Microsoft Intelligent Data Platform | Real-world demo

Work with near real-time insights, augmented with predictive AI and built-in controls that protect your data wherever it goes. See how Microsoft’s Intelligent Data Platform integrates databases, analytics, and governance into a single, streamlined solution. Seize the data!

We’re all trying to work faster and smarter to solve today’s data-driven challenges. Microsoft’s Intelligent Data Platform can help by removing silos around data and providing near real-time insights, augmented with predictive AI and built-in controls that protect your data wherever it goes.

Sharieff Mansour, a leader on the Data and AI team at Microsoft, joins Jeremy Chapman for a closer look at how databases, analytics, machine learning, and governance come together in the Intelligent Data Platform. See how you can use these integrated services to build powerful new apps or modernize your current ones.

Remove silos between data professionals, developers and end users.

Microsoft’s Intelligent Data platform helps you work together efficiently on common data and insights, using familiar tools without a need to learn new skills. See how to remove tech and people silos to make data work for you.

Understand and discover your data at nearly limitless scale.

Intelligent Data Platform brings different types of operational and analytical data together, wherever it resides, in real time. See how to improve the way you work with data.

Integrate database analytics and governance into a single solution.

Microsoft’s Intelligent Data Platform combines individual solutions with our technology stack. See how Intelligent Data Platform solves real-world problems.

Watch our video here.

► QUICK LINKS:

00:00 Introduction of Microsoft Intelligent Data Platform

01:05 Improve the way you work with data

01:37 Foundational advantages of Intelligent Data Platform

02:09 Demo of how Intelligent Data Platform solves a real-world problem

05:30 Tech behind the app

07:37 Built-in AI predictions

10:32 Remove tech and people silos to make data work for you

12:04 Personalize customer interaction with Power Virtual Agent BOT

► Link Reference: Find the latest information on Intelligent Data Platform at: aka.ms/MicrosoftIDP

► Unfamiliar with Microsoft Mechanics?

• As the 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/MicrosoftMe...

• Talk with other IT Pros, join us on the Microsoft Tech Community: https://techcommunity.microsoft.com/t...

• Watch or listen from anywhere, subscribe to our podcast: https://microsoftmechanics.libsyn.com...

• To get the newest tech for IT in your inbox, subscribe to our newsletter: https://www.getrevue.co/profile/msftm...

► Keep getting this insider knowledge, join us on social:

• Follow us on Twitter: https://twitter.com/MSFTMechanics

• Share knowledge on LinkedIn: https://www.linkedin.com/company/micr...

• Enjoy us on Instagram: https://www.instagram.com/microsoftme...

• Loosen up with us on TikTok: https://www.tiktok.com/@msftmechanics

Video Transcript:

- Coming up, we take a closer look at Microsoft’s Intelligent Data Platform that integrates databases, analytics and governance into a single solution. We’re going to show you how this can solve challenges like today’s supply chain issues by removing the silos around data to bring near real-time insights augmented with predictive AI and also built-in controls that protect your data, wherever it goes. I’ll also show you how you can easily scale and build out new intelligent experiences around your data to modernize your customer-facing apps. And to walk us through everything, I’m joined today by Sharieff Mansour from the Data and AI team at Microsoft. Welcome.

- Thanks, happy to be here to help unpack things further.

- Thanks so much Sharieff. We’ve been working with you and your team for a while now and keeping track of all the updates across all of our database services, analytics and governance solutions. And they are are great options on their own, but together they form the Microsoft Intelligent Data Platform.

- That’s right they do. And it’s really by design. We’ve been working on providing best in class individual solutions as well as providing deep integration across our technology stack for some time now. So, let me explain how a single intelligent data platform can improve the ways you work with data. Providing a deep level of integration not only brings different types of operational and analytical data together, wherever it resides and in near real-time, it also allows for greater understanding and discoverability of your data at pretty much limitless scale. And silos are also removed at the people level. You know, your data professionals, developers and end users can more easily work together on common data and insights and they can use the tools they’re already familiar with without having to learn new skills.

- And from an IT perspective, not only are we breaking down the silos and helping people work better together, there are also foundational advantages here as well. So underpinning all of this is the shared identity and access management through Azure Active Directory that allows your users and your apps to safely access your data. And we have end-to-end security through Cloud Defender and also Microsoft Sentinel, as well as unified policy controls that extend to all of your data whether that’s in Azure, on premises, hybrid or across cloud. So, why don’t we make this real though for everyone with a tangible example.

