Cisco Customer Success - Video

Workflows - Network Automation for Meraki

Playback confirmation!

Do you want to resume playing from last left location?

Chapter 1 Thumbnail
00:02
Welcome, agenda, and housekeeping
Chapter 2 Thumbnail
02:05
Network automation landscape and challenges
Chapter 3 Thumbnail
05:22
Introducing Workflow: platform and benefits
Chapter 4 Thumbnail
13:06
Getting started: Exchange and Workspace
Chapter 5 Thumbnail
19:56
Building blocks: variables, targets, and logic
Chapter 6 Thumbnail
29:54
Running workflows and monitoring results
Chapter 7 Thumbnail
41:04
Targets and integrations; SSID demo
Chapter 8 Thumbnail
45:18
Automation rules and webhooks; PoE outage demo
Chapter 9 Thumbnail
59:11
User tasks and approvals; controlled changes
Chapter 10 Thumbnail
01:06:08
AI assistant, best practices, limits, and resources

Transcript

Welcome to workflows network automation for Meraki. This is part of the expert insight series for EMEA, I'm San Bernardi. I'm a customer success specialist for Meraki. And today we're gonna go through a little bit of an understanding of what workflow is and what it addressed, and then we're gonna see some use.00:02Cases where we can use it and then we're gonna deep dive into demos and examples. Just as a reminder, all or almost all the example I'm using today will be available and later I will share a link with you where you can have access to the same exact workflow that I used in the presentation as well to other resources that.00:22You can use to further enhance your ability to use and understand workflows for your environment.00:42What is gonna be the agenda? Well, we're gonna look a little bit of the landscape where workflow is inserted into, what are the tools that we have available for automation, and then we're gonna dig deeper into the interface, the feature, and into a series of demo. We're also gonna learn how to access it although.00:53If you have a Meraki dashboard, it's already there for you. If you don't and you were just curious, hey, how do I tap in in this environment, just create a free account for the Meraki dashboard and workflow will be there, as well as some resources that I will share with you at the end of this webinar.01:10Before we jump in, you found the chat. There's also the Q and A panel, so feel free to use one or the other. I'm gonna try and address every single question on the fly if it makes sense in the flow of the presentation.01:31But if not, we're gonna have a little bit of time at the end where I can go over everything that was left and answered, so feel free to post question in the chat. If anybody knows the answer again, it's a great way to share the knowledge if you already are familiar a little bit with workflow, but if not, don't worry, we're gonna have time at the end.01:44In terms of the landscape and the complexity, well, network management is complex, ok? We're not here to say it's easy. That's why we all work in this space and that's why we have so many different things we have to manage.02:05Sometimes you might be a customer with a single series of product and maybe you manage everything from the Meraki dashboard, but you want a little bit more of automation, and sometimes you are an environment that are more complex than these. Maybe you have Catalyst Center, maybe you have Nexus somewhere. Maybe you use 3rd party interaction like servicenow.02:22Now or you're using the SD WAN solution from the catalyst series, all these things can exist in your environment. And as always, when we design a process, we would want this to be smooth and propagate everywhere. All the elements in the environment need to interact with each other and then.02:42Sometimes lead to error prone process, a little bit of wasted time because we forgot about something. We did made a title somewhere or we forgot to change a specific configuration in order for everything to work, for everything to work. And it also kind of make it difficult to track change.03:02Just because you have various systems and not having a single place where to issue those configuration or to issues those changes and rest assured that will be applied everywhere can get the the the process of tracking what is going on very hard.03:22Yeah. What if we had a way to change that? And obviously that's why we're here here. That's why we are here today. Before we had building and executing ask across multiple domains as a thing you have to do in every single place where you have to do.03:40Keep your hands into. Tomorrow we want to ship to simplifying operation across all the networking domain.04:00Having one dashboard, one central automation solution for Cisco and 3rd party system, and that's exactly what workflow is designed to let you do. What's the idea behind it? It's consolidation. We know, as always different.04:07Networking dashboard catalyst managing Meraki traditional Meraki device, SD one product common policy that we might want to have across our domain. These are all challenges, these are all places where you might have to go and click something or change something.04:27On the other end, we also have security as one of our top of mind. So maybe you have a unified sassy solution, maybe you want to protect your user. Maybe you have protection directly built in in the cloud, whatever it is.04:47It's always another place where you have to go and tweak, change and configure everything. Workflow is the platform where we can converge all these tasks in a single interface where we can define what we want to do and execute.05:02With directly from the cloud. So there's NO need for you to bring in additional pieces or have something running in a, in a cloud provider or on your laptop. It's a little bit of a mind shift. Previously in this series we spoke about API, predominantly API solution.05:22You have to have, you need to have something either running on your laptop or running on your machine or have a machine in the cloud that is running your script. In this case, everything is aggregated within.05:43The Meraki dashboard. With that being said, that doesn't mean that we cannot bring in a 3rd party solution. So as always, we do have an ecosystem that is open. If you want to export information into 3rd party system, if you want to have that additional piece of ma.05:56Machinery that should integrate with your workflow, you can. So again, you can just choose to stay in the ecosystem only and run everything from there or you can add additional component and pieces to make this your orchestration platform throughout all the domains you might have to manage.06:16Okay, that's exactly what we're planning to do here, ok? You have workflow as a platform. It is cloud hosted in the Meraki cloud, so it's there for you, you just need to tap in. The interface is drag and drop, so it's easier even for.06:36For people with not too much knowledge about programmability to jump onto the train. Maybe you don't want to write complex script, maybe you don't know how to build a user interface, and we know we have AI LLM assistant there that can help you with this, but having a more structured environment can help maintain a.06:56Everything's secure, everything's safe and everything within the boundaries of your organization. Also, having this platform allow us to custom build or custom import workflows that are built by somebody else. So once you have something that is working, you can propagate it across different organizations.07:16Maybe you are a service provider and you want to give this as a service to every single user that is buying into your program or maybe you're just a large organization and you want to be able to replicate this across multiple organization. Workflow allow you to do that very easily.07:36And as I mentioned, 3rd party integration is always available, obviously, with some degrees of flexibility. We still leverage API in the backend. So workflow will trigger an API call to a 3rd party system if that system is capable of receiving and responding to those actions.07:55Will come into play or we will go over concepts like webhooks, so HTTPS based communication across automated system.08:15And, we also have traditional API endpoints that we can make or receive an API call from outer entities compared to workflow. In general, the main goal for workflow is to mitigate the risk, reduce manual tasks and have a faster.08:25Deployment time for anything you might have in your environment.08:45And yes, this was possible with API, but sometimes the barrier to entry was a little bit too high and the learning process there could lead to spending more time building the automation than the time you can save by actually using it.08:50The other beautiful thing is once you wrote something in workflow is available to every other