Publish to Web vs. Embedded - Ep. 512

If you’ve ever heard someone say “can’t we just publish the report?” when what they really need is a secure customer-facing analytics experience, this episode is the cleanup your architecture deserves. Mike and Tommy separate the deceptively simple Publish to Web feature from Power BI Embedded, explain where each one fits, and make the tradeoffs painfully clear for anyone responsible for governance, external sharing, or productizing data.

News & Announcements

• Entelexos Embedded Accelerator — Entelexos is a Power BI embedded accelerator designed to help teams stand up a branded, secure analytics portal much faster than building the whole experience from scratch. The pitch is straightforward: it adds multi-tenancy, user and group management, white labeling, export workflows, and embedded report administration so organizations can move from internal reporting to customer-facing analytics without reinventing the plumbing.

Main Discussion

Topic: Choosing between Publish to Web and Power BI Embedded for external sharing

Mike and Tommy spend the episode clearing up one of the most common Power BI misconceptions: Publish to Web and Embedded are not competing names for the same feature. One is effectively public internet sharing with almost no guardrails, while the other is a secured application pattern built for delivering analytics to customers, partners, or other external audiences with authentication, authorization, and product-level control.

• Publish to Web is for public content, period. If a report can be discovered, shared broadly, or indexed without authentication, it should contain only information you would be comfortable putting on a billboard. The convenience is real, but so is the risk if someone treats it like a shortcut for secure external distribution.

• Embedded is the right answer when identity matters. The moment you need to know who is viewing data, restrict what they can see, or tailor the experience for customers, you’re no longer in Publish to Web territory. That’s where embedded architecture earns its keep.

• Security is not an optional add-on. Mike and Tommy keep pushing on the same point: external sharing decisions are product decisions, not just report settings. Authentication, row-level security, tenancy boundaries, and workspace design all matter once analytics leaves the walls of your organization.

• A portal changes the experience from report sharing to application delivery. Embedded solutions let you wrap reports in branding, navigation, user management, and workflows that actually feel like a customer product. That matters because most real-world use cases need more than a naked report link.

• Cost and complexity should be judged against the use case, not against a free button. Publish to Web looks attractive because it’s fast, but it solves a tiny and very specific problem. Embedded requires more design and operational thinking, yet it unlocks the scenarios organizations actually care about when they say they want to “share with customers.”

• Governance gets easier when the intent is clear. A lot of bad architecture starts when teams use the wrong sharing method for the job. Clear rules—public marketing content uses Publish to Web, secure external analytics uses Embedded—eliminate a surprising amount of confusion.

• The strategic upside is bigger than distribution alone. Once you embrace embedded delivery, data stops being just an internal reporting asset and starts becoming part of the customer experience. That opens the door to stickier products, better service, and new ways to differentiate with analytics.

Looking Forward

If your team is debating how to share Power BI outside the company, make the first decision a security one: public content can use Publish to Web, but anything customer-specific should be designed as an embedded experience from day one.

Episode Transcript

0:03 Explicit measures. Pump it up. Be it [music] high. Tommy and Mike lighting up the sky. Dance to the day to laugh in the mix. Fabric and I get your fix. Explicit [music] measures. Drop the beat now. H feel the crowd. Explicit measures. Welcome back to the explicit measures podcast. Hello everyone and welcome back to the show. Tommy, good morning. How are you doing? Good morning, Mike. It’s another great

0:35 Good morning, Mike. It’s another great great day in Chicago. We are recording this episode, so this is a pre-recorded episode, but today in Chicago and Milwaukee area, it is going to be warm for us. It is the end of February for us, and we’re getting 50° days, which is absolutely unheard of in our area. So, very excited about that. That’s going to be fun. I’ll have to actually get out of my house and out of my basement and go walk around for a little bit., they actually said there’s five seasons in Chicago. So, spring, summer, fall, winter, and

1:06 So, spring, summer, fall, winter, and then the spring of disillusment or the spring of [laughter] the spring of disappointment. Yeah. Because you get these days you’re like, “Oh, I guess winter’s over.” Then it’s going to probably snow again. So, it’s not. Yes. Exactly right. All right., we’re our main topic for today is going to be talking about the difference between publish to web and embedded. And we’re going to focus more on security and public-f facing reports. What is the right place to put some of this stuff? Now, Tommy,, I’m very slanted in this direction. I’ve

1:36 very slanted in this direction. I’ve done a ton of work around embedded, but I do want to really unpack what is going on here. And I think there’s a lot of confusion around what is embedded, when should I be thinking about embedded, what’s the right time to use embedded. So we’re we’re really just going to unpack this and also want to layer this lens of like what does security mean? How do we understand licensing in lie of security considerations as you pick through these solutions embedded or published to web. So I think that’s a really important topic. I don’t think enough of people talking about it and

2:07 enough of people talking about it and every time I see things on like Reddit or LinkedIn there’s a lot of questions about like what is it? How do we best utilize it? So, we’ll this will give you the skinny of all the details there. there. Love it. Love it. Before we do that, I do have a beat from the street, Tommy. And,, today is a pre-recorded episode. So, this is going to be airing on., that’s the the day of this episode. So, in lie of that, this is weird because I’m talking about it on the day of we’re actually deploying this, but I set it up right on the scheduling if the schedule’s correct., so we

2:38 if the schedule’s correct., so we I’ve already just completed, we’re at FabCon right now. and at fabric conference I just completed the full day precon session around agentic development. So I just want to point out as I’m preparing for and have given the talk and I’ll all bundled together as a thought agentic development or building experiences around building things with PowerBI using AI models. It’s really interesting and there’s a lot of things you need to understand in this realm. I’m having a lot of aha moments Tommy. There’s as I’m

3:10 lot of aha moments Tommy. There’s as I’m preparing the content,, how do you start doing this? You start thinking about, well, what is VS Code? How do you use it? Where is the MCP server? There’s now two MCP servers that are out. There’s one for PowerBI MCP server modeling and handling things at the tabular model level. And then there’s another one for Fabric, the Fabric MCP server. And I guarantee you there’s probably going to be more announcements around anything MCP related to the Microsoft Fabric and SQL conference all rolled together. So there’s probably going to be more things coming too as

3:41 going to be more things coming too as well. So stay tuned for all the announcements at FabCon. There’s going to be a lot of great things coming out there. Additionally, that being said, I’m having this realization, Tommy, of there’s this concept of what I’m calling the agentic loop. The ability for the agent to see what it built. For example, let me give you some things that we’re working on in the class that that I’m having an aha moment around. Right? When you save a fi a PowerBI file, you can save it as a PBIP project.