- So Jeremy, why don’t we do this together with a scenario that’s top of mind for many of us today, supply and demand. And this is more challenging than ever, because it’s not just about operating at scale and meeting demand, there’s also tension on the supply side as supply chains are disrupted. Ultimately, this is a data problem. In this case, we’re a manufacturer of solar charge controllers. The device that distributes solar energy between the battery units and the power grid. And there’s lots of variability. Starting with the demand side, this is impacted by different government incentives starting and stopping all the time and people considering their options based on events like brownouts. On the supply side, we rely on just-in-time supply so that we don’t overstock, but we often struggle with things like component shortages for manufacturing as well as shipping delays and unexpected facility closures. So this is the app our sellers currently use to look up units available for sale. Now, so that we don’t impact our operations, we do daily batch runs, which means we never have an up-to-date perspective of inventory levels. In fact, as you could see here, the last batch run happened more than 19 hours ago. And as a rule of thumb, we never sell more than half of our current stock levels so that we don’t promise units that have been already accounted for from sales orders that are not yet represented in this view. I think this is something that we’ve probably all been feeling with the recent pressures on both supply and demand.

- That’s right, it’s a common pain point today. So let’s see what happens when we take the same app and apply the intelligent data platform to it. Here’s our new app.

- [Jeremy] Such an improvement.

- Such an improvement. And instead of looking 19 hours in the past for my last batch run, as you could see, inventory is updated every two seconds. So I know immediately what I have available to sell. On the left, I can see the number of units in stock and we know this number doesn’t tell the entire story. To solve that, on the right, I can see what’s available for sale with a data-driven calculation of what will be available a week from today. It’s a dynamic and current perspective. Let’s dig into the details. A lot of factors impact what’s really available to sell. Behind the stock level, I can see what’s already accounted for, the cancellation rate, the manufacturing run rate, shipping delays, returns, parts required for servicing and I have a predictive view of demand surge, which I’ll explain in a moment. The point is, before, using my rule of thumb, I would only commit to selling around half of our current stock level, which is 430 units. And I’d be leaving money on the table, because I actually have around 1200 units available. And if I flip over to Power BI, I can get a macro level view of supply and demand with AI-driven insights. I can see a spike in demand in shipments in places like South Africa, but I know there are a lot of government incentives there so that’s not surprising. California on the other hand, stands out as a big anomaly. And AI gives me a possible explanation by linking the blip to widespread power outages. Now let’s look at the supply side. I can see status of shipments en route to our factories. The map on the right shows me where supplier factories are offline and below that, I can see the areas where inbound supply is stuck. For example, in LA and San Francisco, where the ports are gridlocked. And a lot of our inbound supply goes through those routes. This level of insight is night and day compared to what we had before. So Jeremy, why don’t we unpack what’s going on with the tech under the covers?

- Sounds good, let’s have a look at what makes this app tick. So behind the app was SQL Server 2019 and we’ve since upgraded that to SQL server 2022 and this helps extend our on-premises database to the cloud, but keep the data where it is now. And we’ve also connected it to Azure Synapse using Synapse Link for SQL Server. Now, this has opened up more data sources across all the various dimensions for supply and demand. So let’s see how this comes together. Just to show you how integrated everything is, I’m going to start in Synapse Analytics. And using integration with Microsoft Purview, right here from Synapse started a search for predictive inventory. Now it returns results across my entire data estate. I’ll select this table for predictive inventory, which is our join table. And through automated data classifications right here in the schema tab, sensitive information is already flagged. Now, this lets us protect the data using classifications to ensure that neither our apps nor our reports will surface sensitive data. Now, the data lineage tab provides us great left to right view of the Azure Synapse table showing that it originated first in Cosmos DB and a couple of SQL databases on the left. It’s being all the way used through our Power BI report. And not only are these databases discoverable in Microsoft Purview, they’re also fully integrated in Azure Synapse with all the tools necessary that you need to work with the data. And to be clear, not only does it unify data access, but it also brings together the various people and their skillsets to work with it. Now, our data engineers they’ve brought in all the data and we can now pivot that. And we need to make sure here that our inventory on hand is always current not those 19 hour legs like you showed. So our Synapse Link connection makes it possible to query operational inventory data in our SQL database and it’s also instantaneous. Now our pipelines, they’re also bringing in IoT telemetry data from Cosmos DB. Now, this is globally distributed and coming in from thousands of devices simultaneously, which are installed around the world at customer sites. And this data is also aggregated so that we can calculate and allocate the correct amount of inventory for servicing, for things like parts in the field. So what we’ve shown is how all the data comes together in a single view across our operational and analytical sources as well as the various people involved. And as you saw, the new app also has AI predictions built into it. So let’s unpack what’s behind those.

- Sure, so those also need to be calculated in real time and this is where we can help multiple teams work together even your data scientists. So, right now, I’m in the Azure Machine Learning experience. My data science teams can apply predictive machine learning models using the same data. And there are models here to predict order cancellation rates as you can see at the top and then shipment arrival times as well as units to allocate for parts and servicing. And the one here at the bottom, it’s looking at localized demand spikes. Now these are all regional phenomena that indicate a surge in demand. So I’m going to click into this one here to access our notebook. And here, we’re looking for things like outliers and demand by region, and we’re also doing anomaly detection. So in certain cases, this is going to find things like those power brownouts that you mentioned before in California. And if I scroll down a bit further, there we go. I can also see predicted demand and I can see weekly predictions, which will inform how we plan our manufacturing for all of our controllers.