admin user, so if you build something that could be useful for everyone, just spread the news across your organization and let them know that with a single click, they can do a task always in the same way, always maintaining.09:06Inconsistencies.09:25As I said, it doesn't work only with Meraki. The Meraki dashboard was the logic place where to put workflow. But with that being said, if you don't want to manage anything within Meraki, but you want to manage other Cisco product, the place, the entry.09:30Point it's still the Meraki dashboard. You can just bring in other Cisco product like Catalyst Center, ICE, the Catalyst SD one, AP controller, and so on.09:48Obviously this list is constantly expanding, so more and more is getting added to the platform as we speak. Why workflow is also very useful for, I'm not saying even beginner, I will say it's a good start.09:59Starting point, if you have an idea, if you have a problem, if you have a thing you're trying to address, most likely you're not the only one out there. Obviously, pre built solution, they tend to be kind of restrictive because maybe somebody has an idea in mind and works really well in their environment, maybe worse per per.10:19Perfectly in yours as well. You can just grab it and use it. But also because this is a dynamic environment, you can copy something that was done by somebody else.10:39And modify to fit your needs. We know how it is. There is NO one size fit all. Every single organization, every single network, every one of you will have a different approach in terms of what they consider a baseline in terms of what they consider a correct configuration in term of what they consider.10:49Important for your day to day job. So this ability to.11:09Grab something that already exists and modify it. It's very useful to make sure that you always can get what you want out of a workflow. And it's customizable. If it doesn't exist, if it's not a lift and shift, if it's not modified something that already is available, you can custom build from scratch. We're gonna view.11:14Do a couple of these things in the demo later.11:34Obviously, despite being a drag and drop interface, there is a little bit to learn in terms of what are the expected inputs, what are the expected outputs, so as I always say with programmability in general, it's true for API, it's through workflow, it's true for the next session that we're gonna have about Ansible and terraform. If.11:37You don't have experience, start with something small. Sometimes it feels like hey, I've achieved nothing, ok? I managed to get the list of all my organizations. Well, I can do that already in dashboard with a single click, but that most likely will be the 1st step in order for you to be able to perform some.11:57Thing else and fetch information at scale or configure those network in a single click. So again, start small. I know that the idea is like, whoa, I can automate everything. Well.12:17Step by step yes, if you start already with a huge project unless it's something that is pre built, it can be done thing even with workflow. So again, build the experience and leverage it. I'm sharing from a personal perspective, that's exactly how I learn to use workflow with.12:30Simple task 1st and then moving on to more complex and more interactive workflows. With that being said, we need to talk about user interface and the features and.12:50Again, it's a drag and drop environment, but I don't want to stay here. As always for this EIS session, what we try to do is to do demos, so bear with me as I find my browser and I show you where we can find this information. The slides will be there for you for.13:06Future reference, but I do prefer to show it to you live directly in dashboard. 1st and foremost, let me zoom in a little.13:26We have a dedicated menu for workflow that is under automation, and you can see here that we have a few menu items. Don't get fooled. It's a very deep solution. So for each and every of this menu, we're gonna have plenty of options.13:331st and foremost, I wanna start from the exchange. The exchange is that place where you can install the workflow that are made by somebody else. Some of these are Cisco managed, so these are examples that we provide for free. It's just a marketplace pretty much.13:51With NO fees. You go in, you can see more obviously if it's here, it's probably gonna have a good description of what the workflow does, what are the inputs and the output and what's the use case that it's trying to solve. Some of these, as you might see are also community based, so.14:10In some situation, we get somebody in the community that's built something that's so good that go imported into the exchange. Obviously, Cisco manage all fine community based, as always, take a look at it, understand how it does, understand if it's working in your environment, which is true for any workflow.14:28So you might want to import. One click away, you click install and that gets imported in your environment. You can see here that I already have one, e.g.. If you want, you can uninstall it and just remove it if it's needed.14:47Here's also the place where you can submit workflow for the exchange. So if there's anything that you're particularly proud of, you can submit it for everybody to use. The other place, and it's probably gonna be the place where you will spend most of the time is the workspace. The workspace is where.15:04Where you create and run your existing workflows. You can see here that I already have a few, I will say, but most likely the 1st experience will be, hey, I'm gonna start from scratch. I'm gonna have a blank custom workflow. So top right corner, create.15:23You can click continue and I'm gonna be EIS.15:43Workflow demo, and I can continue. You will see here the interface will change and this is the place where you can build your workflow. What do you have available? Well, 1st and foremost, it's every single integration that we have. As I mentioned before, it's not.15:49Limited to Meraki only. This session will be.16:08But if you have other product, other things that you want to integrate 3rd party system, Splunk Cloud, ServiceNow, I don't know, you, you name it, ok? It's a lot, ok? That's why we have a useful search box if you're looking for something classic example.16:11I wanna go with get organizations.16:31Here we go. You can filter it out pretty quickly. You can open and close these dropping down menu, so again easy way to find what you want. And then it's as simple as dragging and dropping.16:36I'm gonna zoom in a little. The moment you drag the 1st task in it, you get, I mean, it's a flow chart in the end. You get a start and you get an end and then you can add as many action as you want in between. From the start perspective, what is important? If you go into the general, you can write.16:52The display name description is very useful especially as we will be able to see later.17:12This description can be used from the AI assistant to better understand what the workflow does, so always useful to have. Then you have some categorization option and you have variables. Now, each workflow.17:18You can have, I would say three type of variables, let's see if I remember correctly. Okay, actually four. You can have input variables, so it's something that before the workflow starts can be provided or must be provided for the workflow to work. So input variables.17:34You have your types, you have your name for the variable, and then you can say, hey, this is required. E.g., I want to know a specific network name. I want to know a specific user. I need to know the email destination because at the end of the workflow, the record will be sent via email, whatever it is. Or it could be something that.17:54If you put it, the workflow will behave in a way. If you don't, give me this parameter, the workflow workflow will behave in another. You can also have some starting values, so especially when you're testing your workflow, you don't have to input this value over and over. You can just use it as a placeholder and use it for your testing. Other type.18:14The Variable are locker variable. This is something that is used by the workflow at runtime is not influenced by the input of the user. It's just there for you to store information, to accumulate data, to remember what's the organization ID you're working on and so on.18:34And again, same type of interface and in the end you have output variables what you expect from the workflow to be generated. This could be an error message if something goes wrong or this could be the actual value that you were looking for.18:52It can be any type of data and to be fair, what I usually use is JSON data because it's nice to copy and paste somewhere else you can use even