Interview: Microsoft's Arun Ulag discusses updates to the Power Platform, and its future

Microsoft's Ignite conference this week was completely virtual, much like its Build 2020 developer conference and Inspire 2020 event held earlier this year. Among a flurry of announcements and unveilings regarding various Microsoft platforms and services, a bunch of updates for Power BI and the Power Platform as a whole were revealed as well.

Some of the major announcements for this event included the arrival of low-code updates in the Power Platform for Azure and GitHub, Power Automate Desktop and Power BI for Teams entering preview, and new Power BI Premium offerings being introduced. Following the Ignite sessions for Power BI, we got the chance to conduct an interview with Power BI CVP Arun Ulag, going over a brief history of Power BI and the Power Platform, insights into the latest announcements, and what the future holds for Microsoft's business intelligence platform.

Hamza: Could you start off by perhaps providing an overview of how exactly Power BI has evolved over the past few years?

Arun: So, the Power Platform is Microsoft's end-to-end low code development platform. And what we're really trying to do, Hamza, is help customers go from insight to action to automation, and empower everybody from citizen developers to professional developers. And in the Power Platform family, it began with Power BI. So, Power BI was the first of the power platform. And we launched it about five years ago, in July 2015. And since then, it's become a whole family. So you have Power Apps, you have Power Automate, you have Power Virtual Agent. My colleague, Charles Lamanda, runs the applications and automation platform, so Power Apps and Power Automate and Power Virtual Agent. That's kind of how we divide and conquer.

So, that's the Power Platform, and the Power Platform shares a bunch of things in common. We have a common design team, like, it's the same designers that work on my stuff and Charles's stuff. It's a common gateway back to on-premise systems. It's a common set of connectors. Power BI allows you to embed Power Apps into your Power BI report. So you can take an action from right there. You can embed a Power BI report into a Power App, you can bring a Power Automate flow into a Power App or into a Power BI report. All of them are designed together and they're designed to work together. A customer can go from insight to action to automation very, very quickly.

And all of the Power Platform is extensible to Azure. For example, you know, Power Apps, you build a Power App, you can call an Azure function. You use it using Power BI, you can call an Azure Machine Learning model. All of this is designed so that there are 'no cliffs'; it's all the high productivity of a low-code platform without the cliffs that prevent you from scaling out because you can always drop into code, you can always drop into Azure. That's the overall vision for the power platform. Let me see if you have any questions on the Power Platform before I double click on Power BI.

Hamza: I was actually reading your blog post that you published the other day. The term 'data culture' has been used around a lot whenever Power BI is mentioned. Could you comment on what exactly data culture is, and how driving a data culture has been of paramount importance to your platform?

Arun: So, even six months before we launched, we basically said, "Hey, our objective is to help our customers drive a data culture". And what we mean by that, Hamza, is that we are really trying to empower everyone to make every decision with data. Our objective is to really get Power BI in the hands of everybody on planet earth, but at least starting with the billion users of Office. You know that BI has been a very fragmented industry, right? So, if you look at any reasonable sized company, they have at least two, three, half a dozen, maybe even a dozen BI tools that they use within the organization.

In spite of having so many BI tools, if you look at the penetration of BI tools in the organization, it's only about 15 to 25% in most companies, even looking at the information worker population. It basically means that this industry has been in a place in which there's a profusion of BI tools and very few people have access. Our vision for Power BI is that we want to put it in the hands of everybody so that they can make every decision, not just the big ones, but the small ones, the medium ones as well - the sales person deciding which customer to call, the customer service person being able to diagnose the issue. We're really trying to help everybody from the CEO to the frontline worker make data-driven decisions. And it's influenced a lot of our thinking, like, literally every step of the way.

The fact that it's a global cloud service means that we talk about it as 'five seconds to sign up and five minutes to wow!', which is, "Hey, anybody should be able to sign up within seconds and get real business value within minutes." If you don't do that, then you're not going to serve a billion people. We made Power BI Desktop completely free, the authoring tool. It's not free for a period of time or a limited time. It's just completely free. And it was a shock to the industry when we did that, because I think at that point, the tools were like $2000 a license. And Power BI Desktop is completely free. Why? Because we wanted everybody to try it. And when you get value is when you share the insights that you find with others.

That's the point where you create value. And we said, at that point we will monetize Power BI. And by the way, when we monetize Power BI, we said, "Hey, we will pick a price point that, again, allows you to go drive everybody to use it." Which is why we started at $10 per user, per month. And then you get further discounts if you pay by through Office 365 E5 - which includes Power BI - or you want to distribute at scale, in which you buy Power BI premium and the price drops to maybe $4 per user per month. So really, its priced at a point in which you can truly drive a universal adoption.

