Integrate 2018

June 4-6, London


Kent Weare compares Microsoft Flow to Logic Apps, then explains how it fits in the Business Application Platform. He showed 4 demos that displayed the versatility of Microsoft Flow, and finished with a roadmap for the product.

Microsoft Flow in the Enterprise

Integrate 2018, June 4-6, etc.venues, London


Video Transcript

Good evening, everyone. My name is Daniel. I’m a Senior Software Engineer for Atomic Scope. So, I understand everyone is really tired and myself and our next speaker standing away of the drinks that are waiting for you behind the doors. But I would like to kindly ask you for a full attention to our next speaker, Kent Weare. Who’s going to talk to you about Microsoft Flow in the enterprise. Please welcome, Kent.

Kent: Thanks, Richard. We might crowdsource your intro in a couple of days. So, thanks everyone for staying. I will be as efficient as possible with our time as I do know that I’m standing between you and drinks. So, my name is Kent Weare. I’m a Principal Program Manager on the Microsoft Flow team. For those that don’t know me, I’ve previously been in the MVP community working on a lot of enterprise integration using BizTalk as your Logic Apps, API management, and Azure messaging. So, one of the questions that I do get quite often is, “What are some of the use cases for Microsoft Flow in the enterprise?” And I really wanted in this session is to walk you through some of the scenarios and how we actually position Microsoft Flow within the enterprise. So, we’re going to keep the agenda really simple here.

So, positioning Microsoft Flow, it is a demo-heavy session. I’ve got four demos for you and I think I should be able to open up some eyes in terms of how Microsoft Flow is used. And then lastly, we’re going to talk a little bit about the Microsoft Flow roadmap. So, first question, show of hands, who has heard of the Business Application Platform? Not many. Okay. This is a good opportunity. John has. That’s awesome. So, the Business Application Platform is essentially a platform for power users. And from a go-to-market perspective, there’s two brands that we’re heavily leaning on. Number one being Power Apps and that’s a cross-platform mobile application platform, which you saw earlier this morning. And then certainly, Power BI. Basically, an analytics and data exploration tool which once again is targeted towards power users.

Now, hanging off of the sides of these are data stores. You need to be able to store your data somewhere. And so there’s two data stores that we’ve recently introduced under the term Common Data Service. And we’ve got the Common Data Service for apps and we also have Common Data Service for analytics. And then we also have Microsoft Flow, which is what I’m going to be talking to you today about. Underpinning this platform is the common data model. Now, naturally, when you have a data store, you wanna be able to model your entities. And the idea with the Common Data Service for apps is that you can go ahead and design your entities or your tables and create relationships across those different entities.

And then naturally, you need the ability to get data in and out of this platform and you’re going to do so using connectors. And when we talk about connectors, these are the same connectors that Power Apps has access to, same connectors that Logic Apps has access to and naturally, the same connectors that Flow has access to. Now, we are targeting building applications on top of Dynamics 365, Office 365, but you can certainly build application standalone as well. And underpinning all of this is Azure. So, this is what gives you the ability for performance and scale and reliability in many regions across the world.

Here’s another view of the business application platform. So, we’ve got the Common Data Service underpinning and then we’ve got Flow. And Flow is a Cloud workflow and automation platform. But we also have something called business processes. So, if you’re familiar with Dynamics 365 and what’s called BPFs, business process flows, what you’re going to see is those are being wound into the Microsoft Flow brand. So, soon you will be able to access these BPFs from the Flow portal. And eventually, there will be a re-platforming exercise around that. So, what’s interesting about this point is that now these BPFs are being externalized from Dynamics CRM and actually you can now use them inside of tools like Power Apps or within SharePoint and actually have a state machine where you can move forward and backward in a business process.

So, strategy. How is Microsoft thinking about the business application platform? So, we firmly believe that the fastest way to digital transformation is through the use of SaaS applications. Now, one of the challenges is if you wanted to bring a talent application into your organization, nowadays, you probably aren’t going to custom build that. You’re likely going to look for something off the shelf. But the problem with that is that you’re never going to get the full way. Chances are you’ll get 80% of the way and then you need to fill in gaps. These gaps could be related to the industry that you belong to or could be part of the jurisdiction that you belong to, whether it be regulatory or compliance. And so what we wanna do is we wanna build a business application platform that allows to develop very quickly and democratize the people that have access to those tools. And that’s really where we’re focused on.

