Transcript
So today's topic is Mastering the AI Technician. AI is quite a hot topic in 2025 and 2024, and probably will be the same in 2026. This is probably exactly why you're here, you're interested in the technology, you want to know more. Maybe it's the first time that you're trying to understand what it does, what benefits it has for your organization, for your employees, and maybe for your customers. But let's start by introducing myself. My name is Dima, I'm the Technical Product Marketing Manager at Eterra. I've been working here, I've been working at Eterra for more than two years, and I've been in professional services for quite some time, and I was responsible for the most complex Eterra implementations. So I also have hands-on experience about the product, and this is why I'm probably your go-to person whenever you reach out to us on community, user voice, and many other resources where we talk to you directly. But enough about me, quick housekeeping rules. So the recording will be sent to you right after we finish. As soon as it's in the Cloud and we're able to send it to you, it will be done. The Q&A section, as I mentioned, you're encouraged to submit your questions. At the end of the webinar, we'll open a quick survey and we would love to see your responses there, and would appreciate any feedback that you see fit. Without further ado, today's agenda. So introduction, we just made it through. IoT Autopilot, what is it exactly? And because I still get many, many questions about our two AI products that we have, I will talk about this in a second. What is the difference? Who is the target audience for each and one of those two products? We'll cover the differences. For all those people who are for the first time here, I'll do this intro for you. Sorry guys who attended the part two and they attended the part one, but we'll try to switch to the demo as soon as possible. The demo, we'll see the real-time demo of the product, what end-users will experience, what they can expect from the system, and how does it work in their day-to-day life, what use cases you might encounter, and also a couple of things worth mentioning about what makes IoT Autopilot unique. Okay, first things first, AI offering from Atera. We have two products that have AI functionality built in. One is AI Co-Pilot and one is IoT Autopilot. They have slightly different naming, but the main thing that makes them different is the core audience. AI Co-Pilot is designed to augment IT technician performance. What does it mean? It allows you to do diagnostics on devices, summarize IT tickets that are coming in from your end-users, and it's everything that is technician-facing. Everything that helps you do your day-to-day job smarter, faster, just do it better, and spend less time for things like, I don't know, typing even. But beyond that, there are many cool features under the hood, which we'll cover further down the line. IoT Autopilot, what's the main difference? IoT Autopilot is designed to work hand-in-hand with Atera Agent. For the ones of you who are just starting to understand the nature of AI products, agentic AI, and all those insanely popular words in any public forums, LinkedIn and many others, the IoT Autopilot works hand-in-hand with Atera Core product, so the RMM system. And what do I mean by that? It uses the telemetry which is sent by the Atera Agent to allow us to use AI to diagnose issues, install software, and do something, not just to be like a very kind and polite chatbot. And this kind of answers to what makes it different. And so I do this slide on purpose because session after session, we get the very same questions because, hey, I'm confused. Your website says Copilot, says Autopilot. What is the difference? I hope that answers your question. Okay, so what is the benefit for the Autopilot specifically? So in a nutshell, it allows each and every employee of your organization to have 24-7 borderline instant access to an IT technician whose job is to cover tier one tickets before they reach out to the IT team. It's not about gatekeeping by any means. It's just to help people resolve IT issues right on the spot without any type of waiting time, time zone differences. And in many IT departments, in comparison to MSPs, the night shifts are not that frequent and widespread. With that being said, having someone who is 24-7 at his desk, helping people to resolve their issues, which can be a simple reset, software installation, any type of diagnostics, some local actions on the device, helping people to fix things which are happening in the cloud. By cloud, I mean anything related to resetting password for Azure, maybe it's Okta, maybe it's giving the access to a network drive, anything of that nature. And this is what helps people who adopt this technology today, the customers that we're working with, to resolve a decent amount of tier one tickets. And to show you what it actually means to have it and how it works, we need to start with the foundations. By foundations, I mean, where does IT Autopilot start? Before we show the use cases, we need to cover where it begins, where does it get the knowledge from, how is it trained, and what is it capable of? Okay, let me switch my screens to the platform. Okay. Okay, great. So, IT Autopilot, as I already said, works hand-in-hand with a TerraCore solution. And what is really important is that we start with the right foundation. What do I mean by the right foundation? On the screen, you'll see the basic device, and it has