markdown formatting. In the end, personally, what I usually do if I just need visualization, I just add send email or.19:09Send a message to a Webex chat, so I have it nicely displayed in a place where I can recover it fast.19:29We're gonna see a couple of example later. After that, we were gonna look at automation rules later. You have a little bit of versioning if you if you want to add these to a specific repository, we also have the exchange if you want to submit it and the most imp.19:36Important one, the target. Each and every workflow can be.19:56Executed on a specific target or a specific group of targets. Every time we have an interaction like this, because workflow can interact with different systems, we have to tell workflow, hey, work on this Meraki organization, we work on these catalyst center.20:023rd deployment and also these are the credentials that you need to use to interact with that system. We can have the the ability to specify, oops.20:22On the start. So again, prompt the user, so if it's a generic for you can say, hey, user, you want to run this specific task, let me know which systems you want to target. Or you can target or you can have a specific target for this. In this case, you have.20:36The various systems that we can target, including.20:54I want to mention the HTTP endpoint, which is the generic one as long as the other system has the availability to interact through API, we can target almost anything, but we make your life easier if it's a system that we integrate nicely with. So in this case, e.g., I have meraki.20:58The point and you can see here that I have a few defined targets. Now we're gonna see targets in a minute, but the thing that I want to show you is that some are integrated, some are generated and maintained by the system. In this case for Meraki, we have the organization where the world.21:18The Flow was originally launched from. So this is a reference that you can always have. So if you're running these on organization A, this will reference organization A If you launch these on organization B, this will reference organization B automatically. So you don't have to remember, you don't have to paste in the organization.21:38ID all the time and you're also creating a little bit of an isolation. So you know that that workflow, if it's using the current organization, can only access the organization where it's launched from.21:58When we get into defining the block, here is pretty much the same as API, ok? Because in the end, the back end of this solution uses Meraki API to integrate with the system. What we provide is a nice way for you to define things.22:14Like, oh, how many organizations do you want to retrieve? I don't know, 500 at a time, which this is probably not the best. Let me get, get organization.22:32Networks. Here we go, probably a bit more interesting. You can see here drag and drop, we do have, hey, what product types. Now product types you can go in and type it. I wanna know wireless.22:47Or if you use the variable buttons here, you can import something. Now this variable could be, something that is an output from a previous activity. This could be something that is in the environment or this can be one of your input or output variables.23:04So in this case, I haven't defined any input variables, but if I go back in, I go to the start menu, and I get.23:24I don't know, I can ask, hey, hello? Okay, I can define a variable, it could be an input one and I can say, hey, network type.23:32And I can say, I don't know, it's a string or it could be an array. In this case I'm fine with the string and what I'm gonna pass to it is gonna be wireless. Here we go.23:51It's your definition, but once I've defined this, I can just save it, go back to the task that I was trying to perform and I can say product types, input, network type, and I can add them in. Now the definition here is expecting an array, I defined it as a string.24:06You've seen one of these interaction where the learning curve is there. So data types are important. Every time you see the square brackets, it means that the expected input is a list of something, even if it's a single element, it must be shaped as a list with a single element. If you see.24:26I don't know empty fields like do I have organization ID? So here I'm expecting some tests text. So again, it could be an input parameter where it's like I'm gonna give you the organization ID, you are gonna execute it on this organization ID.24:46And once you've done whatever you want, you can then define additional logic. So not only the activities allow you to retrieve information, if you're looking here, we do have the ability to set specific variables or extract.25:04Values from a JSON response. So again, e.g., JSON path query super used at least in my experience when you receive the list of all the organization and maybe you want just the organization ID, you can just.25:21Dump this and just grab the output and say, hey, this is the output from get organization network. It's the output message and from the output message I want to retrieve.25:37The network IDs. So it's gonna be actually gonna look at all the values and I want to retrieve all the IDs. Jason path query use a specific structure. It's not rocket science, it just assumes the structure.25:53After the JSON file and you can tell them at which depth it needs to look and then what type of data you want to retrieve, so I want a specific field, a series of fields. So again.26:12Here a little bit of a learning curve into JSON expert queries. Personally, when I find the thing a little bit too hard or the JSON file is pretty long, an AILM can help you define the exact query within seconds, so if it's too complex, go there or you can read the document.26:25About how jason path queries work and reference a specific field directly.26:45Other things before we move on, we have logics so it will not be a full environment if you don't have access to some logic. As logic, we do have the traditional for each Andy loop that will allow you to iterate over multiple elements. So maybe your collection.26:52In all the networks and you just want to, I don't know operate on the network with a specific network tag or with a specific product in it or with a specific serial number. So again, you can just.27:11Drag the loop in and then you can decide what's the source that I need to iterate over and then maybe I just want to do the JSON path query if the the of or for each and every value that I retrieved.27:25Same so we also have break and continue, so if you want to break out of the loop, you're gonna use a break. If you gonna want to continue the execution because you don't care about this specific item, you just hit continue and the the iteration will go.27:44To the next item. We also have some conditional blocks so you can have obviously you have your start block or your condition block that is pretty much an if block, ok? So you have your condition, you can define.28:00Why would I want to go down that branch? And this could be, hey, I made an API call, I want to check for the status. Is it a 200 ok? Do this. Is it an error? Do that. Okay, so if it's not 200 ok, then something went wrong, I need to perform some other action.28:16This could be used to validate the input, validate the output, so kind of having that level of flexibility when it comes down to understanding or defining different paths in your logic. Last but not least, you do have groups and parallel blocks. These are groups. It's just a nice way.28:36Group the information if you have a workflow that is growing quite a lot and you want to keep it a little bit contained and maybe you want to collapse it and not see it, definitely a thing you can do or a parallel block, sometimes you have actions that you can perform in parallel, so you don't want to wait for one.28:56To be sequential one after the other, but you want to just use them in parallel, you can use a parallel block. This was a little bit of an overview of the editor. From now on we're gonna work on or we're gonna see workflows that are a bit more advanced. I don't want to spend time here just around.29:14Being a single workflow, I want to leave you with some ideas of what you can do next after this webinar. So to do that, we need to actually run a workflow. So how do we run it? We can go back to the work workspace. Oh, one thing, I was actually checking it. There is an import.29:34Important concept here.29:54You can see what is wrong, so invalid action means you're missing something, the data format of the, the input variable you're putting in is incorrect or something is wrong, ok? The workflow will not work if you leave it as this, and you also have the option to validate it. Obviously in this case.29:55Because