Steve Freidkin, CEO of NTIVA our special guest!

All right, we are live. Welcome, everybody. We're in the final Thank you, Gary. Do, can we, can we get that? Hey, Gary is fired up. He's got the fired up shirt on. All right. So, um, I forget who set it out there, but yes, we are in the final weeks there. Pat said we're in the final weeks of 2023. It's really hard to believe. I hope everyone had a fantastic Thanksgiving. And, um, yeah, we got a, uh, heck of, um, a cyber call today. I think this is something that is on everybody's mind.

We've got a lot of different topics to cover in terms of ai, LLMs, however we want to refer to them, but also the executive order, um, what it all means at the end of the day to MSPs. And, um, I thought, um, wow, if I could get, uh, anybody, it would be, uh, Steve, uh, Fred can, uh, CEO of Antia on for his perspective, and we'll introduce him momentarily. A few quick announcements. Gary, one thing I'm really excited about is, and thank you for the help on this.

We're gonna have, uh, Chris Garretts joining John Strand and Ryan Weeks at the right of boom, uh, in their pre-day. And this is a pretty interesting pre-date. It's A pretty good lineup. Those three. Yeah. Yeah. Just to, just to touch a brain power there on. Um, and, but they're gonna be working on something unique, which is gonna be spotting gaps in your, your own security controls within your MSP, how to do that. Um, and, uh, so, uh, again, thanks for that.

That was a new edition that'll go up on the site. Um, and we do have a, uh, a Cyber Monday code out there that I sent out to everybody. I Mean, all these pre-day selling out, like gotta get registered. Yeah. And we still, like, uh, we're gonna sell out Schiz Fest as well. So people looking in January to come to Schiz Fest, you'll laugh, you'll cry, you'll sch nz, uh, go to sch nz fest.com. Very, very cool. Gary, for those that don't know, what the heck is sch NZ Fest? Yeah.

So for the, since 2010, it's been the annual Truman, uh, member conference. But we invite, you know, guests, uh, to come and really a, a unique event because it is completely content based. Um, about half the time is on, uh, main stage and about half is in breakouts, sales, you know, marketing and, and business leadership and, uh, some of the best people you know, in the business.

And having great place to come and bring your team to get everybody on the same page and, uh, you know, get prepared for the, for the coming year. And, um, I, I would highly recommend, uh, why we still have some seats left. People think about it in Phoenix in January. Phoenix. That's an interesting place to have it. Gary. Yeah. Phoenix in January. That's not so bad. Not so bad. Cool. Alright, so let me set the stage here. Um, we haven't had heard anything about AI this year, have we?

Phyllis, you're on mute. Um, it's gonna be interesting. I, I wonder, by the way, Phyllis and the controls, any revisions, if you guys are going to look into that. Um, I'd love your comments and commentary when the floor is yours with Steve on that. But, um, this is something that, you know, you, there, there's so much that we could go into one, you know, and we are gonna talk about the executive order, you know, around safe, secure, artificial intelligence. Um, what does it mean for MSPs?

Can MSPs actually monetize this in a, in a, in a meaningful manner that actually is good for MSPs and Steve and team, I think are, could be on the cusp of doing that. Um, these guys are certainly do things at scale. Um, and I was talking to Steve, uh, offline, I remember when, and the reason, you know, when he tells you the size and scale ofa, um, I, I want you to think about how you look at your MSP, um, Gary, you know, a lot of MSPs say, you know, I'm different. I can't do it or whatever.

Steve, you were small when I first talked to you in 2 0 5. Isn't that fair? Yeah, absolutely. Yeah. So welcome to the show. Um, it's awesome to have you talking about this subject. Can you tell everybody a little about your background, uh, and Tiva story and yourself? It's, it's awesome to have you with us. Yeah, sure. And I'll, I'll keep it brief today 'cause there's a lot to dive in, uh, all around artificial intelligence with, which I'm excited to dig into with all of you.

It's gonna be a fun conversation. Uh, but, uh, I'm Steve Fratkin. I'm the founder and CEO ofa. Uh, we're a full service managed IT service provider. Um, and we, we, uh, are currently on our second round of private equity. Um, we are a little north of 500 full-time employees. And we have the intention to continue to grow our business and scale, uh, throughout the country, uh, as long as we stay true to our core focus and our purpose of growing people using technology to accelerate that growth.

And that's gonna, you know, that's gonna be a theme, especially as we look at artificial intelligence in everything that we've looked to do. Uh, so we are a little bit different in terms of our goals to not just grow for growth sake and make just more money for all of our shareholders. While that's important, that goes along with our purpose driven culture of helping people grow with technology.

And that's really what we're about at antia and everything from traditional managed services to security services, to digital transformation and application development, including, uh, uh, the team that, that can actually deploy custom artificial intelligence, large language models. Uh, so kind of exciting stuff. Uh, so that's, that's really about me. And, um, uh, where we're at now. De before I hand it to Gary, when we first spoke, what, what, what was the size of your Ms. P in 2 0 5?

Oh, so, so in this is, I, I'm a lifelong geek, right? So I started doing this in middle school when girls wouldn't pay attention to me. And So I was starting with a residential customer here or there. And then over time, that grew when we were talking in 2005, maybe we were 15 employees or so. Yeah. Uh, you know, and, and it was really just, just grit and kindness. Up until 2016, we were maybe a hundred employees, a little shy of a hundred employees when we brought in our first sponsor. Yeah.