The last thing I would say, Hamza, is that we care so much about our community, and yeah, it's actually unbelievable. And the community cares back. We started this thing very early on of having the community vote on And that idea itself is not unusual. Some people do it, but what is unusual is how seriously we take the customer feedback. If you go into, you'll find 24,000 ideas that the community has voted on half a million times. Me and my team look at that list of features that they ask for every week, top to bottom; we pride ourselves on how many votes can we take off the table. We've covered 200,000 of the 500,000 votes so far, and we ship features every single week in the Power BI service and every month with Desktop, and we've been doing it for five years. So, the pace of innovation focused on what customers want is something that nobody saw in the BI industry and generally it's been the first for Microsoft as well. So that's kind of how we think about the data culture.

Hamza: So, the community is then very significant in terms of whatever sort of direction, or the decisions, that your team decides to make.

Arun: Absolutely, you'll see that we take the feedback very seriously. And if you go to and you sort by the top ideas, you will see 'under review, 'in development', 'shipped', etc. Each time we do this, it really creates a positive cycle because it says to the community that, "Hey, we take your feedback seriously. We ship regularly every week, every month." Which means that, it gives them incentive to engage more because then they see that Microsoft is listening, Power BI is listening, and we care and we pride ourselves.

I think one of our my most favorite examples is when a community member tweeted saying - this was a really small feature, to be honest - but they tweeted saying, "Hey, Power BI, can you get this done?" And a bunch of other people chimed on and our CTO, Amir Netz, chimed back in saying, "We're getting it done, it'll ship next month." And it was a brand new feature. It just shows that how much we care, and I think they love us back. So, if you see the passion on Twitter, you see the passion on our community, you'll see that this is just a positive self-reinforcing cycle.

GIF via Power BI blog

Hamza: Jumping directly into the Ignite announcements regarding the Power Platform that have been unveiled so far, we've seen more low-code updates being introduced that directly link to other platforms. So, for example, Azure API management connectors can now be used to build Power Platform connectors. And then, of course, the Power Platform has been introduced to GitHub as well. With changes like these, is bridging the gap between the services, by introducing more connections between them, or more ways they can interact in, something that you all are focusing on?

Arun: Absolutely. I think if you think about how we're trying to help customers, let me take a Power BI perspective and then I'll expand to Power Platform overall. So, we think about three things, Hamza, one is, how do we empower every user?

There, we're really just trying to do two things. One is, we're trying to basically make Power BI Desktop PowerPoint for data. And why PowerPoint for data, because everybody knows how to use PowerPoint. So, things like making the ribbon the same ribbon, making sure the theming exactly matches PowerPoint, the way you do grouping, the way you move things around, all of those things exactly matching PowerPoint is because we can get everybody to learn how to use Power BI very quickly. The second part of empowering the user is the AI capabilities in Power BI; this is one place where we really had a breakthrough. Of the 150,000 customers that Power BI has, over 75,000 today use the AI capabilities in Power BI. Nobody else has this level of broad AI adoption. And that's really because we are able to lean on Microsoft Research - which does a lot of fundamental advancements in AI -, basically 'steal' their technology, and create business user and business analysts-focused experiences.

We have introduced two new capabilities in my Ignite blog. One is Smart Narratives, which automatically summarizes the insights. The second one is Anomaly Detection, which we demoed, but it's only coming later this calendar year. And it finds anomalies that you should be paying attention to. So that's the first area of investment, empowering every user.

The second area of investment is empowering every team. That's where a lot of our integrations come in. From a Power BI perspective, we start with Teams; it's probably number one for us, it's probably number one for the Power Platform overall. Hey, you and I are on Teams today, you know. Teams has over 75 million daily active users. And if I think about it from a BI perspective, BI is inherently a collaborative experience. The insights we share are more valuable than the insights we have.

What we've announced at Ignite is that Power BI now has an app that ships as part of Teams, where you can not only do your collaboration, but it's also your home for analytics and data. So, Teams is number one. The next one that we're investing in is Excel. It is the world's most widely-used tool for analysing data, for working with data. I think more customers use Excel today, then all of the other BI tools - including Power BI - combined. And there's probably an order of magnitude difference there. The challenge with Excel is that, because it's so easy, Excel tends to be everywhere with all kinds of data in it. And customers refer to this as 'Excel sprawl'. Everybody has a copy of the data, they're emailing it around, they're putting it on shares. And you don't know if the people in your organization are making your decisions based on the right data. Is it fresh? Has it gone stale? Are people making decisions based on data that they should even have access to? So, all of this ends up being a challenge for customers to use Excel effectively.