there are four invalid action, it will tell me, hey, 1234, you need to fix these four to validate the workflow. Once the workflow is validated, it's ready to run, which means it's good and we can either run it manually, dynamically, automatically through the AI assistant or through.30:16Any other trigger, but the most important thing is if if a workflow is in an invalid state, it will not run NO matter what. So if you go back to the workspace, we can see here every single workflow with the green green. With the positive check mark, without an X pretty much is, is gonna be good.30:35The red one will be red.30:55For this thing, what I wanted to show you, so I'm gonna start with the organization audit. So let me find my favorites because otherwise I will never find it. And I will start with a simple example. So.30:59In this case, I'm gonna view this workflow. You can see here it does look dreading, but if we go over, 1st and foremost, we have the description. So if we go in, it proceeds all network for an organization using pagination and compile the active SSID.31:16Nope, that's not true. And compile organization settings probably it was taken over. You can copy and clone workflow that you've already created. You don't have to start from scratch.31:34Just remember to update the description. You can see here that this is valid if I want I can run it directly from here, but 1st let's have a look at what it does. You go in, we, because we do have this concept of pagination, we have it in within API, we have it within workflow. If.31:54If you're over anything on more than 1000 element, you will need to have some pagination logic. So if you want to execute something over every single network and your organization has more than 1000 network, it's important to remember pagination will be needed here.32:15It's also a nice way to break down, so if you don't want to operate on a thousand results at a time, but there is a guide on how to implement pagination is just a way or a thing to keep in mind, ok? If the element scales above a certain limit, you will need to think about how.32:35How do I access those? Because the API call can retrieve a maximum of 1000 per iteration. After that, I'm gonna.32:54Get all the organization. Hopefully it's gonna be less than 1000. In this case I'm gonna limit that to 500 and after retrieving each and every organization, I'm gonna do a few updates to make sure that the pagination token is updated for the next 500 organization.33:06And I'm just gonna read all the data. In this case you can see here it's a JSON path. I'm gonna read from the top level and I care about the organization ID, the organization name, and the status of the API for that organization. After that, because I care about the security setting, I will go in and for each.33:24Organization.33:44I will fetch the same SAML status, the login security, so if I'm using a strong password, if I'm using password change, if I'm using all those type of things. And then, and this is probably my favorite thing, you can use Python directly inside. So what.33:46One of the block that you have available, if you search for Python.34:06You can execute a Python script. So if there isn't a block that does exactly what you want or if you start to find yourself in a situation where it's like, oh, there's too many parameters I don't want to drag and drop everything. I could do this with a simple python code. You can embed Python code directly in.34:10So in this case, we have this python script that takes six elements as an input, in this case the the oops, the table and all the data we have fetched before, and it just accumulates them in a variable that I can use later. Once we actually we can enlarge it, this.34:30Could be any Python script as long as it doesn't include any additional library. So we can reference any import to what is available in basic python. You almost have everything you need. Sometimes it's gonna be a matter of tweaking something to be able to do it with a traditional library if you're using.34:50Just to use external libraries, but we have it available.35:10And after that, we can define what we want as an output variable. So in this case I've done all my calculation and then I've stored the result in an output variable, which I can retrieve. So I wanna get the result.35:14Variable and I want to put it in a variable I can use later. That being said, I update some variables and I iterate over each and every organization that I have. And after that I use Python again. Why? Because I want to have.35:30A final table that is nice to read. And in this case, we're gonna show another. This one has the NO email thing, the original workflow had the ability to send an email out. So in this case, we can finalize the output, use Python to generate a nice HTML code that then.35:47I can embed into an email and send it to a user when it's done. So 1st and foremost, let's run these without anything else. You can see here we don't need an email address because it's not gonna send it anywhere, but this is an example of an input variable that I need in order for the workflow to run.36:07And I can run it right away. The beauty of workflow is that when it's running, you can go in and see it running live. So you can have an idea of each and every iteration. If it's blue, it's still looping, if it's green, it means that in the loop everything went well. If something is red.36:24There's a failure, and when there is a failure, you can go in and check the exact moment where the failure happened. So it could be, hey, I got an input that doesn't seems to match or I got NO input or maybe you were iterating over something and you put the wrong condition in it. So it really depends.36:44Once you have it here, you can check each and every iteration, so you can see here that for the pagination loop, we only had one.37:04Iteration while for the organization we had 12345. And if you want to go in, you can check each and every iteration just to make sure that everything is working as expected. Also, and this allow me should allow me to go back to.37:12Run monitoring.37:31Maybe you're not running it manually. Maybe you're running it on a time schedule, maybe you're running it through a webhook. There is a menu under automation run monitoring that will show you everything that was run in the past or it was canceled or that errored out or that is waiting for an input from a user.37:33So anything that is happening is available here and you can filter it down. In this case, if I want, I can go back and look at that specific iteration and see why it was a success, what type of data I got out, what was the result of this workflow? As I said.37:53This is a very long HTML table, not the best way to visualize it, but we're gonna see it in a minute in a different shape, just to show you what can be done. But again run monitoring, very useful. In this case because we are iterating over.38:12A group of organization. We said before we have targets, so let me go back to the workspace. For the sake of the demo, I'm gonna run the audit.38:31Okay, this is the one.38:51I wanted to show you it was not starred. That's why I couldn't find it. Exactly the same workflow, the only difference is that at the end here.39:11We're just gonna send an email out with all the data. So I'm gonna run this one just give me a second. Oh NO, that's not what I wanted to do. Let me share my screen again and now everything is gone. So just bear with.39:19With me for a second I need to bring in the chat, so if you.39:39You have any question, I can see lives and the Q and A panel. Here we go. Okay, what I want to do was this.39:44Starting I put in an email address. Hopefully it's gonna do exactly the same task and in a few seconds I will have this in my mailbox There we go. So I'm gonna wait for my mailbox to update.40:07Do I have it? Received an email. Now it's hard to read, but you get the idea, ok? So this is a table that I can use copy paste and reuse wherever I want. So very nice and easy way to export the information. We're gonna see in another couple of demo.40:29That's how we can do it for other systems. Now this, because we were talking about organization level permission, if you remember, we had targets. So in this case, because I'm operating across multiple organizations.40:49I need to target or I need to have access to all of them together. How do I define this target? These targets are defined under automation targets. These can be stored in the system, so you input them once and then you reuse them as many times as you need. So you don't have to.41:04And Remember things, they're gonna be stored securely in here.41:24Hopefully I have it. In this case I have some targets that are managed by myself. E.g., if I want to post to