And then lastly, there’s going to be those situations where you want to build custom applications. And there’s no problem with that. Microsoft and Azure has a full library of different tools that you can use for that purpose. But from a Microsoft Flow perspective, we’re really targeting this middle vertical around enhancing and complementing existing systems.

So, in summary, from a business application platform perspective, we are Cloud first. We run in the Cloud much like Logic Apps. In fact, on top of Logic Apps, we do offer a single data store. This idea now of CDS means that you can actually bring all of your data together and actually manage it in a single location. And using the scale of Azure, you know that you can actually scale and support larger workloads. We also have connectivity. We support over 219 different connectors within our platform that allows you to connect to many different systems, whether they be first-party Microsoft, third-party custom developed or on premises. We are cross-platform. We’ve seen a few Macs on stage today. Certainly, they’re welcome at the party. You can go ahead and build these applications on a Mac, you can build them on Android and you certainly can build them on Windows. We do offer a central management experience, which we’ll talk about in a few slides.

And lastly, we do support this notion of no cliffs. We fully recognize that as a power user, you’re probably going to get so far with this technology set and then you may want to hand it over to people with more expertise. So, we support this grow-up strategy where you can actually start with a Microsoft Flow, run it within the Flow portal, export it, and then import it into the Azure portal as a Logic App. And the reason we’re able to do that is because we are using the same workflow definition language that Logic Apps as well.

So, let’s talk a little bit about customers and who are actually using this platform. So, today we have over 1.2 million active users in the business application platform. Certainly, Flow represents more than half of those users. And here’s some of the companies that have bet on this technology. We also have a number of other companies that I can’t show their logo, but you use some of their services every day, whether it be social media, whether it be entertainment and sports or whether it be actual amusement parks. These are all very innovative, large companies that have bet on our platform. And the goal and the reason why companies have bet on this platform is they want to democratize app creation. And the reason for that is that transformation does not occur while you’re standing in line.

When you have business units that have pent-up demand, as John mentioned this morning, if they get the answer no, they’re going somewhere else. So, something that you can use as an IT organization or a consultancy is actually using these democratized tools in order to help an organization scale. And an organization can scale by bringing more people to the party and giving them a democratized set of tools in order to build out some of these applications. And this is something that is not new for Microsoft. Certainly, if you look back at Excel and Access and Infopath and even VBA, these are things that have been happening for years, if not decades.

Now, suspend disbelief around using some of these tools if you are in IT, because isn’t this like giving scissors out to kids and asking them not to run? And the answer is it doesn’t have to be. And the reason for that is we do have a centralized management experience. So, one of the core issues or core challenges or concerns that IT organizations have is that business users go off, they build these Franken baby spreadsheets or Access databases, and then they actually become dependent upon them. Then something gets corrupted and then who gets the call? It’s like some leader inside of IT and they’re like, “How did you let this happen?” When that leader may not have even known that this was happening.

And so the difference between that paradigm and the paradigm that we have today is the notion of the Cloud. So, using the Cloud, we can federate identity. So, using Azure Active Directory, we know who’s logging into the systems, we know who has licenses for these different technologies. We have a Unified Admin Center. So, this idea that an administrator can go ahead and see what assets have actually been built up within an organization and actually manage those is something that we exist or that we provide. We provide something called the Data Loss Prevention policy or DLP policies which allows you to isolate different connectors so that you can’t have, say your SharePoint data re-going up to Twitter, which is another concern that we regularly hear from enterprises. So, this was one strategy that will prevent data loss or data leakage from within your organization.

From a security and compliance center, we do integrate with the Office 365 security and compliance center where we will track different events that have occurred when someone’s creating a flow, editing a flow, things of that nature. And lastly, we do support and are investing in mobile device management. So, we support conditional access today and we are currently working on MAAM or mobile application management which will be coming soon. So, with that said, let’s get into some demos and show you what Flow is all about.