- And by the way, these models are how we are able to forecast up-to-the-minute numbers and predictions in Synapse that show up in our new app.

- Right, so now all of that machine learning data needs to be merged with all the operational data to be able to inform the predictions that were surfaced in your app. So right here in Azure Synapse, we can access and leverage the models that we just saw to score the combined data table that I showed earlier from the operational side and expose all those predictions and adjust the totals as needed. So these here are the mappings actually from my source data along with my prediction data here in this middle column that’s called Model Input. And now, I can upload the model to my table, which will surface up all the predictions that we saw. And this SQL script here will score the data behind our updated app. And this data is also behind our Power BI dashboards so that our business analysts can also plan out ordering and manufacturing. And if I log into SQL Server Management Studio, because what data demo would be complete without that, I can see that all of our Synapse connected data is here and it’s written back to our SQL Server data store that’s on premises. Now, this allows our operation teams to query it directly as you can see as I make this select statement here.

- And just to put this into context, what we just saw is the reduced churn in connecting operational and analytical resources as well as a layer of governance and understanding of the data. Something that would take months is now achievable in just days.

- Right, and it’s also kind of blurring the lines between all the different services. You’ve got Synapse, you’ve got Purview, you’ve got all the different data sources and also Machine Learning. And it’s also blurring the lines in terms of how various teams can work together.

- That’s right. And that’s really the point here. We’re breaking down both tech and people silos across various processes to intelligently make your data work for you, but let’s even go further. Let’s fast forward in time. We want to expand from being a B2B wholesaler to sell direct to consumer. So here, you’re seeing our new public website that allows almost anyone around the globe to buy direct. As you could see, I’m visiting the site in the US. Everything here is powered by the same near real-time, predictive, intelligent data platform backend we built for our internal app, but now it’s public. We’ve also added recommendations for the highest ranked models in our lineup using availability, compatibility and local service provider data. And as you can see, this is all location specific.

- Right, and under the covers it’s actually looking at a few more things. It’s looking at the units that are out in the field, the servicing records, and also customer feedback to really recommend the best possible unit. So why don’t we take a look at what makes all of this possible? So, behind the customer facing web app is a massive Cosmos DB. Since everything is already live and available globally, here we can see just the scale of the request, the throughput, the data and also index size. Then as you can see from the map, some massive dataset flowing in from around the world in real time. And it’s leveraging the supply and demand insights that we built along with the data from our order and fulfillment processes. Now, ultimately we want to also light up new experiences for our customers. In fact, in addition to our usual text-based chat bot, we’ve also added voice interaction for personalized real-time feedback on things like shipment status and product recommendations using Power Virtual Agents, which is integrated with the Intelligent Data Platform. Let’s try it out by calling the support number.

- [Virtual Agent ] Thanks for calling Rela Cloud Customer Support. In just a few words, tell me how we can help you today?

- I’m calling about my order status.

- [Virtual Agent] Sure, I can help you check the status of your order. Do you know your order number?

- No, I don’t have it.

- [Virtual Agent] No problem. Let’s find your order another way. What’s the name of the company on the order?

- Contoso.

- [Virtual Agent] Thanks. Now, can you give me the last name of the primary contact on the order?

- Vance.

- [Virtual Agent] Great. Finally, can you tell me the date the order was placed?

- Last Wednesday.

- [Virtual Agent] Good news. Your order was dispatched on July 3rd and is currently estimated to arrive on July 8th. Is there anything else I can help you with today?

- No, thanks. Notice the intelligence baked in there beyond voice recognition itself. The bot was able to verify my request and look up the information I needed even though I didn’t have my order number.

- And all this is made possible through conversational AI, which by the way we covered on a recent show that you can check out at aka.ms/PVAMechanics. I’m going to show you how this works as well. So I’m here in the Power Virtual Agents authoring canvas for the topic that gets triggered when a customer asks about order status. So in this action here, you can see that we’re calling Power Automate, which seamlessly connects to Cosmos DB using the pre-built connector. And there are lots of connectors available to connect to the Intelligent Data Platform as well as common third party data. Now, in this condition, you can see how the bot can even escalate to a live agent if there’s a problem finding the order. The voice integration by the way, is using Azure Cognitive Services and Azure Communication Services are used for voice and telephony.

- And what we went through today is just the tip of the iceberg, because your teams don’t have to integrate all these technologies. You can use your existing skills. Governance is baked in throughout. You can respond faster, work together more easily, unlock new insights, and modernize the way you work with your data.

- It’s really great seeing the possibilities that the Microsoft Intelligent Data Platform brings. So, for anybody who’s watching right now looking to get started or learn more, what do you recommend?

- The cool thing is everything we’ve demonstrated today is available right now. And to learn more, you can go to aka.ms/MicrosoftIDP. Thanks so much Sharieff. And of course, keep checking back to Microsoft Mechanics for all the latest updates and be sure to subscribe if you haven’t already, and as always, thank you for watching.