Webex, if I want to post or to use my AI assistant or if I want to post into a 3rd party system, and then there are some that are Cisco managed. So in this case.41:30At the moment, the current organization. Bear in mind that these, all the ones that are owned by you are or must be defined and maintained. The one that are Cisco managed, you have to do nothing. So if something will change in the future, we will just update this target for you. In terms of defining.41:49New target, it's the least I've I've shown you before and then you have to define a few things based on the type of target. If we're going for a Meraki target.42:09Let me see Meraki endpoint, you give it the name, you need to have some account keys. This is the one that I defined, I can define a new one and it's gonna be a display name and the API key that the workflow will use in order to run.42:22I'm gonna cancel it out, few information if the specific destination require proxy to be reached, you can do it via proxy and so on. So again each and every target has a different configuration. I will not go over all of them, but you can see here for sending an email, I use an SMPP endpoint in order to be able.42:37To use a mail server to send the email. So again.42:57Based on the task you're trying to achieve, you might need an additional target. If you're only working within Meraki, the Meraki endpoint is the only thing you need. Just for parallel here, I do have a, another.43:02Here we go. Example that use the fetch network SSID. In this case, what we were trying to do is I have a hundred network, a thousand network, whatever it is, I want to see the SSID configuration. What can I do? Jump inside and here is a bit easier for.43:21From a target perspective because I want to run it on the organization where I am. So again, if I check.43:41The the start and I check the targets. In this case, I'm gonna use the Meraki endpoint. Actually I'm gonna specify it on workflow start so I can say, hey, I wanna run this and I can say, hey, run on my current organization. I can do a bit more. Let me.43:48Duplicate this.44:08And this is required from the workflow I can paste in my organization ID and hopefully, it will run and again pretty much the same story. It goes through, you get the wireless site.44:14Maybe.44:34It searched for what I care. So in this case in the final output, maybe I just want to have the list of the SID name and the authentication mode. Couple of clicks away. If you want to see it, it will be in the run monitoring and it will be available for you to review at a later stage.44:35Now, we've seen how we can trigger them manually, and you will be thinking Sandra, that's great, but maybe these are four tasks that are repetitive for me and maybe compliance is a thing that I don't want to go in and trigger manually every time. I agree with you. That's why workflows can.44:58Can be triggered by rules. So if we're gonna see a couple in action, but 1st I want to introduce you to the rules. We have rules that allow you to define workflow based on either a schedule. So in this case, I have one that is running every hour, every day from.45:18Twelve to 15 in the afternoon. It's as easy as defining the name.45:38Description, turn it on or off. So if you're going into maintenance mode, you can just flip it off. And then you can define your calendar. We have a few. It could be work days or it could be over the weekend. It really depends what's the task you're trying to accomplish. Maybe, hey, every Saturday night.45:45When the company is closed I wanna run an audit on everything because I'm sure nobody's gonna be fiddling with the configuration Saturday night. Great. I'm gonna run it and Monday morning you will have it in your mailbox and you will be ready to check it or it will be just a needle saying, hey, everything based on your check of compliance, everything is compliant.46:05Start your week in the best way possible, there's nothing you have to worry about it. And last but not least, you can apply a specific workflow. Now here.46:25I have one, I can add another one when I select one. If there is any input variable that is required, I will define it directly in the rule. So I have my automation rule and I can define, hey, this will run every.46:35Weekend and send the email to my mailbox, whatever it is, ok? Another type of automation that we can have, if we go through rules, there is a second category and the category is called webhook. Now, if you are familiar with webhook, NO surprise here. If you're not, this is a crash course on.46:50Webhook. You have an HTTPS receiver and that HTTPS receiver is listening for events for somebody to do an HTTP post onto that URL. Traditionally, you will have to maintain that type of service. With workflow, we can have.47:10Bit embedded directly in. So in this case I'm being reactive. I'm waiting for something to happen. And in this case how does it work? I have a web crew.47:30I can jump in, I can define a webhook. This is already defined, but if I want I can add a new one. Again, it's as easy as giving it a display name, the type of content I'm expecting back, and if there is, and, and it will generate for you a webhook API key and a.47:42Web URL that you can use. In this case, I've already done it.48:02And after that I can have some conditions so I can read the content. Usually webhooks are bringing with them some payload. In this case, because I want to trigger this when a device become unreachable in dashboard, I want to look at the body that is coming from that specific.48:06Webhook and I'm looking for other type unreachable and category type connectivity. How did I do that? Well mainly through testing. I've seen, I've launched this webhook 1st because I have the ability to view the content every time I run a workflow, I attach it to a workflow and I.48:26Copied and pasted what was the content of the request. Here again, you can define a few things, including, hey, what are the input variables, what are the workflow attached to it? And you can also turn them on or off based on your needs. So if you're testing something, you can have it on, off or you can have a specific web.48:46For that. How did I use it in my environment? I went under orientation.49:06Alerts and under configure alerts, I did create an alert profile, and if I go in, I can edit this and I define, hey, it's a workflow alert. It's for unreachable device so it will trigger only on.49:12Unreachable device. These are the network where I want it to work and also this is the receiver. The thing we configured in the previous menu.49:32You configured the new one, you created a new one, you can just add a new one. It's just a matter of the name, the web book, the shared secret, and the payload template that you want to use. This will change the way the information is formatted when it's coming through. If you don't like what we provide, there is a way to configure.49:41Web payload the way you want. So again, if you want to fiddle with it, you definitely can. So I know I I have still a lot to show you how this worked. Well, let me go over and.50:01I'm gonna give you a taste of what a webhook can do for you. In this specific case, I'm gonna go to my access switch. I have a port where I have an access point connected. I'm gonna go in and I'm gonna make a mistake. Hey, let's go in and disable power over internet.50:20I can update this port, obviously it's gonna take some time. So I kicked in the process now. We're gonna see the result in a few minutes, but just to show you what's happening under the hood. If I go under automation, workspace.50:38Let me double check. Oh I didn't save. That's bad on me. I could have waited forever for this to happen? So let me edit this board.50:55Yeah, ok. Did I do it on the correct switch? That's a good question.51:10Hopefully yeah ok ok this will trigger the demo. That's the beauty of live demos, ok? Everything we're showing today is live, so it's it's something that you can trust that is doable definitely because it's a live environment, we're doing it live. So.51:31What we were talking about. We're talking about triggered webhook, ok? So I'm gonna show you two versions of this as well. The 1st one is without any output, ok? This is the easy one. And the easy one, what it does, if I go in, I zoom in a little. Now this can be run manually.51:49But it has NO purpose because it expects as an input the payload from the web book I want to know what's the device that is unreachable when it went online. So again.52:08It's a workflow that I cannot run manually is designed to be run on a triggered action. In this case, a webhook receiver for a device unreachable. We extract the information from the payload, we set some variables like the serial number of the