4:11 can save it as a PBIP project. Great. All the files are exposed. You can see everything inside the files. This is really good for letting the AI or agent see the structure of your report, the structure of the semantic model. It’s all there easy easy to use and interact with. The downside is the agent doesn’t necessarily understand everything that it builds. So for example, if you go into a timle file and say, “Hey agent, go add a description,” it thinks a description is a key value pair on how the timle works. But in

4:42 pair on how the timle works. But in reality, descriptions are not that inside. The actual description is a three slash in front of the measure or the column that actually says this is a comment. we’re going to add a comment on top of or the description on top of this cell or whatever. So, it’s a bit more human readable. It’s not pure JSON like you would or pure YAML like you would normally see it. It’s slightly different. It’s been tweaked slightly for what the needs of working with semantical models are. I will say this around the agentic

5:12 I will say this around the agentic loop side, right? So, I can make a change and if you’re using just the agent will say, “Oh, I think I know what you want.” And do the change. that change may or may not be acceptable to the actual semantic model. So for example, example, you can you can make the column called or the the property called description on every single column or measure, but if you try and save that, PowerBI will then freak out and say, I don’t know what this is. You can’t have this

5:42 know what this is. You can’t have this description field. This doesn’t make sense, and you won’t see any descriptions. So there’s actually a hard set of rules that need to be applied when you’re doing this. So how do you solve this? There’s two pro there’s two ways to solve this. One is you can use skills. So skills Tommy you’ve been also exploring with and you’ve been telling me how amazingly powerful they are. So good. A skill is basically you work with an agentic experience a chat and at the end of the chat or the experience I want you to document what you did. What are the steps that you

6:13 you did. What are the steps that you needed to accomplish? What were the pieces of information you needed to go get? get? That’s the skill. So by adding skills to your development environment, you can easily get the agent to understand, oh, this is how I write timle. This is the spec of it. This is how I add comments. This is how I write measures. The formatting is now taken care of by the agent. So the agent has more awareness of how that’s working. So one that’s that’s a solution here. The other thing we were discovering was and Matias Tearbach is also working on the workshop

6:44 Tearbach is also working on the workshop with me. So we’re also going through this together a little bit, both discovering things. The other thing was if you’re using the PowerBI MCP server, this is where the agentic loop kicks in. So you could say,, hey agent, go build me a bunch of measures. And what it can do is it can build a series of measures in the model. It has a skill. The the MCP server knows how to do a there’s a tool basically.

7:14 how to do a there’s a tool basically. It’s called a tool. The tool knows how to write a measure, knows how to go get a column, knows how to write things. So one, a lot of this learning that you’d have to do with skills are now just being one taken care of automatically by the MCP server. But if it screws up and does something wrong, it actually has the ability to go test its own solution. So when I talk about the agentic loop, it’s have it build something, test what it built, and then go back and update,, get get a confirmation. Did it

7:44 , get get a confirmation. Did it work? Yes. move on. If it didn’t work, fail. Go back and rethink what you did. And so there’s this cycle, and this is what we do when we build stuff, Tommy, all the time. We build something, we go test it. Did it work as I expected? If yes, then we move on. If no, then we go back to the proc problem and say, what did I think wrong? I need to rethink it. This is the exact same process that I go through for developing new solutions or new software, all these things. So this agentic loop piece is so powerful., and this is where I think I’m still

8:16 and this is where I think I’m still trying to figure out how to like best leverage this. So, let me give a little more love to the skill or frame it a little more expanded because give it the a credit it’s due. So you’re talking about,, it’s a set of instructions, but it’s a lot more than that, Mike. Because think of a skill in in an a any agentic solution as almost a package that you would install with Python or Spark because really what it is or what it can be, it can expound to patterns. For example, Mike, I I was

8:48 patterns. For example, Mike, I I was work showing you something the other day and not only can it show basically here’s not just a set of instructions how I want you to follow things but it can even process code for example I have a set of markdown I can provide it a script to say hey whenever I say publish to PDF or whenever we want to do a PDF [snorts] use this CSS stylesheet and use the panda doc whatever pandock to publish this and make a PDF out So you can have the set of code actions or even

9:19 can have the set of code actions or even the set of patterns and context. So it’s it’s I would say saying that it’s just a set of instructions is not giving it the due credit that it really deserves. So it can be codebased but it can also be even for marketing for podcast for even for notebooks. You can have basically hey I’m working with notebooks here. Here are the common functions that I like to use. Here are the things that we always connect to a lakehouse. ask me which lakehouse and then use this API to pull the lakehouse ID from. And

9:51 API to pull the lakehouse ID from. And it has that context and set of different files. It’s more than just a single system instructions that we’re used to. So it’s actually a set of it’s a directory of information that it can pull from. And it has a final the main most important part of any skill is the skill. md which basically says how to pertain to the directory a folder of a skill. And Mike, there is a cool thing in Claude where it’s actually a skill builder. They built a skill skill builder which I use all the time. Yes.

10:21 builder which I use all the time. Yes. And so And so just saying system instructions, dude. It is so much more and I think that’s why it’s called I and I think that’s why it’s called that’s the it’s a great mean that’s the it’s a great context adder, Tommy, because I think this when we talk about the skill, right? Skills are more encompassing than to your point than just it’s a list of instructions. It’s right. You can have So, Claude Code, I think, is leading the helm here on like what is a good skill or what skills should look like. like. Oh, yeah. Oh, yeah., they’ve got skills. They have a skill builder., they’re doing there’s

10:53 skill builder., they’re doing there’s a whole framework that they have around how to use like a a folder structure to encapsulate the skill, right? Here’s the name of the skill. Here’s the description of the skill. Okay, if you need more information, here are examples and code and other. So, you have a whole folder structure that can be used to support what the skill is doing. the neat part about this Tommy is like when we think about agentic experiences while that session context window is open that’s all that it knows. So it has a very short memory while you’re talking to it. And if you don’t supply

11:24 talking to it. And if you don’t supply it with a way to go find and discover what it needs to understand before doing the task, the task, it has no way of having a memory of session to session. And I think this is what people are discovering now is look, how do we build persistent memory skills and memory and these longunning tasks? I see people already talking about like using like an ontology to capture every conversation you’ve had with an agent and when you ask it questions, it can then go back and search and have a really highly

11:55 and search and have a really highly efficient search way of finding old conversations. I think this is how my brain works in general. You have to sometimes you’ll tell me something. Oh, hey Mike. Remember that conversation we had,, a year ago? I’m like, I’m not quite sure. Can you give me a couple more details to remind me of what we talked about? And as you give me some more details, I’m able to recall the memory. Oh, that’s what we were. Yes, I do remember. Let me give you the additional context that I know. And I think this is a very similar approach to what we’re seeing now with agents is skills, memory, and being able to recall this when

12:26 and being able to recall this when needed to use. Do you remember the Matrix? Do you remember the Matrix, right? The movie. Oh, I love the Matrix movies. Do you Oh, I love the Matrix movies. Do what a skill is? know what a skill is? Yeah. Yeah. Do you remember the mo one of the more famous famous Neo would put the thing in the back of his head and I now know karate kung fu. That’s basically what we’re doing skill. and not to let us off the hook too easily. Mike will not forget how ridiculous when we sound when we talked about all this. I was I called a friend too to tell him one of the skills I built and he he’s not so much in the AI more graphic design like Mike

12:56 design like Mike another buddy named Mike. I was like happened to be another Mike or Mike I guess. And I was like, “It’s a claw. It’s a skill.” And he just starts laughing. I’m like, “What?” He’s like, “These names are ridiculous.” I’m like, “Yeah, and there’s llamas, too.” So, it it is it is funny how we still use this lingo. lingo. people if you’re listening and I I we will I’m really interested to know how many people are using the tools that we talk about because you and I let us in the chat by virtue I think just by virtue of our own personalities and being able to be

13:27 own personalities and being able to be independent like we’re always going to be on the forefront just because that’s what intrigues intrigues us and we we’re the we like beta testing but I would love to hear from people if they’re using it if they’re struggling to use it if they’re dabbling in it even of tips for us too., by far by no means are we the perfect experts here or close to that. So, I would love to hear if people are resonating with this or have tried using it because we dang talk about it so much.

