a good job when they

50:04 so Microsoft did a good job when they introduced quick measures but I think they should have open sourced that and allow a lot of other people’s to be able to build their own Dax patterns and inject them because that is that is true low code solutioning right yeah here’s here’s 300 measures you can go use inside your data model pick one of these and then those can be injected directly into your data model now it’s a uui it’s a user interface so oh man I’m going full circle here Alex I know you made a comment here back inside

50:34 made a comment here back inside the chat window basically we need more Wizards yeah we do we we need more Wizards and that’s what low code is it’s a basically a wizard that sits on top of some information generates automatic code and then spits out the answer so why not open source all of measures measures quick insights for quick measures should be immediately open sourced give us the standard on how to build those things and that way we can then inject them in all the time and then it’ll make you all the time and then it’ll make immense amount of value and again

51:05 know immense amount of value and again this is this is the two-fold purpose here right it provides you code with with minimal code experience you get something in your data model you can go look at and take apart decompress it change it look at it modify it and by doing those two things you’re now teaching the audience all about Dax because you’re now having good examples up here inside your data model anyways yeah I I think I I think you’re spot on like the more the more you build low code Solutions like that right

51:36 build low code Solutions like that right but all of a sudden you bring into the into the ownership area of of that a lot more capabilities right so rather than it being the 10 of people who know Dax really well all of a sudden you’re you’re reducing that down to like a much smaller number because you have something that’s repeatable and used if I’m looking at our conversation though in terms of like power query and visuals and Chris is pointing out like you can in in the power bi service you can do a quick report where it’s a pretty easy to

52:06 quick report where it’s a pretty easy to generate visuals yeah or the Gap here is the model it’s transitioning from a power query where I have data that I’m

52:14 power query where I have data that I’m bringing in and the model itself and that’s where I think if a better low code solution for folks could be implemented just to ensure that the experience in the visual layer is exactly what they would expect it to be and that requires this layer of hey do you have one table have fun you’re done right and you have two okay here’s some Concepts that you need to understand and you need to set up and you might need to clean your data before we get to the point of visualization and imagine how much better that visualization

52:44 much better that visualization experience would be if you force the user to go do those data cleaning things beforehand right even if it was a process of okay I’m gonna do you want to use the not Auto Builder right and I think that’s where we get into these situations but folks are like oh I’m going to solve all of these problems for you no there are fundamentals that we know about model building that you could absolutely just put into a here’s an error right run me I want to run this

53:14 error right run me I want to run this process that validates that I’m saying these objects should be this thing based on what you’re telling me and does my data match that yes or no if it doesn’t give me an example what what doesn’t align with that because I would assume that it does and then you’re teaching people the data cleansing activities that they need to understand and before they get to the model and I think that’s just where the Gap is right now in terms of the tool set because connecting those lines or just because

53:44 connecting those lines or just because it allows like the the hack here was allowing any type of relationship to go in between two objects and then correcting that people are going to be able to develop what they need to in the visual layer so that’s a great point because theoretically I could go and build an entire power bi report without ever touching my keyboard then except when I click on Save I could literally go through and it’ll be very simple and it may not have the principles but I think people have a misconception that that is the low code

54:14 misconception that that is the low code solution where it’s like no you can yeah you can go through without typing anything however there are that doesn’t mean that you do not have to learn the fundamentals the principles the theory around this and that’s the misconception here yeah and I and I think that’s a progression right I think it’s a matter of add value quickly get people hooked right and that’s what that was what Excel was that was what VBA was for me inside Excel that was power query for me when I started using power query right get the get the 80 of the use cases kind

54:45 get the get the 80 of the use cases kind get the get the 80 of the use cases dialed in figured out just make it of dialed in figured out just make it work right and then once you’ve got people’s attention now give them an easy way to get into that so moving away from the no code move them into low code and then you turn people into Alex powers where everyone just writes nothing but code like so you it’s a progression you work your way through there well there’s the fundamentals again like I I I I’m gonna try this today I’m going to disconnect my keyboard and build a report and I probably could but again I’m missing a lot of the it’s going

55:16 I’m missing a lot of the it’s going like the tip of the iceberg correct yes and people oh well I can do power bi without any code you’re going to get a very vanilla looking report exactly but you can do it yeah you’re you you can do it yeah you’re yeah but other solutions back in know yeah but other solutions back in like when we talk about normal coding items you have to type a lot you have to understand the language anyways as we go through this I think my big takeaway is there’s fun regardless of if it’s a low code or no code solution that does not does not eliminate the need for fundamentals and

55:47 eliminate the need for fundamentals and the theory behind it whether it’s going to be notebooks where I can just ask AI what to write in I still have to understand what I’m trying to get out of it and also how it works so yeah I I agree with Chris Kratos there’s a lot of this conversation that is teamed up to good questions tomorrow exactly right exactly right so can I not use my keyboard anymore exactly right well I’ll be able to model without using my computer awesome so let’s jump into what does chat

