Sourcing Automation with Power Desktop Automate

All right, welcome everybody to Shelly's Alley, and this is a reboot of the last, um, complete and categorical failure from last Friday where I basically forgot to turn on um, the channel that shares my desktop. So I had broadcast my desktop and I thought I was showing it to everybody, but then I wasn't. So, um, yeah, anyway, it was one of those days. Um, today instead of making that mistake, I'm going to make sure that I am sharing my screen right now. There we go. And rather than switching back and forth between the two, what I'm going to do Okay, how about that? Now can you hear me? Tunnel to infinity. Well, hopefully this is now doing both the video of me and the video of my desktop. I hope— can somebody confirm you can see and hear me? What I was saying is that you can join the Facebook chat by going to chat.restream.io/fb. I'm going to put it in the chat again, and if you go there, you're able to connect your Facebook to Restream so that you can put your, um, what was it that I was saying? Oh yeah, your image and all that in there. So okay, it looks like everything's on and I've got both me and the desktop. I'm going to go jump right in. Once again, this is a reboot of my new favorite toy, Power Automate Desktop. I started playing with Power Automate Desktop some time ago and as an extension of the Power Automate tools that are already built into Microsoft 365. But when Power Automate Desktop came out, it was something that you had to pay. I used my trial period, played around with it, and then lo and behold, I found out not that long ago it has now been made available for free. And not only is it available for free, but this is really exciting, it is going to eventually be built into Windows. So on some update, they will— right now you still have to download it, and I've got instructions on that, but they will actually put it into Windows so that it will just be part of your Windows configuration and your Windows setup, your Windows install. If you're wondering why that's a big deal, it's basically because as a Mac user, I use both PCs and Macs, there already exists a very similar tool to what I'm about to show you today, which is Auto the Automator, and that's been in Macs for a long time. So basically, this is the Microsoft equivalent And that's important because a lot of us have to use laptops, Windows, I should say, laptops and PCs for work purposes and may not be able to have a Mac at work. I've been kind of running both, but that's just me and not everyone has the ability to do that. And I can't connect my Mac to work, so I usually do other things on that that are more experimental and stuff. So for me, it's a really big deal. What is Power Automate Desktop? And really, what is Auto the Automator, or what is any of this? This is basically part of what is essentially a, uh, what they call a citizen programmer revolution that is taking robotic process automation, or RPA, and bringing it into the everyday audience so that people with just basic laptop skills when Microsoft Desktop, Automate Desktop, is available for everyone. Right now you can download it, but when it's available for everyone, the idea is that people will be able to create automations of the day-to-day activities without needing to do any code. So essentially, if you remember macros, macros from Excel or just VB macros, and even if you've just heard of them, think of this as really easy super-powered macros that require absolutely no coding skill. You don't have to code to be able to do it. There are some actions that do require that you type a few things, but it's not coding, it's just editing, much like you would edit a document of some sort. So you don't need any code, but what you can do is you can record actions on your desktop. In other words, hey, I'm gonna do this while the computer watches me do it, and then the recruiter can— the computer can remember that and repeat that action, do it repetitively or on demand, etc. You can also do the same thing on websites. So it's kind of like being able to record a macro on your computer or on a website. And the third piece of this is putting applications together. If you were a fan of this, this If This Then That, or IFTTT, they went to a model where now you can only have 3 recipes running for free, the rest you have to subscribe and pay for. But since this application, uh, Power Automate Desktop, is free, essentially you're doing a lot of the same things. And it's way more sophisticated than If This Then That because you can do a lot of connections with things that include recording on your desktop, pulling in data from the web, etc. So what can you do? Here's just 3 big things. You can organize your documents using folders and actions. That means that you can move things around, take it from here, put it there. You can take documents out of your email and put them in a SharePoint or a OneDrive. You can take documents out of your Gmail and put them in SharePoint. You can take documents out of a folder and send them via email, et cetera. You can rewrite documents, reformat them, take a Word document and convert it to a PDF, take a PDF and convert it to a Word document. All of this organization can be done automatically now through these macros. The second major category is extracting data or essentially copying data from websites and storing that data in files like, for example, Excel. So you could grab data from a website and put it in Excel. You could grab data from Excel and put it in a website. You could grab data from you know, basically various different Excels and combine them into something or put them into Power BI to create a nice report that you couldn't do otherwise because you have 10 different Excel spreadsheets with different rows and columns and they don't really match up. So those are all just ideas of extract— what's called extract, transform, and load, or ETL. ETL is a fancy way of saying take it from here, Do something with it, put it over there. And that's what a lot of IT people do when they're converting databases and things like that. So we now have the power to do this. If there's data anywhere, CRM, an ATS, a document, a SharePoint site, a list, if there's data anywhere, you can grab it, transform it, and load it or save it somewhere else. And you can also go