13:58 so much. I and and the I think the reason we talk so I’m going to just maybe this is a bit of a justification of of why we talk about this so much. I think one of the reasons why Tommy I really enjoy this one I love learning new things. This feels to me like when we had PowerBI come out. When PowerBI was announced, I felt this is going to rock people’s worlds. It’s going to change how people want to build things. It’s going to be fundamentally different after learning this thing. And here we are 10 years later. The run is still strong. We’re still growing PowerBI. Now we’re going growing into fabric. And I

14:28 we’re going growing into fabric. And I think in the same way that PowerBI made reporting a commodity like a commodity meaning everyone can now build reports, everyone can now create good-looking interactive dashboard things. Very interesting. Now add to that now this new agentic space where before we were just building reports for the masses. Well, how about building entire applications for the masses? How does that influence now? What is that going to do for us? And so I think as I as I look at this and I go, “Wow, this is

14:58 look at this and I go, “Wow, this is really interesting as to what’s happening now, this is what I feel like is the same wave of interest and excitement that we’re getting around PowerBI when it originally came out.” And so I’m very engaged. I’m very We all got to learn. I feel like I’m so far every day I feel like I’m so far behind. It’s happening so fast. my entire world., I used to do a little bit of like social media more so like publishing what we do here on the podcast because that’s really what we do. I do that on social,

15:28 do. I do that on social, but I’ve basically have dropped out of everything not X and not GitHub. GitHub is my new social media platform. This is what we talked about this a while ago. I said if you had to drop one of the three, YouTube, Twitter, Rex, or GitHub. I said the one I would not drop was GitHub. Yep. I think that’s and and now that that has all the more reinforced what we’re doing. And I think I use the the Twitter, the X feed, there’s a lot of really smart people on there that are sharing what they are

15:59 there that are sharing what they are learning about agents and agentic things. A lot of my articles and videos and things that I’m getting introduced to are coming from very smart people there. And then I’m they’re all sharing stuff on GitHub. And so I’m constantly reading through readme files on project after project after project now because there’s so intriguing and there’s so much happening. I got some beef with people on Twitter. Mike, I I gotta share this. And I I do get a lot of my information there just because it it is a good source for PowerBI and especially AI. But is there

16:29 PowerBI and especially AI. But is there a way I can filter out the people who say I use AI and made $10, 000 today just because I use a senic thing. And if you don’t know because they’re so intriguing and there’s so much happening. I got some beef with people on Twitter. Mike, I I got to share this and I I do get a lot of my information. If you don’t do this, you’re missing out. I I hate I like I don’t want too sales pitchy to me. Yeah, you’re right. right. I I’m telling my kids not to use the word hate. And so I don’t want to be a hypocrite here, but they’re they’re not stand people like that who are like,

16:59 stand people like that who are like, “No, I just did this and I made $15, 000 just because I have 1, 800 agents.” First off, no you don’t. No, you don’t. This didn’t happen. I know it didn’t happen. thing. So like I can tell you that thing. to be real like with you. I agree with you Tommy. There’s also a number of messages is I’ve I’ve run agents non-stop all night long. And one of the skills I think that’s going to be needed here and I’m seeing this as a skill because I don’t have it yet. I I’m still working through this. Right. Right now, a lot of my interactions are like do this and it does a good job of

17:30 like do this and it does a good job of like building something. Again, back to my earlier comment around this agentic loop. Build something, test it, come back, fix it. build it, test it, fix it, and then go on to the next thing. That is a very strong pattern. Yeah. Yeah. And the skill that is required here moving forward is how can you orchestrate one, two, three, five, 10 agents at a time. The new software developer, I think this is very much true. The new software developer or the new developer in general will move

18:02 new developer in general will move towards how can I become the most efficient requirements builder? to clearly explain what we’re trying to accomplish and build and then get as many agents as possible to run consecutively at the same time to build something productive as an output. And so I think the new method here is most software engineers are going to need to turn into a architect a requirements gatherer a requirements describer and this is where I think this is going to be really interesting because we have

18:32 be really interesting because we have already these BI analysts that are already doing this like so let’s take the software side away things from things let’s just talk about BI analysts and going to the business and saying what do you need help offline offline we just this is such a good this is such a good transition for these people Like don’t worry about your jobs leaving. You’re actually going to be needed more and what you’re going to have to do is you’re going to have to interpret a lot of what the business needed. Tommy, we had an argument. We’ve had arguments about this long time, which is your team isn’t big enough to support all the

19:03 isn’t big enough to support all the reporting needs of your business. Yeah, I don’t think we’ve ever argued about that. I think we’ve always agreed on that. on that. Well, maybe we we’ve maybe like unpack the concept like,, so I’ve been pushing more for like a federated approach like you need to have someone in both. I I still think that’s very much applicable. But what about the person who’s now a 10x analyst and what if you start using agentic tools to have com like have the like let me see how do I say this again we need to start planning our projects in what is

19:33 start planning our projects in what is the human capital that we need to get the project done and what is the agentic capital that we need to get the project done because this the agentic side can go way faster than the human side. I still need people to talk to my business to understand the requirements to un to structure the data of how we we represent and build our company. But once we have that the the next skill will be business analysts need to take that skill and conform it into these are proper instructions we can go give our

20:03 proper instructions we can go give our army of agents and we can build hundreds if not thousands of reports that meet the specific needs of these people. And to your to my point earlier, which is we don’t need to build lots of reports. We need to build ones that are making you money or saving you money. I got a question for you. The hard part is if you can build reports now with a prompt, that changes the game yet again and you can test a lot more different kinds of reports to see if they’re valuable. And it’s going to increase the need for quality checks.

20:35 to increase the need for quality checks. It’s going to increase the need to say which reports are actually being used and are useful. So the CRUD process create, read, update, delete or destroy CRUD, we’re going to have to emphasize a lot more on the,, updating and deleting side a lot more because there’s going to be so much more stuff created. We’re going to have to figure out how to sunset that. bring up a very interesting topic here and I’m going to ask you a question because you and I have seen a lot of success I think in the AI realm

21:06 lot of success I think in the AI realm and fabric and PowerBI specifically because we have the 10 years the 20 years combined of doing it manually that’s weird to say manually but no but just doing normally right so so making the clicks doing the buttons so you and I have 20 years of that so it’s easy to transition to that conductor side to the orchestrating the agents around Timol and orchestrating the agents around the workflow because you and I know front to back how PowerBI works, how the,,

21:37 how PowerBI works, how the,, data flows through an organization, where it works, where it doesn’t. We know how to gather requirements because we’ve learned the hard way. And by the hard way, I’ve made a lot of mistakes gathering requirements. And we’ve done that over and over and over again. So now I can set up a workflow that replicates that. So the question I have for you, Mike, just like I I want all my kids to work in the service industry in the food industry at some point before they have,, go to college so they learn

22:07 , go to college so they learn and figure it out. My my son just did that and he he learned he learned a lot and he’s like, “Wow, this is hard work.” It is hard work and you got a lot of people to deal with. with. In the same fashion, is there an argument to be made that you don’t give someone AI until they’re an expert in that area? So, because here’s the thing, like great question. Yeah, we were talking about something offline and like building out some documentation and the only reason I had a skill to build it is I knew what skill to build

22:38 build it is I knew what skill to build and I knew what the format and the structure and and the knowledge there. Now I don’t have to worry about formatting or the tedious effort. So, so let me ask you that. Is is that is there a conversation to be had there around

22:52 a conversation to be had there around I’m not giving you AI until you deserve it, until you’re ready for it in your area area there’s a conversation to be had here around how do you deal with the transition versus how do you deal with once the transitions occurred. I think I think what you’re speaking to is about the transition, right? I don’t want to give you AI because it’s too undefined for what it should be doing because we’re actually we’re we’re standing in two