a contact attached to it for the ones who are seeing a Terra for the first time. This is how each and every device page looks like inside the product. And basically, this is what allows IT Autopilot to understand who is talking to it and help not just by giving the general IT knowledge, but also giving the device-related knowledge and factoring all this data in when responding to questions. I mean, all the hardware information, all the OS and security information, and everything in between. Okay, let me actually switch to full screen instead, because I think we will benefit from it. Okay. Now you see my full screen. Amazing. So, two minutes about the setup. So, the setup begins in the AI Center. So, general settings, IT Autopilot allows you to integrate with your existing help desk ecosystem, which can be, Terra itself can be a third-party ticketing system. However, the overall idea is to work hand-in-hand with all the elements of a Terra, so it gives you the best experience possible. It starts with a basic setup where you customize the name to make it related to your company. You can integrate it with your company messengers, and then the configuration starts, which takes 10 to 15 minutes to configure. As you can see, there's a bunch of toggles which allow it to create the tickets just through Autopilot to prevent users from creating the tickets the old way, create the tickets in parallel with the chats with AI, and assign IT Autopilot to tickets which are created to email. So, basically, these three toggles are responsible for meeting the customers, meeting the end users where they are, because it allows four streams for the tickets and more to come to submit the help desk tickets and get the IT assistance 24-7. So, automated follow-up is something new that has been added. It's like for how long IT Autopilot should wait until it escalates it to the IT technician or sends a nudge to the end user and wait for the next action based on the way the end user interacts with the solution. Certain self-service actions. So, this is something which I will show you when we'll reach the end user's device. Wide-listed software. This is the list of software which is pre-approved by the IT team which end users can install. We'll get to this in a second. One of the most recent features is that in many companies, there are people who require VIP treatment and they always need a live person's assistance. And this is why we added the excluded addresses. But it's not ending there. As soon as you activate it, you can start using it. However, IT Autopilot is trained on the general IT knowledge. So, it will help you as just general consultancy. We need to make sure that several things and several checkboxes are actually checked. What do I mean by that? If we go to the, sorry, to custom instructions. So, custom instructions is where you meet AI and you tell what you do as a company. This is where the company description comes in play. You tell which software are you using inside your company so it knows who is allowed to use what and how you use it, most importantly. And each and every company that we came across so far and probably it's just the beginning of the list and the definitions. This section is responsible for the acronyms, buzzwords and everything in between which is unique to your company. And it can mean something completely different from something that doesn't mean anything to a person who doesn't work for your company. Let me show you the full list. For instance, Missy Elliott has nothing to do with a rep artist. Apparently, it's a conference room located on the eighth floor. Same goes out to the RCN. Well, this is an acronym which is used by some organizations. It's not used by some other organizations but in our use case, it's the receipt control number. And this is where the AI will get its knowledge from. But not only that, its knowledge is fueled by the knowledge base. The knowledge base gives most of the content and makes it available for the AI brain to process and respond with context in mind. And if you're using Atera components which are Helpdesk and RMM, you will see that your system optimizes and shows you how you are doing in regards to progress. It will also show you which of the result tickets can become good candidates for the KB articles. It also gives you a pre-built summarized KB article which you can inject right away. And after 10 to 15 minutes, it becomes available and you can ask questions right away. And it factors in the presence of those KB articles. So basically, you are seeing the KB which is pretty developed already. And for instance, we have public holidays. Let me show you the KB example. And either of those two, the end users or the IT technician can ask, what do we do as a company? What are public holidays for this or that year? So all of it gives them company specific knowledge and it's context aware. Okay, what else? IT Autopilot, as I mentioned at the very beginning of the demonstration. So the Atera agent needs to be present on the workstation or the laptop of the end user. And as soon as it is installed, it knows what the device is, but also needs the contact data to be attached to it. So there's like the end user who is logged in and this user's email and this user's phone number are crucial for the Autopilot experience. What do I mean by that? So certain actions like the chatbot help are available to each and every person. However, the presence of the connection of the device user and user contact data gives Autopilot