on autopilot with some of the things that you already do on your desktop. If you need to open multiple applications, for example, to do a simple task, or if you have to grab something from one application to put it into another, take data from one app and put it somewhere else, or if there are just activities that you have to do on a regular basis, you can automate clicks, a simple, you know, click automator that just automates the clicks on your page. Let's say that you want to click a bunch of checkboxes or click a bunch of, you know, things on your LinkedIn network and you wanna, you know, click on, let's say, remove a contact or message a contact, you can automate that simply right on your desktop. There's lots of other things you can do. So to get started, you're going to have to download the application and install it. This is a, um, basically a desktop app that sits on your machine, and while it does connect to the cloud, it can back up all of your flows and workflows to the cloud. When you create something on your machine, it is on your machine, meaning that if you could— you put it on this PC and you then are using another PC at work, or a different PC, a second PC, don't expect what's on this machine to be automated on that machine. You can, however, open the flows on the other machine and import them to that machine and, you know, set them up. But it is essentially exclusive to the computer that you're working on. That's different than Power Automate on the cloud, which was the original version, which runs independently and had things that you could run automatically, things that you can run based on a trigger like pushing a button or an action happening, or things that you can run on a schedule. With the desktop, it's just on your machine. The first link is to download it, then there's a link with some easy instructions to set it up. I'm not going to bore you with that detail. It's a simple one-page document to help you set it up. And then you can download a couple of other white papers on, you know, things that you can do, ideas on what you can create. And the fourth is the extension that you will need in order to record stuff from a website on Chrome or Edge or Explorer or Firefox. You will need an extension that connects that website basically, or really that browser to the desktop app. And you need that if you wanna do any web-based recording or actions that are on websites. So those are the four components. Let me go ahead and open the one that I installed. And I wanna show you two quick basic things here. I mentioned before that there are cloud-based components Automations, this is where you can find them. These would be if you had any saved. I don't have any in this particular environment. By the way, I'll explain where an environment is in a second. But if you had any saved, they would show up here if they're on any of your machines, but they would have to remember, they would have to be set up for this computer in order to run. And then there's also flows that have been shared with you. So you can share your flows with your team or with other people and have, you know, team workflows, et cetera. That's what these are right here. Then you have the option to create a new flow, which is what we're going to do. You can do that from either one of these places and also change your environment. An environment is essentially your, um, your account. So I have two environments right now. I've got my business one and then I've got my secondary one. You can have different flows on different environments. And that just means basically different accounts. So perhaps you can have a Hotmail account that is your, your, you know, your personal Microsoft 365, and you can have a work account that's your business 365. I'm not going to bore you with a lot of the settings and things like that because you can figure that out pretty quickly, but, um, I do want to show you what we do when we go ahead and create a new flow. So I'm going to start with a new flow. First thing it does is ask you for a name. I'm just gonna basically call this Playground. This is kind of like my sandbox or test environment or whatever. What you'll notice is that it essentially just creates it for you, but then you need to edit it. Now it should be opening the editor behind the scenes. There it is, it just popped up. Once this opens, this is basically where you do all of your editing. On the top here, you see the two types of recordings that you can do, the web recorder and the desktop recorder. Obviously, the desktop recorder is going to record things that you do, click, open, drag, whatever, type on your machine versus the other one, which is the ones that you do on the web. Sorry, just one second here. There we go. All right, so, um, to do something on the web or record something on the web, as I mentioned, you'll have to have the extension for that browser. But don't worry if you haven't downloaded it yet. When you click on web recording, it says, oh, how would you like to record? Would you like to use Edge, Chrome, Firefox? Or IE, and it would take you to the appropriate place to download the extension if you didn't already have it. Let's say that I want to use Chrome. I already have the application— the extension, I'm sorry— added. Then what I need to do is pick which currently open window, um, where I should say tab on that browser, because as probably many of you have lots of tabs open, you don't want to choose a tab that that is something different than what you wanna record. So I'm picking this particular tab. You can also create a flow that opens a new website or opens a new tab, or you don't have to pick one that's already open. Once you do that, then it will take you to that new tab that you just created and the recorder is already configured as an extension. So it's now able to capture just like on your desktop, whatever you do. This is very useful, for example, for when you are— I think I kind of left it hanging, didn't I? There we go. All right, yep. This is very useful for when you are filling out forms on websites. So there you have the two types, the web recorder and the desktop recorder. And then here is the last piece of the puzzle, which are the connectors. This is the part that I ran into a little bit of trouble last time with, so I'm glad I'm doing this over because