56:17 so let’s jump into what does chat GPT say so this is our final segment here but we’ll jump into a question that we ask against chechi BT and gets the answer out and when they see what what value is provided from that answer so my question was why would the why would my organization want to invest in low code or no code Solutions so chat GPT comes back and says low code or no code Solutions haven’t been increasingly popular in recent years I would agree with this I think Microsoft has recognized this direction as well

56:47 has recognized this direction as well couple advantages are increased Speed and Agility with low code or no code you can perform and deploy applications faster than traditional coding methods I’d agree with that you have reduced costs for the low code and no code Solutions typically there was less coding requirement in the development phase this can result in cost savings for your organization and I would agree I’ve built software and so building software the most expensive portion of the software definitely isn’t running it it’s not the service that it sits on

57:17 it’s not the service that it sits on it’s the people that are required to build the logic or the code itself that is the most expensive part about building applications and that’s what these Solutions are doing is they’re addressing the most expensive piece totally agree with that statement as well well improved collaboration low code or no code Solutions and have a typically have a user-friendly interface that allows non-technical staff to participate in the development process I don’t think I would have picked up on that one initially that’s a good point increased Innovation with low code or no

57:48 increased Innovation with low code or no code solution platforms your organization can experiment with new ideas and prototypes quickly and easily and I think that’s an underrated statement right there because the amount of experiments you’re doing with power bi and data I have to be you’re gonna you’re gonna need a lot of people looking at a lot of things to get to value added even Solutions like all the all the all the connections right yeah exactly data sources exactly the other seniors can work out yep and this is that’s a talking point that we’ve made

58:18 that’s a talking point that we’ve made on the podcast is the business should be leading the ITW group because the business should be out front experimenting trying stitching together random things because the business is going to find what adds value once you figure it out then that low code no code solution can be handed over to an IT shop or a bi team and they can productionalize it or make it repeatable or open up for more data that make that pattern of working fast and failing fast makes sense to get to an answer that makes

58:48 to get to an answer that makes that actually adds value later on and then the last statement they hear it’s easier maintenance and updates low code or no code Solutions provide provide built-in testing and debugging tools that make it easier to maintain and update applications over time maybe maybe that’s because Microsoft hasn’t built that into their tool currently so so maybe that’s the next Horizon for Microsoft should be thinking about okay what does data Ops look like

59:21 about okay what does data Ops look like for power bi because that’s a myth that’s a huge Missy nugget as of this point like we’re getting we’re just starting to get to the point where we can do like diffs between Dev test prod with code so we can see what’s actually changing okay good success but it we’re not able to get down to the did my data table change in these different environments and so that’s that’s the next layer of what we’re going to get to is this data Ops layer and I’ll take that back please actually in powerapps they’ve done phenomenal features and tools to monitor debugging

59:52 features and tools to monitor debugging like what’s going on and I use it extensively with users say Okay connect to this let’s see where why this errors oh interesting oh it’s a fun it is full scale but but again it’s about powerapps thing so so that’s a that’s another so it does work that’s a good point by chat GPT that’s what I’m saying yeah okay so the point is valid it just it just doesn’t necessarily apply right so right yeah I would agree and I and I would also argue too Microsoft has probably focused on getting more monthly active

60:23 focused on getting more monthly active users and mouse I think is what they call them Microsoft is very focused on getting more people engaged with the tool and less focused on should I make sure that you have a good data Ops process like that that seems like getting that running is valuable to large Enterprise but not necessarily valuable to the average everyday user they’re not thinking that way yet so maybe there’s a product Gap there how would you rate chat GPT six out of seven times six out of seven I give it 8. 7 out

60:54 times six out of seven I give it 8. 7 out of 10. of 10. I’d give it a a 775 out of 1000. nice yeah no maybe a little higher maybe 850 at 850. I’ll go 850. 850 out of 1000. literally we need to work on our data culture and governance here I don’t know what you’re talking about my way is the right way [Laughter] correctly

61:24 who’s going to govern us who’s gonna who’s gonna govern this who’s our sponsor you can’t govern me I sponsor myself awesome well thank you everyone very much we appreciate the conversation the chat threat has been absolutely Lively so thank you very much chat for for jumping in and really engaging with the topic a lot of great ideas are occurring here thank you very much for the conversation and how much fun this was today our only ask for you if you listen to the podcast is if you like this if you found some value from it if you had a couple Chuckles or laughs around it we only ask that you share it with somebody

61:55 only ask that you share it with somebody else let somebody else know that you found some value from this and and encourage them to discuss and engage with the topic as well Tommy where else can you find the podcast guys we have a mailbox submit your questions we will do look at the mailbag submission so make sure to submit a question or top you want us to talk about also join us follow us on Apple Spotify YouTube Google podcast anywhere a podcast is found if you want to join us live you can do so every Tuesday and Thursday 7 30 a. m Central

62:25 Tuesday and Thursday 7 30 a. m Central on the power bi tips social media channels one time I almost messed it up you’ve only done this a 212 times and you almost missed up I was going off the cliff and I just turned back so well thank you very much for listening we’ll catch you next time

Thank You

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

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

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