So, Flow is the automation tool for Office 365. We are the successor for SharePoint Workflow Designer. Now, granted, we do have some symmetry challenges to work through, but one of our target use cases or sweet spots is really in the area of Office 365. So, as you can see, we’ve got existing support for connectors with Office 365, nine in total. But what we’ve also been investing in recently is first-party integrations with these different platforms. So, platforms like SharePoint, OneDrive for Business, Microsoft teams, Excel and Outlook. And one thing worth noting is that if you have entitlement to Office 365 or Dynamics 365, you have entitlement to Flow. In most cases, that will be 2,000 workflow executions per month per user, which gets aggregated at the tenant. So, this really creates an opportunity for people to discover and try out Microsoft Flow without worrying about large financial commitments.

Also, with Microsoft Teams, we’ve had an existing connector or an existing operation that allows you to post messages to any Team’s channel. That’s not new. But what is new is that within Teams we have a flow app. So, from within the Team’s marketplace or store, you can go ahead and install the Flow app. And what this Flow app allows you to do is access personal flows. If you want, you can go ahead and edit in that experience by creating new flows or editing existing flows. And you can also access your approvals through our approval center. In addition, we’ve also released the first iteration of our flow bot. So, this provides you the ability to call any flow that is scheduled from within the Team’s interface.

And so one of the scenarios that you can use this within is a change in incident management. So, certainly, people in this room, I’m sure there’s a regular change board that exists at your organization and it’s probably a weekly meeting where some people show up, part of the organization shows up and actually goes ahead and reviews changes that are about to take place for this week. Now, for many people, they need to know whether or not their change has been approved before they can go ahead and do that change. Otherwise, they get their hand slapped.

Now, you can usually log into a system, your IT service management system to see if you’re a change has been approved, but that’s a bit of a hassle, especially if it’s you’re dealing with many changes within your teams. So, what if you had the ability to pull this information from Teams itself? And you could go ahead and do that through Microsoft Flow and through our Bot and the connector that we provide with Microsoft Teams. So, that’s what I’m going to show you now. Okay. So, here I’m in the web version of Microsoft teams. I can go into a chat session. And here I have the actual Flow app that has been installed. I’ve got my conversations, which I can start a conversation with the Bot, I have my flows, approvals, and in this case, what I can go ahead and do is say I wanna a list flows. List flows that are scheduled that I can actually go ahead and call from the Bot.

Now, in this case, the Flow has been kicked off and if I head over to my channel, I should be able to see that…whoops. I jumped ahead.




So, now I’ve listed all of the flows that have the ability to access and then I can go ahead and run a flow based upon the index. So, in this case, I’m going to go ahead and run flow three, which is “Show me a list of approved change requests for this past week.” If I jump over to my team and I head down to, in this case, I’m interested in the network team for Contoso. I can actually see the changes that have been approved for this upcoming week. And I can see basically the window, the change window in which I need to execute this and who the changes are actually assigned to. I can do the same thing for incidents, but in the interest of time, I’m not going to bother. It’s a similar pattern. And how this actually is built inside of Flow is I go ahead and create a recurrence trigger. And in this case, it’s really a schedule, but the idea is that I don’t need to wait for the schedule in order to run it. I can just go ahead and call it from Teams.

And I’m going to go ahead and list the assignment group. So, I’m only interested in this context in changes for the network team. And then what I can go ahead and list all of the approved change requests for the network team. So, I need to make a few different calls because we’ve got different system references across these different entities. But I can also pass in queries which is a query language specific to ServiceNow in order to find any of these changes that are approved that belong to this specific assignment group and where the state basically is scheduled. Once I’ve gone ahead and retrieve those results, I’ll create an HTML table and then I’m just going to simply post that into the Team’s channel.

So, a very simple process that actually can provide a lot of benefit for your users without needing a ton of expertise and the ability to quickly build something. So, the next thing I wanna talk about is Flow integration in Excel. So, this is something that we announced at Build last month. And what we have is basically an add-in for Excel. So, today, when this does ship, there’ll be an add-in where you’ll actually have to go and opt-in for this widget to appear, but in the next version of Office, this will ship natively inside of Excel itself. And what we can do with this is we can actually select a row and then basically launch a flow as a result of that selected row. And what’s interesting about this is it runs in the context of the person that has actually clicked the button.