full freedom in regards to being able to give full end-to-end support because it knows who is talking, from which device this person is connected from. And it completes the picture and allows it to utilize the company knowledge to help to offer very, very direct help to the end user. Okay, next. Something that we added just recently. This is the AI assign suggestion. What is it? So sometimes you add new devices to the system and sometimes they miss like the contact data or vice versa, you add a contact and the agent is not assigned just yet. This tool will help you out to assign the contact to the device, which will then in return for the end user will become a very, very smooth experience. Let's switch to the end user's device. Okay, as soon as the end user will be having an Eteri agent installed on a device, there will be a help desk icon located here. Also the tray icon, if the person right clicks on it, it will allow them to create a new ticket, manage ticket and trigger AI system by just pressing F1. And for the first time the person logs into the portal, they will get a welcome tour. The welcome tour exposes the end user to what Autopilot can do. And if device mapping is in place, the last action is to perform the device diagnostics. So this is where the Gentik AI nature comes in and it tries to understand who is talking. And yeah, while it's doing the diagnostics, let me zoom out. Yeah, let's go to the hardware and just do, let's just wait. Okay, let me check it because it, okay. I'll skip it for now, but we'll check it. Maybe we need to do something there. Okay, as soon as the end user is on the portal, it can do many things and they can start a conversation. However, what do I mean by how critical is assigning the device to the contact data is to that this menu will be exposed only if the mapping is done right. And this basically shows the full functionality of the Autopilot in regards to more like the flagship features like resetting the local user password, like Azure password and whatnot. Because some of it will require an SMS verification just to make sure that the person, the person's actions are intentional and they're like, there is no security concern involved. And let's start by a simple question. What can you do? So this will expose end users to basically the general knowledge about the product. And they also can ask away something like what do we do as a company? And this is where the custom instructions come in. And you'll see the, yeah. So as you can see, Atari is an all-in-one platform, which matches exactly the custom instructions that we've seen. I can type a question which is like, I'm looking for Missy Elliot. I'm intentionally not capitalizing letters and making typos. So I can demonstrate that it reads quite well. So, yeah. So Missy Elliot is a conference room on the eighth floor. So even though I was like looking for Missy Elliot and it caught exactly what I meant, which is great. But let's go to some use cases which do not involve the chatbot capabilities. By the way, while I'm chatting in the background, let me switch back to Atari itself. For the IT technician, the ticket has already been generated. And you can see that the conversation is fully captured. So what can you do? I'm looking for Missy Elliot. And as you can see, the currently assigned technician is AI, not a real human. Okay. Back to the end user's device. I want to talk to IT technician. This is a feature that has been introduced, what I'm about to demonstrate. So escalation to the technician existed even before. So I want to talk to IT technician. I want to talk. I want to talk to a real person. Okay. I need a technician. Escalation. Okay. So the ticket has been logged and one of our technicians will take it from there. As you can see, right now we don't have an automation. So let's imagine that I took it myself. And I just generated, let me like, hi, like, what's up? And the conversation keeps on going with the real person. And as soon as the real person responds, the end user and the backend will see the response from the technician. And the end user can respond, it's all good. I'm just testing how it works. And then as a technician, I understand that I need to involve AI assistant. So I send it back to IT autopilot and this is a new addition to the system. And as you can see, the end user is now again in AI's hands. So basically if the IT technician feels that it can be done by AI, it still can. It's still what it can do. It can toss it back to IT autopilot. It can handle it from there. Okay. Let's tackle some real use cases. So I think that it would be valid to do something. I found a corrupted file on my laptop. Can you check it for me? I don't remember. Sorry, typing in another place. I don't remember the file name. I mean, remember just scan the hard drive. Okay. So as you can see, I gave it zero context and it started to do some action-based work. So in the background, what IT technician sees is that the chat keeps on going. Okay. IT autopilot ask for, I say yes, start the scan because this is one of the sensitive actions and the IT autopilot needs our permission. And what it does in the background, it executes a command, which probably many of you know, is SFC scan now. And it's not only doing that, it's also recording the progress. So while it's doing its job, I'll switch back to Atera and I will go to reporting and recent processes. So I'll do just today. I'll wait till when the process is finished. Okay. So while it's doing its job, let's check