I didn't have the updated version and I couldn't find all of them. There are over 400, I believe closer to 500, at least 460 connectors and counting. This is similar to the connectors on If This Then That or Zapier where they're like the tiles for the product. If you go to this web page here, you will see all of the connectors and you can see a summary of them and what they do. I encourage you to browse through there to start thinking about what you you can automate, give you some ideas on what's possible. Most of them are free. Some of them are premium if they have a little diamond on them. And those premium ones require either that you have that subscription already. For example, let's say you use Salesforce. Well, Salesforce is premium connector, but since you maybe already have a Salesforce account, it may be included with your account. So premium doesn't mean that you have to pay extra necessarily. It means that it belongs to something that you have to pay extra for and you may already be paying extra for that. Your company may already have a subscription to QuickBooks or whatever. So there are connectors that are built-in and there's connectors that are premium and then there's also connectors that you have to buy that are separate, almost standalone apps that people develop to sell through, through the Microsoft Store. What are all these connectors about? Well, just to give you an idea here, There's a bunch of them and I realize this is probably really teensy teensy, but you can essentially drag and drop. So I'll just give you some examples. Here is the PDF connector, and what the PDF connector allows you to do, for example, is extract images from a PDF or extract text from a PDF or extract PDF files or even merge PDF files. So you have a bunch of PDF pages that you want to merge together, These are all actions that you can do, and if you wanna do one of these, you simply just drag it over here, and then it starts to ask you some questions like, where are the files? You point to the folder, and you can do some other, you know, advanced configurations if you need to log in, or if you need to rename the files, et cetera, but it's pretty drag and drop. So that's just one example. Another example that I've been playing around with is playing with text. Add to text, you can trim text, you can convert text. So let's say, for example, that you're looking at an Excel spreadsheet that has a cell that contains a bunch of information in it and you want to separate that information out. There's multiple ways to do that. You could use Excel and just do text to columns, for example, but the problem is that if there's not a clear delimiter, then you might end up with more columns in some fields and people and others and it's a mess. But with this, you can use this text processor to split text and tell it specifically where to split. So you can say, hey, go to this Excel file on this column and every time that you find text that has this, split it into this and put it into one cell, one row, whatever. You can put it into another document, et cetera. You can also parse text. This is where you start now scraping information from anything that you want. You can scrape from a website that has a list. You can scrape from a PDF document. You can scrape from an application. Wherever there is text that can be captured, you can use this to parse that text. And parsing means breaking it up into the different columns. You can also do find and replace. So when a particular text appears, it can be replaced with something else. Different ways you could use this, for example, you can have your signature file. So when you type in SS signature or, um, something like, you know, asterisk signature, then the flow will automatically take that piece of text and replace it with all the information in your regular signature so that you can use your signature on any social media site or texting app or CRM or whatever You can do the same thing with more than just your signature. Texts that are, for example, descriptions or links or directions and so on. This is pretty powerful. So powerful, in fact, that it even allows you to use what are called regular expressions. That means that you can take pieces of text, convert them, extract them, parse them based on formulas like only numbers or only capital letters or only numbers when they follow a capital letter or only numbers after they follow a hyphen. And so on, lots of different things. Similar to text, there's also date and time, which allows you to do math with time, which can be very useful for calculating when things are due, how far away they are, those kinds of things. There are optical character recognition examples, and you can see that these are coming from a couple of different places. This one here is extracting text. So if you have a document like a PDF, that has been scanned in but doesn't have any text that can be copied and pasted or extracted, you can use optical character recognition to virtually scan it and get the text and extract it. You can do that with the built-in engine, or you can use the Tesseract engine or the Microsoft engine. There's 3 choices depending on what kind of text. This is a very intro-level course right now, so we're not going to cover that, but you can grab text out of files that don't even have the ability to copy and paste text. You can encrypt information or decrypt information. You can control flows. So these are things like conditions such as if there's an error, do this. Think of it as if this happens, then that. You can split actions. You can have subroutines that work when you have certain conditions that are met. There's an FTP that allows you to upload files, download files, convert files, delete files, rename files from a server. And one of the ones that I find the most impressive here is cognitive. Where did it go? There it is. Okay, here. So there are 3 types of cognitive automation. There are services from Google, from IBM, and from Microsoft. Each of them have developed refinements in specific areas. Everyone knows Google has a good search engine. So one of the things that Google has done exceedingly well is analyzing sentiment and entities and syntax. So you can use this cognitive solution or this cognitive action to grab the sentiment or the entity or the syntax of a bunch of documents