So, if I go ahead and share this Flow with other people on my team, it’s not running underneath my credentials, but they all will actually be prompted to provide their own credentials and this will run underneath their context. So, as mentioned, this is an Office add-in for now, but it’ll be in the box soon. And so the scenario that I wanna run through is in the energy industry, and this is really around basically bidding into a market in order to basically sell energy. And then also once you’ve gone ahead and made those commitments, you need to be able to manage that obligation. So, today, Excel is still a tool of choice for this type of work where you might have a commercial analyst that’s going to go ahead and model out what are the revenue expectations for a specific week based upon what’s known capacity.

Now, the idea is that you have to, once you’ve made these commitments to market, you need to hand that off to an operations manager because someone actually has to manage the generation of this power. So, in this case, this could be a wind farm operations manager who now needs to ensure that he’s in the boundaries of the commitment, otherwise, there’s fines that can occur. So, once again, we can go ahead and use Microsoft Flow to manage this process and then we can also take advantage of Approvals. So, Approvals is a first-party experience inside a Flow where I can, before I actually go ahead and make that commitment to market, I can have a supervisor or someone else within the organization go ahead and approve what I’m about to publish and provided they do approve that, the business process will go ahead and continue.

Once we’ve actually made that obligations, we can provide a PowerApps to people in the field because if they’re going to go ahead and provide any sort of maintenance on a wind turbine, it could actually jeopardize or have an impact on the commercial commitment that was previously made to the regulator. So, I’m going to go ahead and now I’m in Excel Online. And I’ve pre-populated basically what I’m expecting from a commitment standpoint. So, for this upcoming week in the British Columbia region, I’m saying that, I’m the commercial analyst at this point, that the minimum energy that we will provide is 100 megawatts. Our target, this is like from a company perspective, we’d love to be able to generate 120 megawatts, but the max capacity that we can provide is 130. Here’s the expected price, the expected revenue, and profit as a result.

So, with this cell selected, I’m going to go ahead and kick off this flow. And I’m simply going to go ahead and click the Play button. I can provide a message, in this case, “Please take a look at this week’s forecast.” Click on the Run Flow. I can see the row that’s been selected, and then I’ll hit Done. And at this point, the actual flow has been kicked off. If I go and look at what that flow is doing, we should see that it’s hung up on an approval stage right now. This wasn’t the runtime. So, we can see that it’s running and that it is currently in approval stage.

Now, I can respond to this approval in three ways. One, through my Outlook inbox. I can do it through the mobile application, which we’ll show soon or I can also access this through the Approval Center inside of the Flow portal. So, I will see now I’ve got a new received request. In this case, it’s this week’s energy forecast. If I wanna reassign this, I can go ahead and manually reassign this if I wasn’t comfortable making that decision. But otherwise, I can approve or reject and provide some comments. So, I’ll just say, “Please proceed.” And I’ll hit Confirm. And really what will happen now is that Flow will continue. It’ll pick up and it’ll process the rest of the actions within it.

Now, in this case, we’re going to go ahead and if I was the operations manager, I could be looking at my team and see that there’s been a new record at 5:35 saying that the energy targets have been posted for this week. And if I head over to SharePoint, and I’ll just do a quick refresh. I’ll see a new entry here for the week of June 3rd to 9th where I can see what the expectations are. The target is 120, the min is 100, the max is 130. Now, what I need to do is I need to basically acknowledge this and select the turbines that I want to use in order to address this requirement. So, I can go ahead and use a multi-select feature in order to select these different turbines that will fulfill this need. Despite popular demand, wind turbines don’t spin all of the time.

So, I’m going to go ahead and save this. And what we’re going to do in the background is I need to be able to calculate what the output of those turbines are. I need to manage to the 120 number and I can actually create another flow where I’m going to go ahead and do that. And really all I’m going to do is listen for a new event or a new record inside of SharePoint that has been edited. And then, what I can go ahead and do is basically loop through all of those turbines that were selected and add up the total output for each of those wind turbines. So, if I head back to the SharePoint site, I’ll see that the number is 120. So, I know I’m in line with what my commercial analyst is expecting.