What we're doing with the Excel-Power BI integration is, firstly, Excel now just ships with Power BI. You don't have to install anything, you don't have to pay anything. Excel now has native PowerBI integration. What it means is, when you bring a data set into Power BI, it has a lot of security associated with only certain people having access. It has role level security, so what you see might be different from what I see. Like, I might see U.S. data, you might see Asia data; it has certification. I.T. can say, "Hey, this is certified." It's I.T.-approved. It also has MIP labels - Microsoft Information Protection labels. You can say this is highly confidential, in which case the data should always be encrypted. It's continuously refreshed, you know, 48 times a day or through Direct Query, etc. So we're saying Excel can now natively discover those datasets, without leaving Excel. Now, when people do that, you know that they're using the right data, you know the data is not stale, you know the data is secured based upon permission-based access and role-level security. And if the data has a MIP label like 'highly confidential' applied, Excel will inherit that label and encrypt that Excel file. This basically means that now you can solve Excel Sprawl, and everybody can use Excel with the right data. So, that's a major investment for us. This is again something nobody else has an industry. That's the second major integration.

The third one is about the Power Platform - so, across Power BI, Power Apps, and Power Automate. I didn't spend a lot of time on the blog on this one, but we are actually doing a lot of work to make sure that cross-solutions building all three work well together. That's the second point about empowering every team. And the third one is about empowering every organization. And this is really about all the capabilities that I.T. cares about. Large-scale semantic models, real-time data analytics, application lifecycle management, with GitHub and Azure DevOps. All of these capabilities that I.T. cares about, you can scale systems end to end, reach tens of thousands, hundreds of thousands of users, those are the things that we are focused on.

And this week, we announced major improvements to trust, enhancements to Information Protection, but also deeper integration with Azure Synapse Analytics. Between Power BI and Synapse, it actually becomes a self-learning, self-tuning system. We'll work with data where ever it is, but if you happen to use Power BI with Synapse, the two systems will work together where they're constantly learning from the end-user usage patterns and then optimizing themselves to render the best performance. So, these are the major announcements for Power BI.

Image via Power BI blog

Hamza: I also had a specific question regarding Power BI premium. Basically, you guys introduced two new Premium offerings, and one of these was the Premium Per User capability. Interestingly, this is said to provide Premium features' access to only a select subset of users within an organization. So, how exactly does this differ from Power BI Pro, which already provides licensing to Power BI on a per-user basis?

Arun: Yeah, that's a really, really good question. So, we introduced Premium about three years ago. And it's grown like crazy because it gives two major capabilities, Hamza. One is, it allows unlimited distribution to consumers, as much as the capacity can support. And because you're paying for your own capacity, we allow you to run advanced workloads that potentially require a lot of compute: things like building Automated Machine Learning models, you know, things like Cognitive Services, so you can analyze text and images; things like pixel-perfect paginated reporting, or to be able to run really large data models - the 400GB data models -, or do large volume data preparation at scale on Azure Data Lake. We've introduced a whole bunch of premium capabilities that really have moved the state of the industry for BI forward.

However, the challenge has been that because it starts at $5,000 for P1, it's not been accessible to a lot of departments and a lot of small and medium customers. The ask has been, how can you get an entry point that is much, much more affordable? Which is why we're introducing Premium Per User. So Premium Per User will go to public preview before end of this calendar year and be completely free during the public preview period. Anybody can just sign up and experience all the capabilities of Premium: large models, 400GB data models, about four to eight terabytes of data in memory, Cognitive Services, Automated Machine Learning, all of that stuff on a per-user basis for free during the public preview period. And then once it gets to a GA, we will announce pricing and it will be very, very competitive in the industry. So, that's how we think about Premium Per User. I was surprised to see how much excitement there is for Premium Per User. I think that just because so many people wanted these capabilities, but the $5000 per month price point was just too expensive as a starting point, now this really unlocks a massive new market for us.

Hamza: Right, so basically it's providing a more affordable entry point.

Arun: For the capabilities that customers have had, but they haven't been able to get their hands on, yeah.

Hamza: Last year, Amir Netz, your CTO, mentioned how Azure targets more of the technical community - developers and the like -, while the Power Platform focuses more on serving the business user community - that may or may not be well-versed as far as the coding part is concerned. So, as a whole, are your objectives now as the CVP of Power BI still the same, or do you feel that you will expand a bit on the target community looking ahead.