23:22 actually we’re we’re standing in two worlds right now. We’re standing in the old world of like clicks and building reports and doing things manually manually, right? Even that is somewhat automated nowadays. But like we have this manual world of like non-aggentic development and now we’re stepping into this new world which is the agentic development cycle. So I want to look at it two lenses. I think to your point, Tommy, in the transition phase as we move into this new world, you’re right. I don’t think I want to give you tools that are straight up super ultra powerful and can get you really like there’s nothing worse than

23:53 really like there’s nothing worse than giving you a tool that can do everything, but you just don’t know how to use it. Here’s a Ferrari 16-year-old. Go drive to school without crashing. Like, you’ve given too much capability to that person. Now, I would argue after a number of years and some training and education and like so,, you’re the 16-year-old grows up, he’s now older or he or she’s now older and then they now have more experienced driving. Now, you can give a more performant car or,, a Ferrari to to again, aside

24:23 , a Ferrari to to again, aside from the price of that, you never give anyone a Ferrari because it’s just so expensive. But the idea but the idea is like as the skills increase, what you do or what you have capability of doing will change as well. So I would argue in 6 months to a year from now when everyone is just using agentic development patterns what you need to give people will be totally different. We’re not going to be saying you need to understand how the report is built. You can literally talk to the report or talk

24:53 can literally talk to the report or talk to something and say report I need three visuals on this page with this stuff and you don’t have to go through the manual clicks anymore because that actually will be evacuated from the system. It won’t be there anymore. It it it’ll that that that mentality of building will cease to exist. exist. Does that make sense? I I like it. I’m going to make a statement I don’t think you’re going to like here. It’s going to be a blanket statement and all those never go over well. well. [laughter] Well, give me a percentage because

25:23 Well, give me a percentage because that’s usually what you do and that makes me like I’m like I think about 15 and a half% of people will say always use quad code that bit didn’t work. Those bits don’t work. How do we even measure this? I tried those bits. Those obviously did not re resonate. So, well, not with me. They may resonate with other people, but I I like you got to measure it, man. We got to measure it. There’s got to be some way you start with anyways. I think I think 100% 100% in any industry and I’ll take PowerBI. I

25:53 in any industry and I’ll take PowerBI. I don’t think you should have AI capabilities until you’re certified in something so you can be a conductor. Honestly, I don’t as I’m saying that in my head, it sounds from an AI responsibility, which we haven’t talked about a lot. You and I are just like, “Oh, look, a new open source thing. Download install.” And that’s great. But for an organization, if you’re going to hand the range to your point, you give someone a Ferrari, but also take someone not to drive it off the cliff. cliff. So, I would challenge or I’m going to

26:23 So, I would challenge or I’m going to make an argument here and and then we’ll I move on, but that you don’t get AI capabilities until you are certified in a thing you’re going to use AI for. That works in the transition stage. Yes. Once you I agree with you. I do agree with you. Like I think that is the right approach for the transition stage. I I but it’s it’s there’s two stages we’re talking about here. There’s a transition stage of like the early adopters when

26:54 stage of like the early adopters when you’re starting to do this and this would be this would be what I would agree with, right? What is a skill? There’s a lot of knowledge that’s coming along with this, right? That’s the adoption stage. Once we get through the adoption stage, I would then probably say that statement is not necessarily true because you’re already going to have a lot of mental context around how do you use the agent? What can you do with it? And the tooling will get better and the amount of things you can say to it will get better and you will stop having to say you won’t have to learn all the things. This like imagine stepping into something that is so one

27:26 stepping into something that is so one of the one last thought here Tommy we should move on because we do have an actual topic to talk about. The last thought here is in my class when we’re talking about adjunctive development, one of the things that’s very interesting is when you use the PowerBI MCP server, you can use chat bots to talk to the model directly. Hey, what measures do I have? What tables are facts? Which tables are dimensions? I was going through this experience and actually I didn’t really understand what can the MCP server do and so the trick here is you actually ask the agent what

27:58 here is you actually ask the agent what does this MVP or this MCP tooling actually do and how can I use it what are good examples and the documentation of the tooling of the MCP server is actually enough to have the agent tell you hey I’m here these are the things I’m capable of so think of it this way as we get out of the adoption phase and we move into the this is part of our build process. This is this,, standard way we think the tools themselves are the teaching elements

28:28 themselves are the teaching elements you’re going to use to get going. And so the barrier for you and I, Tommy, to teach more people how to use the tools is going to continue to drop to be easier and easier and easier. And as tools get more advanced, we will build into the tools their own documentation because if the agent needs to understand how to use them, it needs to be able to communicate that back to the user who’s trying to use the tools of the MCP server. So that’s I I’m thinking like it it starts today with a we got to have

28:59 it starts today with a we got to have some knowledge so you don’t do unsafe things, but in the future it becomes everyone’s going to know how to do it and the tools themselves will be the educators of how to use. We just got a new topic on the board. I think you’re going to like this one and then we’ll This is our new topic. We’ll put on the schedule. Does AI drive your center of excellence? I think I I think I have it on I have it on our board. So, let’s talk about it. So, all right. We got a mailbag, Mike. Dig it. Dig it. Again, mailbag mail. Let’s hit March Mania. Mailbag Mania., and again,

29:30 Mania. Mailbag Mania., and again, it’s about embedding. So, let’s dive in. Again, guys, people, girls and gals, guys and gals, put your name on. We like,, shouting out who list your first name or just put Tommy’s name on Carlo. Yeah, [laughter] I don’t care if this is Yeah. built generated by AI. Just put your name. We appreciate knowing who you are even if we can’t see your face. Yeah, thank you. We go. Great question. Hello. I heard you guys are making a series about PowerBI embedding. I don’t think we are., but I would love to know the difference of published to web and embedded in a scenario where report

30:02 and embedded in a scenario where report is embedded to a public site. Security-wise, embed is embedded any safer safer when it’s shared with the public. public. Is there a way to get into rowle details like in publish to web? Are there measures one can take to make it more secure? For example, disabling URL query parameters if the underlying data is very granular. So, Mike, right off the bat, there’s a few things here I think we got to correct., really appreciate the question, but a little Yeah, great question. But this is this is

30:33 great question. But this is this is exactly why we’re talking about this one, right? It’s it this is where I think exactly what we were talking about earlier, which is there’s a little bit of confusion around what is PowerBI betting, how would we set it up, what is possible there, and then let’s overlay the security conversation on top of all that. Yeah, Tommy, you want to correct a couple things? Let’s do a few things here. So, first off, p you can’t embed it to a public site and embed a report or if you try to, it’s not going to work for anyone unless they sign in. Embedding in PowerBI is the basically with the

31:03 PowerBI is the basically with the security blanket of Entra ID. You do not get access unless you have access in PowerBI. Even if you put it on your blog, if you put on your WordPress site, it’s going to show an error. It’s not or you got to sign in with your Entra ID. So right off the bat that’s one thing that publish to web available to anyone doesn’t matter who it is if you haven’t got the URL or if you can find the URL you can do it embedded you need a way to sign in in some capacity and I want Microsoft has added so when publish to web so there’s only two

31:33 publish to web so there’s only two options initially right let’s do a little bit of history here there’s another little history here that a little history lesson we’re going to go through this one again this this is the 20 years of knowledge between Tommy and I that we’re going to be drawing on here from this one I’ll be I’ll be pulling from my ontology of knowledge to to get this together here. So, published to web and embedded was initially the landing spot for when we wanted to get reports out to non PowerBI. com users. Okay, so let’s let’s be very clear. Yes. Yes. Yes. When you’re licensing,

32:04 When you’re licensing, okay, so licensing to get to powerbi. com a user to show up there, there’s always licensing involved. And then the the you need like a free license to get to my workspace. You need a pro license or premium per user license to use workspaces. And if you are on F64, the same license. Yes. Correct. And if you’re in a premium per user workspace, all the users of that workspace must be premium per user.