device, the time when, what do we.52:19The organization ID, the organization name, the network ID, the network name, the serial number of the device, the device name, the device MAC address, how bad is the alert, what's the alert type, what's the ID and when the issue started? These are all things that we are getting from.52:41The, the payload of the web. After that we do some calculation for time differences because maybe actually we're gonna see it in a minute it will check for status of other things in the network.52:585 min before, 10 min before. These are parameters that are defined. So if you go here, you look at the input variable, there is, how much, how many seconds should I look back for configuration changes? How many seconds should I check for device availability? So I get.53:14A little bit of information about the environment, but not everything that is available historically. I just get something that most likely is relevant to the event. After that, I do have a huge parallel block here where I start doing a specific series of things. And this series of thing is.53:34I want to get every single configuration change in the organization around that time. I want to get the CDP and LLDP information for that device. I want to get the availability change history for that device. I want to get the network events.53:52For in this case is an access point for the wireless environment in that network and I wanna see if there was any firmware update scheduled for the time because maybe I scheduled an update and the device has just rebooted so it's not an issue. After that, I can grab all the information.54:08And once again Python is our helper here because it can help us formatting the information beautifully. We can feed everything into it and then we can set some variables. Now, this is one way to over it. Another thing that you can do is based on the information that I collected, I want to do some.54:27Something. I wanna try and cycle the port. I want to send an email to a specific user because maybe I have an admin list for that specific network. I want to trigger another workflow or another webhook. I want to open a case for ServiceNow.54:48You name it, ok? I'm showing you the easiest of the example here, but what you do after with all this information is up to you. Another, use case here is if you go down this road, I have done something different. So.55:04This is device status unreachable device. Let me view the webhook here or sorry, the workflow. The workflow is exactly the same, so it doesn't change anything, but what I've decided to do because I have access via API to a large language model, before composing the message, I'm gonna send.55:22And all the data that I collected on the previous step.55:42To the AI assistant. And the AI assistant has one task in it. It's like you can, I don't know, where is there are some parameters but it's like, hey, who are you? You're a network engineer, you have to fix the issue. This is the information that you have. Include the symptoms, give me and give me an.55:46Hypothesis on what went on there. This is the data that are available. Figure out something and return me a 200 words summary of what was going on.56:06After that, I'm gonna do exactly the same thing. I'm gonna compose hopefully an an output message and I'm gonna both post it into a Webex room and I'm gonna send an email with the same content. So hopefully.56:191203, ok. There are a few new messages. Okay, so in the background I you remember I turned off the power over ethernet on a port.56:37And this is my Webex message that I received from the workflow. This is the device name, this is the network, this is the time when it happened. You can view the device directly there. This is the summary, became unreachable around this time, ISO time following a configuration change that disabled power over internet.56:53On the switch. This is the switch, this is the port.57:13LLDP confirms that the AFP switch history shows online offline around the same time, around the outage. Likelihoods power overs to the access point is why the access point is not working. Do you trust the AI analysis? Good. You don't, hey!57:17Because you formatted this message, this is what happened in terms of availability, this is the configuration change in the timeframe. These are the CDP and LLDP information. These are the recent event on the device. I designed that. I decided that this is what is important for me.57:35And I can just go back in, go to the switching device, I have the information, maybe I could be even smarter and say, hey, if it's a power overages, just go back and change the configuration on the port, ok? This really depends on the degree of.57:54Information or type of actions that you want to take. So in this case I can just go in enable it and update it again. Actually, let me go through the all the workflow in one place, monitoring.58:11Progress for active run, we're gonna see these in actually in a minute because obviously you might have a run that lasts for a few seconds and that's absolutely not a problem or you might have run that requires a little bit of user interaction. So one of the things we saw before is hey.58:31And maybe I want to ask the user, ok? So I have again two nice example for you because there is an additional capabilities within workflow. It's under automation and it's called users tasks. User task is the.58:51Ability to stop the workflow at a certain point and either wait for an input from a specific user or group of users.59:11Or, a wait for approval. The workflow that I showed you before, hey, this is the full analysis of everything. Either find that the problem is power over internet change, I would re enable it on the port. Do you want to do it?59:21So instead of having just the agent doing stuff, just the, the, the software taking decision is I'm gonna stop. I'm gonna provide you with a possible solution, but I'm gonna wait for a human input to confirm that everything is in line with my analysis and then you can just say yes, approve or NO.59:37So disregard this change and move on.59:57How do we import those and what they can be used for? Two, use cases here. 1st one, let me go to the workplace, let me grab the six water.01:00:01Okay, we go in and let me look at my favorite again because otherwise I will never find it and reboot device with approval, ok? Let's jump straight into this very quickly. In this case, I'm not gonna ask anything to your user when I see.01:00:22Start the workflow, I'm just gonna start querying the user after. And the beautiful thing is that workflow has a conscience of who run the workflow. So we can target the user that made the request and we can say, hey, I need additional information from you. Yes, you started this workflow, but there is something more.01:00:42You'd need to give me. In this case, it's just going through asking for a tag for the device that is fetching all the devices with that tag and is just returning to the user an approval message saying, hey, based on the tag you gave me, if you want to reboot all the devices with that tag.01:01:03This is the list of devices that I'm gonna reboot.01:01:22Do you want to do it? Let's see the next question very briefly. I can run it here, fetching some information and then it's creating a prompt. You can see here that because I'm logged in, it's telling me, hey, there's a workflow with a pending task approval, select the view task.01:01:25Button. I'm gonna view the task and I'm gonna say, hey, select the device tasks to reboot. I'm gonna have this list. I have a dedicated one that is not gonna make any mess in my environment.01:01:45And then I can I mean I can multi select if I want more than one. I'm gonna save and we're gonna see the workflow resuming exactly from the same point. So we created a prompt, we gathered all the serial number that we want to reboot, but we don't want to reboot straight away.01:01:57I have another task and it's like, hey, this request approval is associated with this serial number with this AP. Do you want to approve it? Yes. Do you want to reject it? No. I'm gonna reject. I'm gonna save. The workflow will continue and it will just close successfully without a reboot.01:02:16Now, this prompt can be very powerful. I'm gonna show you a different example here and and this is kind of ties in for those of you that were here in the previous session, we created a basic network deployment with.01:02:37One MX and one access point using API. Again, can you do it with workflow? Absolutely yes. One way to do it is I'm gonna run a workflow. I'm gonna input the serial number of what I want and I'm gonna let it go. Another possibility is I can build a workflow.01:02:52Network creation with prompt and approvers. And in this case