Now, in the event, I do have some sort of an outage, uh-oh, here we go, I can go ahead and now use a Power App. So, because this is a mobile-enabled application, I can provide this to my wind techs and they can actually go ahead and see what the expected outputs are. And in the event that there is some maintenance required, they can go ahead and fill out this form and actually determine whether or not this is going to jeopardize the financial targets that they’ve had. So, in some situations, the maintenance may be able to wait to a future work so it could be scheduled around it. In some cases, that may not be the case. But all possible and this is actually a real-world scenario based upon one of our customer business processes.

So, that’s demo number two. Now, let’s talk a little bit about Cognitive. Now, Cognitive Services is another way that organizations can actually go ahead and scale. And the scenario I wanna talk about in this case is what’s called intelligent customer service. Now, nowadays, there’s no shortage of channels. In order to get feedback to a customer or to a company, all it takes is you got to go watch Twitter and see someone with a lot of followers complain about a late flight and see how quickly people jump on top of this. Now, one of the challenges for many businesses these days is that labor costs account for a lot of their actual overall costs. And it’s not easy just to throw bodies at a problem. Now, once again, in order to transform your business and actually scale, you can actually use cognitive services in order to help with that.

Now, another thing we wanna do is we need to be able to fulfill these service requests. We can go ahead and use Dynamics 365 services and we can certainly use Teams and that collaboration in order to elevate with these requests. And this idea is that some requests are more urgent than others. But the challenge is how do you detect that? So, historically, if you say had emails coming in that were customer service related, a human would have to scan them. And they would use their own judgment in order to determine what is a higher priority. So, in the case of like an energy scenario, if someone has their power outage, that is a higher priority than someone that may have a billing dispute over a dollar. So, what we can do is we can actually use a service called LUIS, which is Language Understanding Intelligence Service, which will actually go ahead and understand the intent of a request in order to route it accordingly.

So, now we can elevate scenarios that require say, someone to deal with a power outage versus something that might be a lower priority. We can actually scale without actually adding headcounts to an organization. So, to demonstrate this, what I’m going to go ahead and do is I’m going to mimic a customer and send in a couple requests. Now, in this case, I’m going to send an email in that says, “My lights are flickering and I can see a trash can has fallen on a power line next door. Can you please investigate?” So, we’ll go ahead and send that in. That’s been sent. And then we’ll send another request in that says, “Why is your service so expensive? My invoices keep going up and up. Please take a look at this.” So, we’ll go ahead and send that in as well.

And then in the meantime, while that’s processing, I’m going to show you LUIS. So, in this case, I’ve gone ahead and created an app called Customer Service. And as part of defining this app, I go ahead and I create different intents. So, it’s really the intent of what my application is supposed to do or provide. So, in this case, I’ve got billing, customer service complaints or power outage. When it comes to power outage, I can go ahead and give examples of what a power outage might look like if a customer was emailing me. So, I provided a lot of these phrases, it’s very simple phrases, my power is out, my lights are flickering, my lights keep on flashing. And then what I go ahead and do is train this model, this machine learning model. And once I’ve trained it, I go ahead and publish it. And at that point it becomes available.

Now, with LUIS, you do get a REST endpoint. So, if you are a Custom Dev, you could call that endpoint yourself, but with Flow and Logic Apps, we have this connector. And this connector actually makes it a very simple experience for someone to go ahead and use. So, in this case, let’s go check on the process itself. And what we’re going to do is when a new email arrives, we’re going to go ahead and convert that email, because most of the emails come in HTML format nowadays, but we don’t need to send HTML to LUIS, we’re just going to send text instead. We’ll then pass it off to the prediction model. We’re going to get a response. We’ll go ahead and find the customer inside of Dynamics 365. We’ll go ahead and create a new record for it. And because we understand the context of it, we can actually create that ticket with the right subject, with the right title, and we can also go ahead and publish to the right teams so the right team is aware of it. And in this case, we can also go ahead and make a voice phone call using a connector called Infobip, which will actually take text and convert it to speech.

So, let’s just go back and look at the run history.