something else. So let's go back to the ticket itself. Okay. This is another ticket. Okay. Okay. The scan is complete. Okay. So the scan is complete. So it waited, like it took about two minutes to complete because it scans the full drive. Now let's go to the reports, recent processes. It will take some time to refresh. Okay. Let's refresh the page. And as soon as we're done with the recent processes, it should, okay, should display. Okay. Interesting. Let me check on the device level. Okay, let's see. Gentlemen, apologies for that. My demo environment might be not feeling that well today. So let's switch back, like while it's refreshing, might take a minute or two to display. So I want to get back to, like the action has been completed. So let's do something on the device. On the last session, we checked the date in time settings, the action that was like performed on the previous session was the ability to sync. Now, like the clock, for instance, like the clock is out of sync. So the one-time passwords can be out of sync and everything can basically, like the person will be prevented from doing their work. So let's start a new chat for a clean experiment. My clock is out of sync. So as soon as this question, okay, the time is off. So in the background, let's take a look right here. We'll be looking at the sync. So, okay, this could take up to a minute. And while we wait, this is the use case that I demonstrated on the previous session. And as you can see, it synced to exactly the time. However, we have many employees who are traveling for business and something that I didn't show just yet, but I want to show it to you. So yeah, even though it responded, but it still retries to execute the clock sync, but it did what it was supposed to. However, I want to send it a message and like set my time to PST. Oops, sorry. Okay, now let's take a look at the clock itself. And as soon as it's done, my time zone should change. So if I'm traveling for business and if I'm stuck and I cannot use the one-time passwords because they don't match or maybe some like SSO issues that I'm experiencing, and it's very frequently happening right before the meetings for the people who are traveling, especially for the management. So they don't have to wait until someone else comes to their laptop and fixes it. And this is exactly how you might do it. So it can be also not that IT related, but still it's on the edge of helping you out in the blend of IT and facilities. I'm looking for a printer. So, and this will give us an understanding. Okay, I need to print and destroy the document immediately. So I'm giving it very little to no context. And as soon as it's okay, and as you can see, it's asking me like questions and I'm responding as vague as I can. It's kind of the knowledge base. And if it finds something, it gives me the very brief answer and also guides me to the full version of the document. So I'm giving it very little to no context. And the article is stored here, how to resolve printing issues when a physical printer is installed, is not installed. I need that nearest shredder. Let's take a look. Okay, private printing and secure disposal. This is what I was looking for. So private printing and secure disposal. So how the desks are numbered and how to locate the printers. All these IP addresses, of course, like simulated, but it will give you an understanding which printing station has what capabilities. Also, like, is it color? Is it black and white? So basically it gives a person an understanding what they can expect from this. Okay, so also, like, my Windows search bar stopped working. Windows search bar. Very common request stopped working. And these use cases illustrate perfectly how the end users might tackle it. And it doesn't do anything. Yeah, so, and as you can see, like, it stopped working. And as you can see, like, autopilot starts to look for some actions. And currently we're supporting more than 300. I'm exposing you to, like, the tip of the iceberg. However, the use cases may vary depending on your company's use case. And then, and as you can see, I said yes, and it just does it. And while it does it, I can start a new chat and ask it to install something silly. Because this is another use case that we haven't covered on the previous session, but I wanted to show it to you because Winamp is one of the software that I have whitelisted. It's checking before doing its thing. It's also checking if it's already installed. And I kept empty space here for a reason. It will first look for the software and then it will do what I asked and it will try to install it. And we should be looking around this area for the software whenever it's done. It may take up to 10 minutes, but let's make sure it does what it's supposed to. Okay, I need to give it a confirmation. It's thinking for a second and then does what it's supposed to. So the use cases, as I said, they're very broad. As soon as you have a very specific one, we can test it on the next session. And as you can see, it started installing it. It may take a minute or two. Oh, there it is. So it was installed right away. And now let's check if it actually did install it. And yeah, it's definitely installed. And I felt old school today. So for whoever still remembers those times, WinAMP is now running on this machine. The same goes out to any software that is whitelisted. It can be Chocolaty based. It can be a private repository. And it can create the library of whitelisted applications. And this is exactly how you