Built-in AI in Power BI. Find the Insight You Can’t See!

QUICK LINKS:

01:41 — Microsoft Teams integration

04:10 — New Goals functionality

05:41 — Smart Narratives

07:15 — Other smart visualizations with AI

08:48 — Set up Synapse Link for Dataverse

09:24 — Hybrid Tables

11:12 — Wrap up

Link References:

Learn more at powerbi.microsoft.com See additional capabilities on our blog at powerbi.microsoft.com/blog

Unfamiliar with Microsoft Mechanics?

We are Microsoft’s official video series for IT. You can watch and share valuable content and demos of current and upcoming tech from the people who build it at Microsoft.

• Subscribe to our YouTube: https://www.youtube.com/c/MicrosoftMechanicsSeries?sub_confirmation=1
• Join us on the Microsoft Tech Community: https://techcommunity.microsoft.com/t5/microsoft-mechanics-blog/bg-p/MicrosoftMechanicsBlog
• Watch or listen via podcast here: https://microsoftmechanics.libsyn.com/website

Keep getting this insider knowledge, join us on social:

• Follow us on Twitter: https://twitter.com/MSFTMechanics
• Follow us on LinkedIn: https://www.linkedin.com/company/microsoft-mechanics/

- Up next, we’ll take a tour of the latest updates that integrate Power BI inherently with Microsoft’s collaboration analytics stack, including the new Power BI app for Microsoft Teams that makes it easier to access, share, and act on insights as you work, the latest AI-driven visualization capabilities with smart narratives, and achieving near real-time analytics over your enterprise data using hybrid tables. And to walk us through all of this, I’m joined today by Vanessa Fournier, from the Power BI team. Welcome back to the show.

- Thank you. It’s so nice to be back in person.

- It’s so great to have you back on. So it’s been a while, actually, since we’ve done a deep dive on Power BI, and it’s really great to see that it’s continued to evolve really as a canvas for insights on all your data, regardless of its size or scale.

- You know, the whole design principle behind Power BI is to make it accessible for everyone to use, from your novice to your data scientist. We want to build a business intelligence offering that supports everyone with gleaning insights that drive their personal decisions or their business. We have literally hundreds of interactive visualizations at your disposal to build reports and dashboards today, and we’re continually working to make the Power BI experience better, so it’s the single canvas for business insights, powerful visualizations, and AI-driven analytics. And together with Azure Synapse Analytics, it works across all of your data sources, wherever they sit and no matter the size. Also, within the new Microsoft Teams Power BI app, we’re helping you quickly discover and cascade actionable insights, collaborate with others, and make faster, more personal, informed decisions the way that you want to work.

- And of course, Power BI is part of the Power Platform of low-code tooling, making it easy to integrate insights into your Power Apps and Power Automate business processes. That said, though, let’s take a look at some of the new capabilities, and why don’t we start with the Microsoft Teams integration?

- Sounds good. Let me show you Power BI with the new Teams app. All right, here, you can get to the app from Teams’ app store. Or in my case, our Teams admin has actually pinned it centrally for everyone on my team, so it’s easy to find on the left rail. I can see recommended reports where we use intelligence from the Microsoft Graph to surface what’s trending, along with everything I’ve recently accessed. I’m interested in the Sales Pipeline report, so I’ll open it. And if I go into the Insights tab, I can see some AI-generated insights that show key influences that affect the pipeline. Here, I can see that Anne Weiler is taking the longest to close her deals, 38 days longer than average, or around 144 days total. So it looks like I need to give this some attention, and I can easily call in the expertise of my team to investigate. So right from the report, I can share this insight, along with what could impact our forecast. In fact, I’m going to share it via a Teams chat with you, Jeremy. You’re a member, so maybe you could help us investigate.