And in this case, this was the power outage scenario. So, “My lights are flickering and I can see a tree has fallen.” If I go down to LUIS, let’s see what LUIS interpreted the intent as. So, in this case, LUIS felt that the top scoring intent was power outage and it was 98% confident that it was a power outage. So, based upon that, it was able to do some intelligent routing to create the case in the Rights category. But was also able to go ahead and send a message to the right team indicating that this is a customer issue, the ticket number, and so now that they’re aware that this is actually an issue itself. We head over to the team itself and head over to field services. We could see at 5:41 that this ticket has been created and we get the original request from the customer. We also can email the customer with the ticket number indicating that, “We’ve received your request. Someone is looking at it. Here’s your reference number.” Which would allow them to track it and feel confident that someone’s actually looking at this.

If we head over to Dynamics 365 here, now, this is going to be in the Pacific Time Zone because that’s where my tenant is. But we should see two tickets were created. One was billing, so 9:41 AM and then one was a power outage. So, that’s intelligent customer service. And the last demo, who has seen the TV show “Silicon Valley?” A few. Okay. So, I highly recommend it if you haven’t. Now, within the show last year, he taught…this is Jian-Yang and he’s an entrepreneur and he creates this whizzbang application called “Hotdog or Not Hotdog.” And you can take a picture of a piece of food and determine if it’s a hotdog and if it is, you get a check mark saying it’s hotdog. If it’s not, it’s a red X. And so in the show he fictitiously sells this app for $15 million to Periscope.

Now, the idea is we thought, “Well, could you build this in Flow?” And so we thought about it like, “Yeah. We probably could build this in Flow.” And the idea is you can build this in Flow within probably 10 minutes once you know what you’re doing. And so what we’re going to show today is the fourth demo is “Hot Dog or Not Hot Dog.” And part of the purpose of this is to demonstrate our Flow mobile application and what we call virtual buttons. So, here’s the Flow mobile. You download this from the app stores, Android, iOS, and Windows. And you go ahead and any flow that has basically a manual input or manual trigger inputs, you can actually…it gets represented as a flow button here. So, what we’re going to take advantage of here, is not that Flow is so smart, is that Flow can connect to Custom Vision API, which is really smart. And based upon what it determines, it’s going to give us the green check mark or the red X indicating whether a piece of food is a hot dog or not hot dog. So, just to demonstrate this, I’ll just flip over to my mobile.




And so here’s the Flow mobile application. I can go ahead and see my activity, including my approvals. If I had outstanding approvals. I could go ahead and use my virtual buttons. I can actually see all my flows, I can edit my flows, although I wouldn’t suggest that as your primary editor. And I can also switch accounts. So, in this case, I wanna go ahead and use the Integrate 2018 Demo Hot Dog or Not Hot Dog. So, what a colleague and I did as we went around the Advanta Campus at lunchtime, we took pictures of food, and then we actually tested our application using this. So, in this case, that looks like a hot dog, when we go ahead and submit this, we should get an indicator which will give us the confidence. In this case, it was 99% confident that it was a hot dog. And so this works. But let’s test it, let’s do a negative test. And in this case, this is a pita with a Souvlaki on it. So, let’s go ahead and submit that.




This file is a little bit larger so it might take a few more seconds to…




So, in this case, it was 36% confident that that was a hot dog, which I’d say is good because it’s obviously not a hot dog. Now, to really make this work, as I mentioned, the magic isn’t so much within Flow, but the magic is within the Custom Vision API. So, how you actually go ahead and build this out, is you actually log into the service and you upload a bunch of sample pictures, then you go ahead and tag them. So, in this case, I’ll click on this hot dog. It’s been tagged as being a hot dog and is having relish. And you would go ahead and you can do this for other attributes as well, including ketchup, mustard, et cetera. So, I’ll just show you the flow now. So, in this case, we’ve got to manually trigger a flow. In this case, the input is going to be a file. So, basically, we’ve got native integration with the camera itself. We can upload a file, then we’re going to go ahead and call the Custom Vision API and just provide a project ID. So, basically, I get that from the website where I uploaded all of those different images. I went ahead and trained it similar to LUIS and then I’m just going to pass the image content right to it.