do it. To finish off the demonstration, so I'll just lock my screen because I finished the demo and it's time for the Q&A. Let it think for a second. And let's see like that. It takes about 10 seconds to do it. But then it does it. Three, two, one, and we're locked down. Okay, so this demonstrates how IT Autopilot works in real life. Autopilot works in real life. It is capable of solving many, many IT use cases, depending on what you're trying to achieve. And before we finish, a quick Q&A. So certain questions that I received. Hi, I didn't know of the part one. Could you please send me the recording? Yes, it's available in the on-demand section on our website. However, we'll make sure you'll receive it. Jeff is asking, in the bit when you're asking Autopilot for a technician, can we somehow teach it our phone number to include in the chat? So in the coming weeks, we'll be having, in addition to custom instructions, which is called playbooks. So it will additionally guide IT Autopilot how it should respond, what information to include. And this can be something that can be definitely achieved because the troubleshooting steps might be specified and the information provider can also be specified. The second part of my question, Jeff, again from you, second part of my question, where you're asking AI to get a technician, how is the real technician notified that a customer is actually modifying the ticket? The short answer is that there is a built-in ticket automation, which is called escalated by the AI technician. So whenever AI escalates it to the end user, there is an automation to assign it to a specific technician group or a specific technician who handles exclusively those type of tickets. Before it becomes their problem, AI tries to tackle it. And also, as I mentioned on the demo now, you're also capable of sending it back to AI if you did your part, which requires a human to do it, then you can send it back to AI to finish the job. Can the AI system be used in conjunction with other ticketing tools such as ConnectWise? The short answer is it's doable. We have built certain integrations on demand and it's a case-by-case scenario. And please, please click on the request a demo button for this. Josh, this is a question from a user. Josh, this is a question from you. So this is something I would encourage you to discuss with our account managers, account executives, and professional services. So it's a case-by-case scenario. So this is doable. However, depending on your setup, it might require certain types of customizations. Okay, before we wrap up, that was all the questions so far. I hope I answered all of them. I'm checking the chat, nothing in there. I hope it was useful. Let me know in the chat. And in the meantime, I'll launch the survey and you will also see it at the right-hand side and you can submit your answers and your feedback is more than welcome. I'm still here for the next couple of minutes. Oh, another question from Jeff. Jeff, I noticed there is a reset Azure AD password in Azure AD MFA. How does this work? So IT Autopilot can be connected. So technically, Eterra can be connected to Azure AD. We offered Azure AD sync even before. Right now, Eterra connects as an enterprise business app, which gives it a certain permissions that allows it to perform actions like a reset the password and reset the MFA. This requires a bit of a setup, but you will find a very like a KB article on the topic. I'll share it in the chat. Enable password reset Azure AD. So I'll send you the KB article on the matter. So you can read it. There you go. And you can read a little bit more about it. How is it done? What does it involve? But the TLDR version, it's available for Azure AD and Okta, both of them. And you can reset the passwords. Even if the user lost their password, they have just the email left. They will receive the reset link to their like backup email and they will still be able to reset the password and come back in. In case of the local password loss, like IT autopilot will, even another employee can try to help their colleague. So basically what it will do, it will reset the local password and also will send a temporary password as an SMS message. This is why, as I mentioned, the contact data, both emails and phone numbers are critical. So we can do complex tasks like the one that I just described. And Jeff, you're more than welcome to click the request demo to get even more information than what I shared today. Guys, we have some more webinars coming in. There will be a part three. We have some more exciting AI additions. And not only AI additions to the platform. So you're more than welcome to be our guests to be part of the upcoming webinars. We do them every other week on different topics. I will share the link in the chat. One of the things that I'm personally doing on December 11th, same time, I'm doing the analytical reports power user workshop. And for whoever is interested in how you can elevate your reporting, it's especially available for the existing customers. But you're more than welcome to join me. And we'll answer all the burning questions about the analytical reports, which are powered by Looker BI, which is integrated into Eterra. So we can retrieve all your hardware, software, and other interesting pieces of information from Eterra. That basically concludes today's webinar. I'm very thankful and honored to answer all your questions. I hope I did a great job. And you're more than welcome to join us on the upcoming sessions.