32:36 that workspace must be premium per user. Then there’s this other level which is like, okay, now I’m moving over to the fabric SKs. Fabric SKUs allow embedding from like external embedding, right? Non so these are these are generating reports and sending reports out to the web and nonp powerbi. com experiences. See a report not in powerbi. com that happens at F2 all the way through all the FQs. You can embed things anywhere at any level there. However, there’s two types of embedding that we’re talking

33:06 types of embedding that we’re talking about. There’s so many there’s so many nuances here that’s going on. This is literally like a rabbit trail. That’s why we’re talking about this one. Inside the embedding space, you have a embed for my users or my app owns data and a user owns data methodology. A user owns data methodology means Microsoft is going to supply the identity provider to you to help you authenticate into the report. report. Good point. The user owns data embedding scenario is

33:37 The user owns data embedding scenario is supported at an F64 and higher for internal users of your organization. Anything below an F64, you’re not supposed to use embedded for embedded user owns data applications. If you’re using the embed for external users, so that’s the other the other embedded scenario that runs through all the FSQs. But Microsoft is saying we are no longer going to manage your identity providing. You as the app developer or whatever you’re doing are going to have to roll your own

34:07 doing are going to have to roll your own identity. That’s the advantage of using embedded. Okay, I’m going to come back really briefly to publish the web.

34:13 really briefly to publish the web. Should I pause our time? Is that too much? We need to react. Really the the layman’s term or the elevator pitch there is embedding needs a level of authentication at some any capacity. So So yeah. So let’s let’s Well, yeah. Yes and no. like yes I’m going to say yes but there’s and that’s so good let’s talk about that next okay so we’ve talked about the different licensing who has access to it and again a lot of this embedding thing is what I want to hang on here is what is available for users

34:43 on here is what is available for users to access at powerb. com level that’s that’s a big differentiator for a lot of this this let’s go back to publish the web publish the web was initially you publish this report it goes to the internet that’s it you can’t there’s no restrictions there’s no sign in you publish to web and it immediately publishes the report for anyone who has the URL to go look at it. Okay, it. Okay, inside publish to web that was not a great solution all the time. Sometimes people actually wanted people to sign in

35:13 people actually wanted people to sign in internal organizations for whatever reason. Here’s a publish to web link web link but I need you to sign in to make sure we have access to your thing. So publish to web actually comes in two flavors now. There’s true publish to web and there’s secure publish to web. If you just use publish to web, you’re making the report wide open to anyone. This is great for like government organizations,, anything that’s has to be publicly available. Hey, I’m going to be doing an earnings report on our company and we are a publicly traded

35:44 our company and we are a publicly traded company and I want to just public everyone can have access to Exactly. Exactly. Exactly. Hey, I’m doing a fun little side project and I’m going to build a stock report. stock information is already publicly available. I’m going to go scrape the internet. I’ll build a report and I’m going to publish it. Right? Anyone can see it. So, anything that’s nonsecretive or data that you’re allowing to be shared publicly, that’s a great candidate for just the standard publish to web. to web. The flip side of this is there’s a second option called secure publish to web or secure embed. And that’s where

36:15 web or secure embed. And that’s where you have you basically get a URL, an iframe you can go use anywhere you want. But when a user goes there, they need to sign in. They have to sign in. There’s going to be a button that says, “Hey, I see the report. Here it is. User, you must sign in and authenticate who you are so I can verify that you’re the right person and the you you correct. And so that’s that’s the second wave and that was recently added. So that one’s a newer thing. Secure publish

36:46 that one’s a newer thing. Secure publish to web.”, and I’ll just pause there. any comments around this? No, I think right now you’re it’s spot on in terms of I think I’m just going through the mailbag to see if we were missing any again anything that’s misleading. So the other thing too is with embedded solutions. one of the things that was mentioned in the mailbag is getting the using query parameters. This is a great feature that can be used on PowerBI [snorts] report existing reports or anything that’s embedded in organization

37:16 anything that’s embedded in organization where what I can do is filter the report using the URL. This is this has been around I feel like since the early days I know at least since 2017 where if I want to access the report either in powerbi. com or embedded I can add to the frame or to the URL link a filter query on the model table column and what it equals and this is a great solution to my knowledge Mike and I’m not going

37:46 to my knowledge Mike and I’m not going to be all knowing here but I am pretty dang sure that is not aail available in publish to web for a on a public facing. This is embedded or in PowerBI only. This is a great distinction. So now we’re going to talk some of the security pieces, right? So all of this was just where can I put the report. So let’s Yeah, let’s unpack what we mean by embedding. Right. So let’s talk about the data. Yeah. There’s one thing about the access and what you see and then there’s the underlying information. Exactly. This is the next layer of this

38:17 Exactly. This is the next layer of this onion. So the first layer, the surface layer of this onion is okay. How do we get the report to other people? Where do we want to distribute the report through? So, let me go back to this question here. The the first question was publish to web and embedded in a scenario where the report is embedded on a public site. You most certainly can do that. The standard publish to web experience, the one I talked about, the nonsecure publish to web, publish to web will let you land the report as an iframe with just a simple URL on any public-f facing website. If

38:48 URL on any public-f facing website. If you can put a URL somewhere, that link will work will work bar none. bar none. It doesn’t cost you anything extra. You’re not spending a lot of extra compute. It’s just a service you get as part of that. And you can embed that from anywhere you want. And I recommend when you’re doing things like this, I would highly recommend build a dedicated workspace for all the reports that you’re going to publish to web and just collect them there. That’s that is your publish to web workspace. And the idea is everything the the the design of that workspace is

39:18 the the the design of that workspace is everything we need to publish to web goes into that web into that one workspace. Why do I say this? I say this because it’s a best practice to have a single point of contact so where everything is getting published to web. Also you can give access to create content creators to the workspace and you can then also add in the publish to web setting in the admin settings you want a limited number of people who can actually publish to web. So I think if you’re

39:50 publish to web. So I think if you’re going to turn on publisher web and you’re going to do it on a public website, you need to do some substantial training and even have an agreement with your employees, people that are going to be doing this because if you don’t do that, that, someone may publish something to web thinking it’s just easier and actually be publishing secrets of the company. That is a big no no. And you need to make you’re very clear with that individual. This workspace is only for publish to web., you’re only allowed to publish

40:20 web., you’re only allowed to publish things that are publicly available data. And I need to make sure that you understand the implications of what would happen to your job if you accidentally publish something that’s not published away. And I think that you’re getting to really the model part of the publishing side because there’s one the display. So when I embed a report, Mike, I can still have the ability to do row level security or even column level security or object level security, excuse me. Object level security. Object level RS or OS. Yes, I was like that doesn’t sound right. Correct. that’s

40:51 sound right. Correct. that’s available with embedded again because it follows the authentication publish to web while the model itself is technically technically not available like to see the underlying data. in a publish to web there is no ability because there is no authentication to restrict what’s available or seen in that report. What you see is wizzywig. What you see is what you get for everyone embedded. And that’s with and that’s with publish to web. And then I think