what we're doing is oops ok wow. For a moment I thought it was not there.01:03:09We're gonna retrieve the organization, we're gonna get the inventory of the available device, we're gonna provide it to the user and the user will be able to select. In this case, we have, where is it?01:03:28Prompt user, we can prompt the user that ran the workflow, so it's like the only user that will receive it will be that one. And then we have a request approval that actually, let me see. Here we go. I want a minimum number of approvers. It cannot be an.01:03:42Action that a single person will be able to figure because maybe it's something disruptive so I can build in some kind of safety mechanism. So in this case, I'm gonna ask for minimum two number of approvals. I'm gonna run this one once again.01:04:01Hopefully I'm gonna receive a message. I'm gonna view the task. I'm gonna be the name EIS as network I'm gonna lease my security appliance. I want to put six seven access point, I know I have only one and I'm gonna save. This process will continue with request for.01:04:20Approval I'm gonna view the task.01:04:40I'm gonna, hey, this is the list you can write whatever you want in it. I'm gonna approve it and save. And you will see the task is going nowhere. Why? Because I've set a minimum approval list of two. Luckily I do have another identity in dashboard and if I refresh, I actually received an email that said.01:04:42Says hey, there is an approval task that is waiting for you so please go in and somebody do something. You can see here that I have a pending task for me.01:05:02It's a new network creation, I can review it and I can approve it being 2nd approver for the task. Now that the task has been approved by two user, hopefully, here we go. It resumes.01:05:11It continued and it created the network. Just to prove the point, let me refresh the page here.01:05:32And we can go here and see that we have the EIS test network that is now being created. So again, you can use these to standardize the process to have check in place to define how things should happen in your dashboard.01:05:39But this is not the only way. Now, we do have another couple of more slides which cover the AI assistant. So 1st and foremost.01:05:59You can use the AI assistant to recall a workflow. If you have something that you know it's there, you can, hey, is there a workflow for this specific task? The assistant will search, will show you the option. That's why the description and the name are fairly important, and you can run a workflow directly.01:06:08Directly directly from the AI assistant.01:06:28Now, let's try a new conversation and we're gonna go with, hey, I want to perform an organization audit, any workflow available for the. It's gonna search for some of the information.01:06:31Hey, I found one, two, and three. I'm gonna do the 1st one. So please run the 1st one. It's gonna read the request and it's gonna say, hey, these are the list of tasks. So this is what will happen. It will finalize the output table, it will generate, it will set the final output. That's the idea of what's gonna happen in the workflow, do you want to confirm?01:06:55To proceed with that? Yes.01:07:15And this case is like, hey, I do require some input, I need the email address to continue, so I can type in the email address, I can click continue. Would you like to proceed? Yes. As you can see, we're very thorough to not let everything run randomly. You can see the workflow.01:07:18Running live here.01:07:38And you will be able to see the result. As always, it's not gonna be the nicest thing because it's an HTML formatted table that was designed to be sent over via email, but you get the idea. If you want, you can download this, you can format it the way you prefer.01:07:40This is just an adaptation of the traditional output that you will get out of the workflow. This is one way you can leverage the assistant, but there's also a second way, and this was released not too long ago, it's the ability to ask the assistant to generate workflows for you.01:07:59So I'm gonna have a new chat.01:08:19And here, I'm gonna go with a different prompt that is please create a workflow that lists all the wireless devices in the organization based on the organization ID. This is a bit of a short prompt, I have to be honest with you, but it's mainly for demo purposes. I have a slide with guidelines on how to better write prompt.01:08:21For the AI assistant, but it's just to show it to you in action and for the sake of time, we need something that will give us an output in a reasonable amount of time to be able to do it live.01:08:41It will fetch the information, it has the information of what's available in workflow. It will propose you an idea of what it's trying to do. After that you will be able to tweak it so you can continue the conversation say hey maybe do this thing in a different way. And when you're happy with the.01:08:54Result you will be able to just ask the assistant to generate this workflow for you.01:09:13I wanted at least the 1st step to be here. Here we go. Test the plan. We're gonna input the organization ID. We're gonna output the device wireless list and step by step, we're gonna get the organization detail, we're gonna get the list of wireless device, we're gonna format and present the result. Are you happy with this?01:09:21Generate a workload Jason for me. Yes, it will take a little bit more, and hopefully.01:09:41I do have a backup one by the way, but I wanted to, I know it's a little bit of time to wait, but I think it's most definitely an interesting team thing to see live, personal experience here. Sometimes it might need to adjust.01:10:01Don't go over, over the top, ok? Try to have fairly small tasks. Remember that once you created a workflow, you can have two workflows interact together so you can build a workflows.01:10:19A workflow that contain more than one workflow. So if you have a stepped approach with hard tasks or that is trying to accomplish something very hard, you can go task by task and get it generated here. I can just click import workflow here.01:10:34And we can go through and it's there. I can click on it, I can probably close the AI assistant and you have it, it's getting the organization is listening to what device is storing the results, it's formatting the results through python.01:10:51It's also handling errors for you which is not bad. And if I want I can run it again target, I want my current organization, I want an organization ID, and now we're gonna see if it works or not. Fingers crossed. It does.01:11:12Here we go. So as easy as this, you can start leveraging workflow today. It can also be used as a canvas. You don't want to start from scratch. I used it a lot to, I have an idea. Give me, you know, at least you already have the blocks. Maybe the intern.01:11:30Actually is not quite what you want. You want to tweak it later, but you don't have to start from a blank page and this is in my opinion.01:11:50Amazing. So bear that in mind, it's it's amazing that you can start from here. There is currently a limitation. I don't know if it will change in the future. So if you're looking at the video recording in the future, please double check the documentation, but for today, the limit is ten workflow creation per day.01:11:58So again, don't go crazy right away, but you do have a budget there and you can start playing with it right away, which was my next slide in term of a good prompt.01:12:18Okay, because what I showed you was a very short one, but what you want to put together is what's the intent? What's the outcome that you want to get? What are the input that you want to have or that are mandatory for you to have for that workflow to work? What are the required outputs and into context aware, do you want.01:12:31Want to use specific workflows that you already have. Do you want to use specific section where it's like, oh, do this with Python if you can. And if there are any technical constraints. Bear in mind that the limit of the input in the chat, I think is a thousand characters. So sometimes you might be in a situation where you do.01:12:51The rough definition and then on successive inputs you can modify the plan that was proposed to you to adjust it to what you need. How did I learn that? The 1st time I wrote the full prompt on a text file and I went and copied over and was way above like I put a lot of spaces in it, a lot of new lines.01:13:11And I had to tune it down a little so bear that in mind if you're planning to go down this way. In terms of the best practices, so there are a few best practices to keep in mind and a few limitation, same.01:13:31Information as before applies, so I'm gonna post them in the chat both because I think they're useful. We