- Okay, sounds good. In fact, I can see your notification pop up right here. So I’m going to jump into the chat and take a look at the link that you sent. And it takes me directly to that Power BI report, and from here, we can start to work together to figure this out; and I can see what you mean by Anne Weiler’s closing times being a lot higher than average, so let’s take a look at what that root cause might be. So I’m going to use a decomposition tree to look at some of the differences between the sellers who are represented here by the owner value. Now, first you can see our top seller here, Riley Johnson. So I’m going to select the fields I want. First, the product category. Then I’ll select the industry field, just to see the types of customers they’re working with. And so Riley, you know, we can see that with her, accessory sales are roughly equal to the sum of devices and warranties. Now, computer software, business services, and healthcare appear also to be the highest customer industries that she’s working with. With Anne Weiler, though, who’s pulling our averages down a bit, you know, with the longest time to close, I can see that their mix of accessories is a lot lower than Riley’s. In fact, here, their industry mix also is highest for medical, restaurants, and industrial machinery, and we happen to know that in our case, these industries take a lot longer to close deals. So now we can maybe work with Anne to really focus on accessories or target different industries, so we can start to bring her closing times down.

- Right, and this is a great example of how you can collaborate with your peers in a data-driven way to highlight and resolve issues within the context of Microsoft Teams, by sharing real-time insights. Next, I can also drive more progress against the issue by using the new Goals functionality in Power BI. This allows me to curate core business metrics that I want to track in a scorecard. I have one here for sales goals that I’ve added to my weekly team meeting, and based on the previous example, I’ll set up a new goal to ensure that our seller’s deals are closed within five days of our target of 100 days. I’ll edit this and I can create a new goal and give it a name of Close Rate. Now I’ll select the data I want to track, and in this case, it’s Sales Pipeline, and there’s the report and pages we saw earlier. Now I’ll head back into Insights and choose the data point I want, the average days to close. Then I’ll connect that to our goal. Here I can see it populates the current value in our report of 104, and in target, I’ll set it manually to 100 days. Over here for status, I’ll set up a new rule by clicking. And in my case, I’m looking for a value change that is greater than 5%. Then, if that condition is met, I’ll change the status to at risk. Otherwise, if it’s less than 5% from our target, I can definitely say the status is on track. From there, I can save it and confirm for the all up set of goals. Since I just added this close rate goal, I’ll go ahead and add a note with a new check-in for today to say that we’re on target. And this view gives a great tool to the team to track progress against the goals that are important to us.

- Right, I can see this driving a lot better collaboration around specific insights and data, but why don’t we step back and take a look at the Power BI report itself? So what are some of the recent updates there?

- You know, Jeremy, we’re constantly adding visualization and AI power-derived insights. In fact, one that I love is called Smart Narratives. This allows you to add verbal narrative to the reports to land key takeaways and point out trends. I’m going to open our win/loss report in edit mode. I can add a smart narrative which uses built-in AI to summarize the whole report for me, to describe the top takeaways. Already, you can see that it’s found the month of December was almost four times better than July, along with more insights. I can also customize the narrative with my own values, text, and formatting. Here, I’ll add the title ‘Smart Narrative.’ I want to create a dynamic value, so I just start typing what I want, like revenue + revenue in pipeline. I can even add some text around this. I’ll also add a second dynamic data point that calculates its percentage goal attainment. Now I can add some formatting. So here, I’ll just bold the key values. You can also create a smart narrative around a single visualization, for more tightly-scoped insights. To do that, I’ll right click on the visual I want and select summarize. Now I have a smart narrative for the close rates of the last eight months, and I just need to place the summary where I want it on the page. What’s also great is that when I interact with this report, like selecting the category devices, I can see that both narratives automatically change based on the filter I select.

- And this is a really amazing technology to summarize, you know, all the report insights into words, and I can really see this being used, for example, with cut and paste to put it in the email or Teams chat when you’re sharing report insights with other people. So do you have other smart visualizations that have AI built in?