41:21 web. And then I think with publish to web. Yeah. The Yeah. The the other side of this is secure publish to web. And that’s one I got to just double check on myself. That’s one I I’m pretty sure relevant as long as there’s authentication that has an authentication like you have to sign. And to be clear when you’re using secure publish to web it has to be a user in your entra ID and has access to see the report meaning if you’re using relative security and that is set up that user needs to exist somewhere so

41:51 up that user needs to exist somewhere so we can tie who is the user apply the filters to the semantic model and then filter things down the rule of thumb when we’re talking yeah I’m going to call this other example here like you’re saying embedding I’m going to call it pure embedding right pure embedding means I’m not using anything related to publish the web, there’s a property in the embedding experience called custom data. Custom data is what you use to pass the user information into the semantic model using the user principle name or username. And so that handshake

42:24 name or username. And so that handshake between I’m this user showing up to this report then allows the semantic model to say, oh, I now know who you are. I will then apply filters before rendering the report. So there’s no security holes here. The the act of getting the approval to even render the report, you’re already sending in a credential that says I know who you are and I know what you should see. I will only resolve data back to you using your identity and

42:55 data back to you using your identity and what you’re filtered down. Mike, this may be the only time in the podcast where people are actually wishing we went back to talking about AI. AI. [laughter] This is going to get I know but this is necessary. It’s necessary. So it’s necessary but it’s also highly complex. so this is where this is not an easy one. You really not you need to understand the handshaking of like what’s going on at powerba. com? What’s going on in licensing? How do I embed it into this thing? What is this role of security? Where does the identity of the user come from? How do I pass that back to the report? And

43:25 pass that back to the report? And ultimately the semantic model. There’s a ton of stuff here and some other you ton of stuff here and some other really useful things. publish to know really useful things. publish to web and secure publish to web you can only look at reports view only that’s all that it is when you move into what I’m going to call pure embedding pure embedding allows you to view reports add rowle security security you can also have edit reports create new reports you can like what we do in one of our apps that we have inexos we actually let you connect to the semantic

43:56 actually let you connect to the semantic model I can run queries using role security and embed the output of semantic model details. So I have a table explorer where you can explore your data and look at them just the way you would do exploring inside PowerBI. the capabilities of you as a developer that open up when you go into embedding is like 10. This is the misnomer. This is the misnomer a lot of people have on the two options where they’re like well if I’m embedding in PowerBI for most I remember we had the conversation about report developers for most most report

44:27 report developers for most most report developers when they consider embedding they’re considering SharePoint or teams or some maybe internal application where we just fill in the iframe and just simply displaying the report and that’s the idea of embedding most organizations who work pure like very indef F embedding solutions to your point they’re more they’re more developer than report developer if that makes sense right so correct there’s a heavy developer influence on like making it work

44:58 influence on like making it work correctly correctly and usually it’s more client facing you and usually it’s more client facing we have our external tenants or we know we have our external tenants or we have the external clients and then you have the people who are just dumb and they publish the web without the secure and they just say well that’s an easy way to get that available to everyone in the organization not realizing that they have just made all their internal data available when anyone can scrape their PowerBI URL and it’s available even if they no one has the URL. Well, one person shares it or people have found ways to scrape all the available published web URLs. Don’t do that. Don’t do that unless you’re sharing baseball

45:29 do that unless you’re sharing baseball data. Don’t do that. But I think the misnomer is to your point most people think I’m embedding a report or securely embedding a report. I’m not I don’t have

45:40 embedding a report. I’m not I don’t have the ability in a sense they don’t think of what you’re talking about. I can securely embed embed the model and the underlying data which opens up a lot of features. Most organizations that I work with when they say embedding they’re embedding in SharePoint they’re embedding in teams they have their own internal application they want to display a report internally for the organization and Mike you’re talking about another level here of to use the affformentioned onion on the other use

46:10 affformentioned onion on the other use cases here because I I’m going off the mailbag and then I I know I’ll let you I’ll let you drive let you spin Tommy you’re going you’re great the user here I’m I’m curious curious as I I always whenever I look at the questions people ask us I’m think about what’s their situation where did this question come from like what there probably something’s going on in their organization that they have a project on or there’s a conflict and to me I’m looking at this person and I I hear someone just basically saying I need to get this data out to a lot of people and

46:40 get this data out to a lot of people and and I need to do this without a lot of peer development work and that’s where that question comes from but Mike I want to I’ll stop there because I want to hear from you. I know you built some products and I’m curious why would someone take the next step outside of SharePoint and Teams because one if I have PowerBI I’m always going to drive people to PowerBI. Sure. So what are the other situations here?

47:10 situations here? Yeah. Yeah. Yeah. What’s the big move? What what So what’s the benefit? This this is going to come out in Yeah. This is going to come out in probably two different points. The first point is you’re removing yourself away from a lot of the PowerBI autobuilt things that just makes it easier for you., so there’s,, they have the the portal, they have a lot of these things that are there that are easy to use, a lot of,, bookmark, you use, a lot of,, bookmark,, all these things that you can kind know, all these things that you can kind know, all these things that you can can can get for free because of can can get for free because you’re using powerb. com. There’s a lot of things you have to rebuild. Embedding is very easy to do.

47:42 rebuild. Embedding is very easy to do. The complicated part is the management of the authorization part. And so what let me explain what by that. When you publish a report to the web, you have the report which is an item. You have potentially bookmarks you want to use against that item. You have multiple users who may or may not have access to the item. You may have ro security you want to apply to the item. The embedding is simple. The management of all the data and the objects of who has the right permissions to the right

48:13 has the right permissions to the right things. Can I look at this report? Am I able to make a copy of it? Do I want to build another one? Do I want to be able to edit this report? There’s a lot of other data points that are being used to make the experience really good. Right? You have to manage that. So the easy part is embedding. The more difficult a part is getting the right authorization patterns that let users create, generate, and reuse content that you embed. So that’s the hard part. What are the advantages of going towards embedding? Embedding allows you to

48:44 embedding? Embedding allows you to authenticate with anything you want. So one of the major value ads I think here is when you go to pure PowerBI embedding you have full control of the role of security you have you can do multiple key partitioning you can do advanced security settings but you also don’t need any users in your entra ID a lot of organizations have this really hard line as to I need to share data with my external customers but I can’t add them into my guest network into our entry ID

49:15 into my guest network into our entry ID they just don’t want to do it. And I agree, you shouldn’t you shouldn’t be bloating your PowerBI or your your Entra ID tenant with a whole bunch of guest users to use your apps or to share data with them. That doesn’t make sense. So the biggest advantage I see here with embedding is you’re able to roll your own security. You can use Azure BTOC. You can use Google Analytics or Google signin. You can use whatever you want. You can use GitHub.