or text. IBM has done a really good job of tone analysis, and that's different than sentiment. So if you want sentiment analysis, you can use the natural language Google processor. If you want tone analysis, you can use IBM Watson's tone analyzer. And with Microsoft, they have really nailed it when it comes to transcription. So you can take— use the audio function to capture audio and convert audio to text. I am most impressed with Microsoft's ability to do live transcription. That's when you're speaking, let's say on a PowerPoint presentation, and it is typing what you're saying in real time, like closed captioning. Really accurate. Conversely, Google isn't as good at that, but it is really good at translation. So if you want to translate, you could use Google. If you want to transcribe, you could use Microsoft. And with computer vision, there's also some really important and powerful things. Google is really good at detecting patterns and properties of images, finding other images that are similar, whereas Microsoft is exceptionally good at describing an image. So if I give it a photo of myself, it comes back and says, man with glasses and a beard, for example. Very descriptive. With Google instead, if I give it a picture of myself, it'll find other pictures that are similar that perhaps are other pictures of me or people who look like me or other people that have beards and glasses and so on. So it's very different functionality. You can see that Microsoft also has optical character recognition. You can do things like automatically tag images, etc. So these are just some ideas of things that you can do with cognitive automation. There's much more that you can do. You can have automation for your user interface, like data extraction. This is basically scraping. So there are already pre-configured activities that allow you to extract data from a website or fill out a form. How many times do you have forms that you have to complete repetitively, like to create a rec or to create a record on your CRM or your ATS? You can use this to extract record from your CRM and parse it and paste it into your ATS. You could literally run a process that grabs the candidates from a particular list and imports them into your ATS yourself through a robot without having to hire a developer to do that behind the scenes. It's doing it on your machine with all of your security and permissions. It's basically doing what you would do when you copy and paste, but it's doing it for you, and it's already pre-configured so that all you have to do is point it to here's the field that I want to grab from and here's the field that I want to paste to. You do the grabbing and pasting once and it remembers all the field names and everything. And when you execute that, it will then do the grabbing and pasting automatically for you. You can automate email responses so that when people apply for certain jobs, they get a reply that is tailored for that job. Or perhaps there are different stages in your e-screen. You can For example, ask a question, they answer it, you can use the sentiment analysis to determine if it's appropriate to respond, and then you can respond with a variety of templates. You can automatically scrape information from a website that's already structured, like say a list of speakers. The scraper will recognize the fields and extract them for you into essentially an Excel, and you can then take that and put it into a list or put it into a CRM, whatever you want. You can also move or scan files. For example, what if all of your resumes were in SharePoint or in your Outlook inbox? You could use one of the analyzers like the sentiment analyzer or the pattern recognition cognitive analyzer to look at all the resumes in a particular folder and detect similarities. This is similar to the word cloud exercise that I talk about. When I say put 5 resumes that are similar into a, you know, a word cloud generator and see what the patterns are that emerge. This can do that, but with everything in a folder. So essentially, this is what you are paying for with machine learning. All those AI apps, this is all they're doing. It's not really AI. It's not even really machine learning. It's looking at a bunch of patterns based on analyzing a bunch of text. Well, you can do that. You can say scan these 100 resumes and tell me what the commonalities are. Boom, machine learning, not really, but you can do that with this. You can also extract candidate interactions from social networks. Say that you are reaching out to people via Facebook, Twitter, Pinterest, Snapchat, how are you capturing when they respond? Well, you have to copy and paste that into your log or whatever. You can use this solution to capture the text from all of these responses and merge them into a single document. You can say, when this person with this name responds in any of these places, grab those responses and merge them together and tell me where you got them from and put them into this file, which could be like, say for example, that person's note field in your CRM or whatnot. So those are just some ideas. Hopefully I got your creativity juices going. This is really just the tip of the iceberg. I'm going to show you a few of the, um, actual flows that I'm working on, and this is more of a lab, so some of them may or may not work, but I'd love to go on this learning experience for you. Do me a favor, if you have any suggestions, ideas, or requests for what I should try to automate, drop me a note you know how to find me, not that hard to find. Just shoot me a text message or write a comment in one of these, either LinkedIn or Facebook, and let me know if there's something that you've been thinking, oh, you know what, I wish I could really automate this because it's like a big drag on my day, right? I spend hours a day doing it. Let me know what it is and, you know, I'll look at it because if it's causing you pain, it's probably causing a lot of other people pain, maybe even people on my team. That I can help, so I can use the idea, and then I'll talk about it on our next session. So please let me know if you have any ideas. Don't forget to go to the Facebook group and join so you can get the notes and the handouts. And that's all for now. Thank you very much for coming, and I will see you next Friday on Shelly's Alley. Bye.