It’s going to respond back with all of the predictions. So, there’s five different sets of tags that I had there. I had hot dog, I had relish, I had ketchup, and then what we wanna do is we actually wanna find the tag that is specific to hot dog. It’s not relish or not relish, it’s hot dog, not hot dog. So, we wanna find the tag for hot dogs. We’re going to loop through those and then what we’re going to do is we’re going to get the probability. So, we need to provide some sort of a threshold here. And we’re basically going to say, “If the threshold is greater than 85%, then we’ll call it a hot dog. If it’s less than 85%, then we’ll say it’s not a hot dog.” So, that’s exactly what’s happening here. If it’s greater than 85%, we will just send the mobile notification saying hot dog and the probability, so, that was the 99%. Otherwise, we would do the same thing if it was no.

So, you might be thinking, “Okay. This is kind of a bit of a gimmick.” But the reality is we didn’t really come up with this idea. The Silicon Valley was sort of an afterthought. We actually came up with the idea through a customer. And in this case, it’s a pipeline company in Canada. And what they have to do is they have to manage the vegetation around their pipelines. So, there’s specific vegetation that you cannot pick or remove, you actually get fined for doing things. So, previously before this PowerApps and Flow solution, is they actually carried around a binder. They’d handle a binder to their field staff and they would actually have to clean up the vegetation because they don’t…around the pipeline and the right of ways. And they’d actually go through the binder and say, “Oh, is this something we can pick?” And they would actually write it down and they write down the location. And then they take that back to the office and they would create work orders. And they will then contract that out for someone to go out.

So, they had all of this manual work. Today, they actually have a Power App that uses Flow, and they go ahead and they take a picture of the plant and they’ll see if this is something that can be picked or not. And then what’ll happen is if it can be picked, they use the GPS from the phone, and they actually record that in their LAN system, which is Dynamics 365. And they’re now automating the actual, like the workflow process of all of this stuff. So, instead of their users now carrying around these binders, they’re actually using this technology to enrich the dataset every time they go out. So, they’re using this Power App on all of their field staff and people are now encouraged to go basically catalog this type of information.

So, it doesn’t just apply to pipelines and weeds and things like that. You can think of many different sorts of inspection scenarios where you can actually use this data to indicate what something is. And this is the whole sort of idea of this democratization is that you don’t need a Ph.D. in artificial intelligence to build these solutions anymore. You can go ahead and take advantage of a lot of these different tools and use your creativity in order to come up with these compelling solutions that actually drive business value. It’s funny. Whenever we show this hot dog, not hot dog, it always generates a lot of ideas around this technology because it can have an impact. I did ask them what the name of the app was. It’s not “Weed or not weed.” But it’s a much more corporate-friendly name.

So, last slide. This is a roadmap, both sort of immediate and short-term. And so if you want. ..I wanted to highlight a few of these. So, one is the business process with stages. I talked a little bit about that before. This is something that should light up here soon within a few weeks sort of the first iteration of that. Microsoft Teams trigger with integration. So, I showed you sort of our initial Bot. Obviously, you couldn’t really pass many inputs to it or get direct outputs in the chat. So, those are things that are being worked on. So, I thought this would be interesting to this crowd is the data residency within the U.K cloud geo. So, we’re not in every single Azure region. It is something that we are continuing to expand upon and the U.K is up next for us.

Azure functions integration is something that is being worked on right now. And the ideas that we will sort of provide an experience where you don’t have to go to the Azure portal in order to build the function. We’re not going to support Visual Studio, that’s a Logic App’s use case or just functions native. We will provide it an experience within our portal to do very simple functions. And the idea is that they should be simple. If it needs to be more complex, there’s other pro-Dev tools for that. And then also additional admin and governance capabilities. So, this is something we’re actively working on, including admin analytics where we have a Power BI embedded experience for both environments and tenant admins to look across their entire organization in order to understand how people are using Flow, what are the connectors that they’re using? Who’s creating flows? Who’s sharing flows? And really bring all of that data together.

So, with that, that’s my presentation. I’ll be around all week. So, if you are using Flow, come talk to me. I’d be really interested in your use cases. I do have some Flow stickers and physical buttons as well. I only have a handful of these. So, if you show me a really interesting flow, I might be able to get you one of these Flow-branded flick buttons which are similar to the virtual buttons, except it’s a physical button. When you press it once or twice, it’ll actually kick off a flow. So, thanks for your time and enjoy the rest of your conference.

Download Presentations

Fill the form below to get all the presentations delivered as a single zip file in your mailbox.

please wait …..

Interested about future events?

please wait ...