49:45 whatever you want. You can use GitHub. You can use anything you want. If it has an identity provider, you can get it out of there and use it. A lot of companies have existing applications with an identity provider already built in. It’s already there your own sign on into your solution. So in lie of that right in lie of having that solution already pre-built for you there you have to do a build around how do you manage that that that group of users

50:17 manage that that that group of users with that that identity and that’s that’s the key part you get the Ferrari but it is stick shift and that’s what you’re telling me here because you have all the capabilities but you better dang well know how to drive that because you don’t have that net, right? Because you got to really all build that. And again, I’m I’m failing to see Mike though, unless I am publishing this for a client where I already have the set users, where does something else come into play or where does the added benefit, right? Outside of the identity,

50:48 benefit, right? Outside of the identity, right? Because the identity alone is a good point, correct? Really solid solid one, but that’s not the only case. It’s a typical stopping point for most. So that is the probably the most common story the stopping point of most organizations around doing embedding. Look, Look, we want to do embedding with power. com. It makes a lot of sense. Let’s do it. Great. You’re going to need to add all these new users into power into your entry. And you’re like, nope, that’s not acceptable. We can’t do that. So the big advantage of this is roll your own security. Let’s also think about the cost advantages of this as

51:19 about the cost advantages of this as well. So I’m paying for an F2 up to an F whatever skew it as well. When you move from userbased licensing, when you bring users into your PowerBI environment, you need to pay for licensing for them to use the workspace. So if your external sharing experience is I will make a workspace and I will add those users to my entra ID, any users coming into that workspace, unless it’s on an F64, you need to pay for a pro license for them to go use that., and that’s not a

51:50 to go use that., and that’s not a great experience for those users and price points. So now when you do embedding, you now have control of I’m paying for dedicated capacity and if you build good models, efficient design things, you can service models to a much higher audience. So let me give you an example here. I run the website themes. p powerbi. tips. Themes. parbat PowerBI tips renders a real visual for users when they use the tool. You go to the bar

52:21 they use the tool. You go to the bar chart, you can set the properties of the theme file for the bar chart and your bar chart renders for you. A a sample report. It’s a sample bar chart. My model is very efficient and I service well over well over 3, 000 plus users a month on these calls to go get these visuals out on an F2. So the user base is really large on a very costefficient skew. And so that is an advantage here as well is you can then

52:52 advantage here as well is you can then better tune your financial model around getting the right users the right models at the right time for not paying a per user licensing. So there’s actually a great reduction there on the licensing agreement as well. And you can the other advantage I would also argue here too is there’s a lot of companies out there that have built accelerators to make the solution easy to use. Right? Embedding is not easy. It’s quite difficult. You can actually go into the open marketplace and find real solutions. Mine happens to be one

53:23 real solutions. Mine happens to be one of many that are out there and it’s called Intellexos. And you can do a lot of these other things. so for example what we’ve built in addition to just rendering the report we’ve built copy report edit report build pageionate reports data exploration data shares. So our embedded solution actually incorporates a lot of other what we think are really valuable feature sets that companies will want to use all built on fabric. It’s all on top of the

53:53 built on fabric. It’s all on top of the fabric ecosystem. And so you license our application that sits on top of PowerBI, your FSQ, and then you turn on the FSQ, you pay for it, you license our software, and then from there you have access to all these really rich features, things that even PowerBI. com doesn’t give you. I would argue though, Mike, that I’m hearing great use cases, but very specific use cases, right? So like to your point, the PowerBI tips, it’s a great use case. Not everyone has a

54:23 great use case. Not everyone has a PowerBI. tips type of website or or theme theme but everyone wants to share data externally like so there there’s a m there’s a maturity curve in every organization and and this is how I see it you start with internally we need to figure out okay what are we doing now let’s move away from SSRS let’s move away from Excel a little bit let’s incorporate PowerBI you find wins internal to your organization around using those features great love it the next step is your business is almost always and

54:53 your business is almost always and there’s a lot of business you’re almost always working with other vendors, suppliers, and other people who need your data. And so this is it’s happening. People are going to your PowerBI reports. They’re exporting large chunks of data and they’re sending it via email to your customers or your vendors or whatever. You have to share data. The PowerBI embedded experience is a much more catered experience to helping you share and distribute data external to your organization. And I would, again, I’m comfortable with it,

55:23 would, again, I’m comfortable with it, but I would choose anytime I’m sharing data externally, I’m always using an embedded solution to do so. Really Really 100%. It’s so much easier. There’s so much extra friction and licensing and cost by letting external users into your tenant, tenant,, but it’s part of I think if you’re No, I honestly this is where I would agree with you because anytime I’ve seen any embedded solution or actually let me let me even make that simpler to say. Sure. Anytime I’ve seen an organization have a need to share data outside their

55:55 have a need to share data outside their own organization and usually B2B correct correct there’s always there’s short-term what people think but I always have people because it always happens is think about the scalability of whatever they decide to do. This is the critical point for me Mike because I had organizations like well we only have five people right now so why do we have to set the simple I only send three reports a week. I I only send three reports a week. That’s all I got. But it starts with three reports, but when does it turn into like 15 reports? Four reports. I always ask them like, “Is your plan to

56:26 I always ask them like, “Is your plan to have five users in a year from now?” They’re like, “Well, no, because we can make some money off this and we make it people.” Okay. It’s like, “Okay, does this work for five people?” Yes. I would say if you’re only going to have,, five cases. the sharing the fabric the workspace is great great but odds are that’s not what it start what it stays with and I I think embedded especially I think a lot of things we do are this type of situation but none more than

56:56 type of situation but none more than embedding because it’s so hard to change that once you start so whatever solution you do it really is always it’s always

57:05 you do it really is always it’s always the case of let’s say you decide to do the single users in a workspace fine once you want to change that. That’s a tough thing to do once you have that set up, once that ball’s in motion. And honestly, yeah. So, I I completely see this where and and this is what I always push people. I always ask them, what’s the plan a year from now or six months from now? It it’s it’s the adoption curve, right? You first adopt PowerBI. You get comfortable with that solution. You start having successes and wins internal to your company with internal dating

57:36 to your company with internal dating data. data. Everyone loves it. They’re like, “This is working for us. It’s making it easier. We’re automating a ton of data loading, right? Love it. That’s the great step. But as m as organizations mature, they’re going to need to share data externally outside of that. And the the way the PowerBI environment does it is just not as easy. And again, Microsoft made PowerBI embedded for this exact reason. And the embedded accelerators is the sweet spot because you can get the full embedded multipleyear development cycle on embedded solutions and start them in

58:08 embedded solutions and start them in less than an hour. like you you can go from zero embedding to embedding in minutes almost like so I’ve I’ve done these deployments with customers I’ve done well over 50 plus custom embedded solutions multiple environments dev test prods all the things so I’ve taken all these learnings of like what people are telling me that is useful and valuable to them I’ve distilled them all down into here’s a product off the shelf go to the Azure marketplace click get now it will install everything you

58:38 it will install everything you It runs in your own infrastructure. I’m not,, privy to any of that information. It comes right off of the store. So, this is this is where our product intellects, I think, really shines is we’ve really tried to provide the white glove experience. Another thing, Tommy, I’ll point out here. Let’s imagine you want some custom branding. Companies want to use PowerBI, but they don’t want to tout the fact that PowerBI labeling is all over the product. So you need custom label branded solutions and that’s what these embedded accelerators do. Like I don’t want to have the log. I

59:09 do. Like I don’t want to have the log. I don’t want to go to app. power. com. I want to go to like reports. pulabi. com. Like that’s what I want to go to. You can’t change the domain name in powerbi. com. It just does not happen. So what this is giving you, it’s giving you a high level of white labeling customization to really and the again the end of the day here is think about making your customers sticky to you. You give them good data like every product I’ve ever used Tommy on social media to try and get data into stuff. It’s a pain to get data out of

59:39 stuff. It’s a pain to get data out of stuff. The sooner you can make getting data out of your systems or working with data in your systems a dream, the more your customers are going to love you and they’re never going to leave because the next guy who doesn’t give them that experience, they will not leave. They will want the UI and make it pretty, but at the end of the day, if they can’t get their hands on the data they need to do their job or their business, that’s a that’s that’s a no-go for me. And no a amount of AI you can throw at that will give them the data they need. Yeah, this is a this is your proprietary