Arun: Yeah, it's a really, really good question. And thank you for asking this. Our focus is primarily on serving business users and business analysts, right. And business users are business users, you know. They're trying to get their job done. All of the glories of Power BI are not that interesting to them. They really want to understand what's going on with their business. But this is analysts - we see a spectrum, Hamza. You know, we have folks that are very likely skilled, they're basically Excel users. And then on the higher end, we see folks that are, you know, pretty data savvy. They can build very advanced Excel Workbooks, they have a reasonable familiarity with data science. Some of them have probably done some R and Python coding. So, all of them are rock stars with Excel formulas and pivot tables.

We see that spectrum, but all of them, you wouldn't consider them as developers. And that is our primary audience. However, when we see a large customer using Power BI with the data stack underneath it, we do see a need for us to operate well with the data stack. There are data engineers or BI professionals that also use Power BI, but when they're using Power BI, they're typically not just serving businesses, they're building a large end to end system. There, the tools that they would primarily use would be, Azure Synapse Analytics, Azure Data Factory, Azure Machine Learning, you know, professional tools. What we're trying to do is saying, "Hey, while you're using those professional tools, you can also use Power BI and we've integrated our development environments." So, you can build a project into it. That is what we're focused on. We don't want to confuse customers by saying, "Well, if you're a pro data engineer, you start your job in Power BI," - you would start your job in Azure. But when you build systems end to end, we want to make it easy for you to build a full stack.

Now, when I expand beyond Power BI and look at the Power Platform, we actually see a lot of pro app developers who are looking for a high-productivity environment. But if you want to build an app within your enterprise and you want to have it run on your iPhone, and you want to run it on your Android, if you're gonna use the native APIs of iOS and Android and get each app for two different stores, that's a very expensive process, even for the pro developer, who knows what they're doing. Now, Power Apps really helps you there because it gives you a high-productivity environment. We are seeing a lot of pro developers who have started using Power Apps and Power Automate as a place to start. But, what's great about it is that because we have done a lot of the work with API Management and Azure integration, they can actually build both on Power Apps as well as in Azure Functions. It gives them the ability to build the whole system end to end using a high-productivity environment and a high-code environment.

Image via Power BI blog

Hamza: Thanks a lot, Arun, I'm just going to wrap this up. As far as the future of Power BI is concerned, what more is in the pipeline for the longer term, and how do you see the platform evolving, moving ahead - especially in the wake of the coronavirus pandemic?

Arun: Yeah, that's a really, really good point. I would say three things, Hamza. One is, it starts with trust and security, right? Everybody's working from home these days. Corporate data is now living on so many machines that it never used to see before. Enabling customers to be able to work securely in a remote environment is paramount because often customers put their most precious information in to Power BI. That's why we have invested deeply in Microsoft Information Protection; that's where we invested in cloud app security, that's where we added 'bring your own key'. In terms of the security capabilities that Power BI has introduced, they really have been massive over the last 18 months. And it has really helped customers, especially when COVID has hit. That's why I would say security and trust are number one for us.

Number two is making sure that Power BI is just deeply integrated into your collaborative experiences. We really do see Teams as being the home for Power BI, where you can find your analytics, you can discuss your analytics, you can discuss your data, all of those things, just living together. Every team that's formed has a goal that they want to accomplish. Often the progress of that goal is tracked in Power BI. So, it's just such a natural fit.

The third area is really AI, you know, with the amount of data that's available, with the frequency at which it's changing and the volume, you cannot have humans figure this out anymore. You have to have machines help humans. So, AI is where we're making a lot of progress because we're able to do things like, "Hey, automatically find the influencers that drove an outcome; automatically decompose your KPIs into the dimensions that drove the results; automatically generate a natural language summary; automatically identify anomalies." All of these things, our vision for Power BI there is that, "Hey, we should be able to tell you what matters to you. You shouldn't have to go hunting around to discover it." That's the third area we are investing in.

Power BI has, no doubt, evolved into one of the premier business intelligence tools globally over the past five years. The focus on community-requested features has also been quite apparent in recent releases. Arun Ulag detailing how the platform currently functions, and how he sees it developing in the future under his role as its CVP certainly provided us with some unique insights. As to how the latest assortment of introductions is perceived by the business intelligence community moving ahead, only time will tell.

The transcription of the interview has been slightly edited for brevity and relevance.

Report a problem with article
Next Article

Leak suggests Google Pixel 5 could be priced at $699

Previous Article

Samsung Pay launching in Germany, works with almost every bank account

Join the conversation!

Login or Sign Up to read and post a comment.

2 Comments - Add comment