- Yes, we are invested in bringing out-of-the-box, easy-to-use AI to all of our users, like key influencers, text and image recognition, and the decomposition trees we saw earlier, just to name a few. If you’ve ever wondered why there is a spike or dip in a graph or chart, Analyze will explain the increase or decrease without you having to dive back in to the data and explore. Here it found that the states of New Mexico and Idaho had significant increases that led to this particular spike. It’s also found the variances by industry. And this is data you can’t see from the report. AI has discovered the root causes in the underlying data. There’s also AI built into the line charts that you can automatically detect anomalies in your time series data, providing explanations for the anomalies, again, to help you speed up root cause analysis. Here, it’s found the correlation between daily sales discount rates and overlaid those against revenue per day to explain the increase. Again, these are things that could be super difficult to find without this anomaly detection.

- And it’d be a massive time saver, I think, especially compared to manually sifting through and digging through that data. So now we’ve actually seen how Power BI can unify data and the stories around it, but can we take a look at what’s behind your report?

- Absolutely. Behind the scenes, we’re querying billions of rows of data, from Dynamics 365 to Azure data services, as well. And we’ve used Azure Synapse to bring all of the data together. You can connect to data from anywhere, across clouds or on-prem, and with Synapse Link for Dataverse, you can now easily bring in your operational data.

- And by the way, setting this up is pretty easy. In fact, I’ll show you right now. So starting in the Power Apps Studio under Dataverse and Azure Synapse Link, I’ll add a new link to a data lake. I’m going to choose to connect to my workspace, and then I’ll fill in kind of the standard Azure fields that you would expect, along with my Synapse workspace. Now, here, I’m going to choose the tables that I want to link. I’m just going to pick a few here. I’ll pick account, then I’ll search for user, then select it, and now I can go to my Synapse workspace. And there on the left, you can already see the two tables there, right from my Dataverse environment. So with all this operational data now flowing in, we’ve got our billions of rows of data. How do we make sure that our query performance is fast enough?

- The key to this is Power BI Gen 2, and it now supports large scale analytics with fast performance using a new capability called Hybrid Tables. These let you query massive dynamically growing datasets using Power BI without any lag. And the best way to explain this is by showing how to set it up. Here is a report with our website traffic. It has an underlying dataset of hundreds of millions of rows, and is growing quickly. From the data model view, I’ll show you the DemoClicks Hybrid Table. As you can see under incremental refresh, the single table has been set up with multiple data segments. The diagram at the bottom is a great visual representation for how data is flowing, stored, and refreshed. We’ve set the first segment in the table to archive data for the past one year. The second is to set incremental refresh data just for the past one day, and with get the latest data in real time with direct query selected, there is also real-time data coming in. I’ll close this view and then publish the report. With the Hybrid Table set up, it means that even though our table contains a year’s worth of data, it’s like querying one day’s worth of incremental data.

- So with all that data flowing in, then, how do we take advantage of this then, with our report itself?

- To do that, I need to change it to automatically refresh every few seconds. So I’ll go to the page Refresh Control and turn it on. First I’ll change the interval to seconds, then make it five. Now I just need to save it and open the report’s reading view, and you can immediately see that the total visits and unique visits are growing every few seconds. Below that, you can see the raw data rows. I’m going to change the sorting of the datastamp column so the newest rows are on top. And now you can see the most recent site visits refreshing every few seconds as well, and even though the underlying data is changing and there are hundreds of millions of rows, Power BI returns the results in near real time.

- And this is really powerful stuff. I mean, we’ve just seen now the fusion between Power BI, the collaboration tools, AI, and also our analytics stack.

- It is, and this is really just a taste of everything that’s to come with Power BI. In fact, a lot of recently added capabilities that we didn’t have time to cover today can be found in our blog.

- And I know a lot of people probably watching are really looking forward to getting their hands on this and trying it out.

- Well, the good news is that you can try everything that we showed today. The Power BI app in Teams is available now, Goals and Hybrid Tables are in preview with Power BI Premium, and you can always learn more at powerbi.com.

- Thanks so much, Vanessa, for joining us today and sharing all the updates with Power BI. And of course, keep checking back to Microsoft Mechanics for the latest tech updates, and don’t forget to subscribe to our channel if you haven’t already, and thank you for watching.