60:10 this is a this is your proprietary information. Well, yeah, and I know we’re getting to this. I’ve really enjoyed having this conversation because I again a lot of people mistake not just the complexity of it, but again what they’re trying to do with it. And I think I’m glad that you and I are are leaning that way where if you are planning some embedding, there’s a need for that for client facing or whatever the case may be. be. Yep. Yep. Have what you need now. But understand what the company’s trying to do with that because anytime you plan for the

60:41 that because anytime you plan for the short term with embedding it’s gonna fall on his face because it cannot scale. Usually you got to pick a scalable there are easy ways to get started right away. You could get the people you can give people access tomorrow today and before lunch you get people but if your organization has a bigger plan for this product usually it’s a product that they’re you usually it’s a product that they’re they’re servicing here. know they’re servicing here. Correct. Correct. what’s that? or you have an existing product that needs more data in it. Like again, this is another one of these things like you’ve got a data product that’s doing things, you need a

61:11 that’s doing things, you need a reporting. So, I’ve done a lot of these too. There’s a there’s an app that runs that does all the data things that that’s like actually generating the data. It’s sitting somewhere on the back end. end. You need another reporting portal that helps all the analytics of it. So, you don’t need to build it back into your same product. You can add these accelerators on on the side. So, it has and I I I’ll give you I’ll give you some flowers. I don’t like doing that often, but the the Intellectus app is a great starter because previously previously if you wanted an embedding

61:42 previously if you wanted an embedding solution like that, well, we’re in it for two months before your boss sees anything. anything. Not even not even No, I’m saying previously. No, I’m just saying saying the previous developments I’ve seen companies spend anywhere between six, nine and 12 months of development time of trying to get the requirements together to understand how it works to build the features they want and that the the six month side of things is like I’m just getting embedded stuff up and running in my application with permissions and then you’ve got to

62:13 permissions and then you’ve got to figure out how do I do the edit stuff? How do I build automatic page reports? How do I build functionality in the all the other things, right? Yeah. you and and I think that’s usually why people go with the short term because he’s like listen I know we need to do the embedded solution and I know they want to get 100 people on here but my boss is going to kill me and no one’s going to be happy if I tell them six months before they even see anything. This is why the accelerators are so useful. They’ve already done a lot of the leg work and it’s just pull it off the shelf. Yeah.

62:43 Yeah. Light it up and get and get going. That’s why it’s such a good So this is a great topic. Yep. I do want to cover off on the last couple questions here. I want to make sure I directly answer every question here., security-wise, is embedded any safer when it comes to sharing with the web? The answer is no. You don’t get relevant security., it is a publish to web is whatever they see, they can see in the web. If you need security, you’re going to need to think about the embedded the the custom embedding solutions and adding your own real security. So, it’s not any I’ll call quote unquote safer. You’re still using

63:14 quote unquote safer. You’re still using HTTPS to service the data. Mhm. All that is standard. It’s standard web protocols. All that is the same level of secure. It’s the difference of are you needing to add real security and publish to web isn’t more safer when you’re going to share with public. You really do want to have a layer of authentication on top of your applica on top of the reporting. Thousand% thousand%. Is there a way to get row rowle details in to publish the web? So can we get rowle security details into publish web? No. The answer is directly no. Publish

63:45 No. The answer is directly no. Publish to web does not support rowle security into that reporting. Secure publish to web I believe passes the credentials from users into the report and so if you did have row level security applied secure embed will at least allow users to sign in in order to see the report with some level of security there. So that one would be one that would be an option. But again, you now need to add every user who needs to see the report into your entry ID. And yep, if you’re a guest, you’re gonna have a really weird like URL or email address name of them.

64:16 like URL or email address name of them. And you’re going to have to make sure you add that really complicated email address into the report. So that’s another one that you be mindful of., so the answer is directly no there. Are there any measures one can take to to make it more secure for embedding or disabling URL filters, query parameters, and underlining data if it’s granular? If you’re using embedded, so publish to web, you can use URL parameters. There is no guarantee that those will be more or less safe. They’re just going to assume everything you publish and publish to web is 100%

64:48 you publish and publish to web is 100% exposed to the internet. All the report, you’re not going to use URL parameters. You cannot change those to have it you can’t restrict them when you’re using publish to web. The only way you can add security is apply the security before the report is data rendered. And what by that is that’s the whole rowle security adding the additional layer of when you ask for the model and report to be rendered you need to give the credentials of the user to the model

65:18 the credentials of the user to the model pre-render the data and have it sent through the embedded experience. So embedded, yes. Published to web, no. So I think we’ve answered all the direct questions. Does that cover everything? We’re Dude, I think we’re perfect, man. Okay, good. Well, great topic. I love this topic. As you can tell, I’m slightly passionate about embedding. I think this is so cool. I think this is really what people need., and also, honestly, Tommy, I love powerbait. com, but it’s a bit busy in my opinion. Oh, this is where we differ. I think I

65:48 Oh, this is where we differ. I think I think in the way that Tommy you say there’s so much training that needs to happen inside powerb. com I agree agree you do need to tell people how it works you do need to get people to navigate through the system sometimes people just want go to URL see list of reports click on report and so I do think that embedded also gives the the ability to really simplify the user’s experience about how do I want to go see a list of reports and get them to my eyeballs.

66:18 reports and get them to my eyeballs. Right. And I do think you can have direct links, you can have workspaces, but powerbi. com has a lot of additional features that are shown to standard. You get a pro license. There’s a lot of other buttons you can click on. Useful, yes, but a lot of other things. I think sometimes in in organizations there is still a need for super simple, super easy to get access to. Here’s the list of reports. Just go get them. get them. Okay. Another day, another topic. I like that. I like that. But no, great

66:48 that. I like that. But no, great conversation today. Awesome. Well, thank you all very much for listening today. This has been a super fun episode. If you like this, if you want to hear more about embedding, let us know in the comments. I What is not clear to you? If we said something that didn’t make sense, let us know in the comments. We We’ll be happy to to watch them. Tommy and I watch the comments on the YouTube channel. We’ll we’ll respond if we can, and if we need another topic, we’ll do it again. I think this is another great topic to continually unpacking with the community. There’s a lot of opportunities for organizations to deeply integrate PowerBI into their apps

67:19 deeply integrate PowerBI into their apps and their applications using embedded. Off-the-shelf solutions are a great accelerator. It gets you started right away, but you can also build highly customized solutions in your product that makes your customers sticky. That’s the story here for me. Make your customers love you with the data you provide. That’s what embedding is for me. That’s that’s what my I think that’s exactly why you do this. Anyways, Tommy, where else can you find the podcast? Well, you can find us on Apple, Spotify, or wherever you get your podcast. Make sure to subscribe and leave a rating. It helps us out a ton. Do you have a

67:50 helps us out a ton. Do you have a question, idea, or topic that you want us to talk about in a future episode? Head over to powerbi. tips/mpodcast. Leave your name and a great question. And finally, join us live every Tuesday and Thursday, a. m. Central, and join the conversation on all PowerBI. tips social media channels. Thank you all so much, and we’ll see you next time. Explicit measures. Pump [music] it up. Be it high. Tommy and Mike lighting up the sky. Dance to the day to laugh in the mix. Fabric [music] and A. I get your fix. Explicit measures. Drop the

68:22 your fix. Explicit measures. Drop the beat [music] now. H feel the crowd. Explicit measures.

Thank You

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

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

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