have the best practices here and we do have the notes and limits. So once again, a couple of useful link for you in the chat. I'm not gonna go over all of these.01:13:48But the thing that I want you to keep in mind is.01:14:08It will not run forever, ok? There are some size maximums of specific variables. So be bear that in mind. It's not infinite space that you have. You also have a maximum run run time of 30 min, so if the workflow runs.01:14:12For more than 30 min it will run out, it needs to free up resources.01:14:32And, and this is probably the most important one, the, where is the loop limit? Okay, loop limit four and a while loop can iterate for a maximum of 500 times. Now you will be thinking, ok, when do I hit this limit?01:14:37The example is the one that I showed you before. I'm gonna go over all the organizations, all the network. If you have a large organization, you might have more than 500 network to check.01:14:57Where or how can you handle this? Well, you can leverage pagination as I showed you before, and instead of setting the limit to 1000, which is the maximum, you can set the limit to 500, iterate over 500 results at a time and then move to the next.01:15:09Next page for another 500. I've done that in an organization with a thousand 200 networks. It works pretty well. 500 at a time, I'm just checking the next page, the next page, the next page, and then for each and every network I'm gonna perform action A action B action C.01:15:29But I can rest assured that those operation will not exceed the 500 limit because I don't have more than 500 results coming back from my request. In terms of the best practices naming, if you want to publish, make sure that is done well, predictable.01:15:47So make sure you handle corner cases. What happen if I ask you for a serial number and there is NO serial number available? What happen if the device is unreachable? It's also very important to think about even potency every time you create something automatically. It's also good to have.01:16:07The reverse because what happened if I create 200 networks by mistake? Well, it will be nice to have a workflow that does exactly the opposite and clean up if I made a mess and categorize and organize everything you can and name variables in an understandable way. I know you have input variables, output variables.01:16:27Local variables. I do like to put the suffix on at the beginning to make sure that I always remember what it is, where it is, and how it's called and what I'm acting upon.01:16:47Yeah. But I'll leave it to you to read because definitely in the interest of time, we cannot stay here forever. There's, I mean, there is a demo. We've been doing demos for most of this session, which this was a simple one with cloning an existing organization.01:17:00So again, it's the same story. Define what you want, start maybe with the AI assistant or the exchange where you can go and check, hey, what can I do? Is there anything that I can use that already exists or should I just move on and create it from scratch?01:17:19And you can start playing from there and define your constraints. In terms of the access, it's included, ok? It's there. There's NO reason to adopt and deploy. The the adapters for Cisco are maintained by us. It could be used as a local.01:17:40Information builder. But if you already have in place an infrastructure, think of how workflow can standardize the way you access API. I had plenty of conversation with customers, they're like, oh, I'm already using ansible. I'm already using Terraform.01:17:59Yes, but who can access that? How do you make it available to the rest of your organization? And workflow can be that building block of standardizing the entry point through the services you already have. The other thing is when you use workflow.01:18:17Or how can you use well what what can be done in term of standardizing the access is you've seen how we could leverage the AI assistant to do specific tasks, but AI has this problem of.01:18:36Maybe changing the results, ok? It's not always the same answer. So to have something that is predictable, what you can do is AI assistant 1st to do an analysis and then ask, do you have a structured task I can run? Because as a network admin, I know that a workflow will always run.01:18:54The same way, the way it is defined. It will not change his mind halfway through or provide me two different ways of interacting. So it's nice to have the.01:19:14I don't know, the starting point, sorry, I lost track of it. Yeah, the starting point being an eye conversation and then move on into something that is structured by default.01:19:29Also, this is probably there's plenty of resources. I've shared with you the documentation. The other one that I want to share with you is a lab. There is a Cisco live lab that is publicly available that you can run because workflow is free for everybody. This will give you the basics of creating a.01:19:46Variable as we spoke about passionation, we spoke about Python, all the things we discussed about in this training are covered in the lab. Obviously, it's not taken hour and a half, ok? It's it's designed for you to guide you step by step.01:20:05And get you to the point where we can confidently use the solution.01:20:25There's a question coming through and it's about if there's an ETA to have native integration from Meraki events to be used as trigger for workflows without the overhead of using webhooks? Not that I've heard about, but keep in mind that despite workflow living inside the Meraki dashboard.01:20:35Word we want to keep it open to everything else. So will there be a shortcut for Meraki to get into workflow without webhooks?01:20:53I'm not a hundred percent sure. I haven't heard anything, but for integration with any system that is not in the Meraki dashboard, webhooks will still be the main trigger. So at the moment, I, I haven't heard anything because it's in the same platform, maybe there is some evolution there. Remember.01:21:02Workflow exactly the same way as API. Every month we release something new. So new API called new integration, new atomics that you can use in your building blocks, new functionalities, smarter AI to build the workflow, so it's an.01:21:22Ever evolving landscape, I would say as I said at the beginning, start small because that's definitely the best place to start and if you want to continue your learning, actually, you know what, probably more depending on what you prefer. I think most of the example or more.01:21:41Most of the thing are coming here from videos so we also have a video library, definitely use that. We didn't have time to cover it in the session, but it's important for you to remember that workflow can also be exported and imported. So everything is in a.01:22:01Json format, so if you want, you can just copy paste it and import it somewhere else or you can integrate with Git. So if you want to have a repository somewhere, actually, that's the point that I was about to miss.01:22:21And here it is. All the example that we use for today's session are in here. If you want, you can navigate to it, you can import it manually or you can import it via gate. It's a little bit sanitized, so not every single thing that I've shown you is there, but as you.01:22:39Usually just a matter of I need to have email dropping the block that does email and you should be good to go.01:22:59This, kind of give us the end of the session. If you have any additional questions, please post them in the chat. I'm gonna stop the recording now.01:23:07

About this Video

Author : Cisco Customer Success

Uploaded : May 22, 2026

NetworkingMeraki
AI summary icon

AI summary

Explore Cisco Meraki Workflow to consolidate network automation across dashboards and third‑party systems. The recording illustrates how to access the Automation menu, use the drag‑and‑drop editor, define targets and variables, apply loops and pagination, embed Python, and integrate HTTP endpoints for third‑party tools. See demos of scheduled and webhook‑triggered workflows for device‑down alerts, multi‑organization audits, SSID reporting, and change tracking. Learn to publish and reuse from the Exchange, monitor runs, and enforce consistency with User Tasks and multi‑approver controls. The session also showcases the AI Assistant to find, run, and generate workflows directly in the Cisco Meraki Dashboard—no extra infrastructure required.

  • Increase operational safety by standardizing changes with approvals, input prompts, and run monitoring.
  • Accelerate cross‑domain automation using built‑in integrations and HTTP endpoints for Cisco and third‑party systems.
  • Build and reuse faster with the Exchange and AI Assistant to generate templates and import or share workflows.