Well, speaking of kindness, I wanna introduce the Don Rickles of the cyber call. Necessary kind and true. Mm-Hmm. So Steve, let's start, uh, maybe with giving us a little, like your opinions about, I guess what we're trying to see governments around the world don't know really what to do yet, right? With large language models, with ai. Um, I think everybody, well, most people agree, and there has to be some guardrails on it.

And so one, one step was the Biden administration, um, with their executive order, uh, that came out in March. What do you think about this? Does it make any impact? Like where, where from a regulations and, you know, guidelines, guardrail standpoint, what's your opinion? Yeah, so look, I, I, to to say any of us are experts on this topic, even though some of us have spent a little bit more time in it, is, is really hard right now. You know, things are changing rapidly and moving quickly.

So, you know, today I'll just talk from the experience that I've gained, understanding that we're learning more things every day. Looking at the most recent executive order that, that Biden put out there, you know, it, it requires, it requires developers to share safety test data, right? So, so ideally, um, anything around what's being done is getting reported up.

If the company that's creating it as of a certain size, it, it's been designed to, to create safety standards and tools and tests around ai. It's been designed to protect against any model from, uh, allowing folks to engineer dangerous materials, whether it's chemical, biological, radioactive nuclear, or that kind of stuff. Um, and it tasks various government agencies to go, well, what are the social implications of using artificial intelligence?

And how can we use it to actually benefit society and, and not cause harm? And then ultimately, to create ways to reduce the likelihood of deep fakes. Meaning how can you, um, how can you tell when something's artificial intelligence versus not that that was the idea. Um, and, you know, I think the order itself was met with a lot of support from various parties. It was one of the few that, like, everybody got to have something in there that they wanted.

So almost everybody was pleased with what was put out there. With that said, it's an executive order without, as far as I understand, I'm just an it geek, but without the Congress, like actually supporting it and there being laws in the states that enforce these things, there's not much that can be done here. Um, you know, short of following through with the guidance that's been put in place, much of it is very good.

Um, I, I also think, and, and maybe you guys can chime in on this, but I think he invoked, if that's the right word, the Defense Protection Act, which was kind of clever in, in, in it is one of the few tools the White House has to pull commercial companies into the fold and, and really ask them to do something, uh, to protect the country. Um, all of this to me is a great start.

And I'm, I'm grateful for the US stepping up and doing something, but I struggle to see this really take root and have legs unless the entire world agrees to do that. And understanding right now that, you know, in the entire world can't seem to agree on anything. Um, I, I struggle with how, um, how enforceable this will be or how realistic it'll be to keep our country in a strong powerful position if we're taking a backseat to other countries who are moving faster.

So, lot, lots there, and hopefully we can have a conversation around that. But that's, that's kind of my understanding of the order, uh, the, the tools that, or, or the, the protection, um, uh, the, the, the Defense Production Act that, that Biden put into drive it, and then the reality that it's not exactly enforceable. Yeah.

And you know, it's interesting now we've been through so many cha changes so fast, like things happen so much quicker now to, to be at the beginning of something this big after we've been through other, you know, other technologies and seeing both good and bad. It's, uh, it's super interesting, but it's definitely gonna take some, it's definitely gonna take some time to, uh, you know, to find its way. So let, let's turn this a little bit back.

You know, as an MSP, I know you guys have built an extensive, you know, cybersecurity, uh, offering, um, and you deal with a lot of regulated clients. Can you give your perspective where you see AI being used for the SMBs, the clients, the cus we'll start there with the customers. Yeah, so look, I, I think we, we gotta separate AI and, and talk about chatbots, right?

So yeah, first and foremost, obviously the chat, GPT and equivalents that have come out over the last year, um, are being used regularly by many people, uh, SMBs included. And we think there will be continued adoption of that to take, uh, you know, repeat tasks and make them simpler for the end user. Uh, now of course we want to provide guidance on how to properly secure the data and use it the right way.

But our, our perspective on that is we're gonna see this continue to roll out more, um, our advice to our clients, and frankly our team has been to just go out and experiment with this, play with it, understand that there are security risks and, and teach and educate around that. Um, get into options that have DLP or data loss prevention put in them, and then just play experiment. And I, I, I can think of many uses of that in the next year.

And I think we're gonna have significant adoption, if not complete adoption across most SMBs. Then I would see that the line of business applications for these SMBs are starting to incorporate ai. I mean, like, I, I joke about this all the time, but I think I was at a Denny's the other day, and on the menu it said it was powered by ai, right? Like, so everybody's gonna say it, whether it's true or not, I don't know. And that's mainly how they're gonna get this, right.

Like mainly most SMBs are gonna experience this through their software vendors, Correct? Yeah. And, and so they're gonna have these line of business applications that are now, you know, either included with their updates or for an extra fee. They've got these new feature sets that are gonna come out that ideally would take steps out of repeat processes and or give more visibility into their data and what that data means.

And, you know, that, that ties to the, the other area that I think it'll be used the most, regardless of line of business applications, it's really around business intelligence. It's around getting, you know, meaningful, meaningful output from all the mass amounts of data that we have. You know, we've got lots of tools that are out there to help with that today.

Uh, but, but having an artificial intelligence layer on top that can sort through all of the data, look for patterns and, and present information in a meaningful way, I think is, is really, is really awesome. And, and honestly, I think it's something that SMBs today couldn't really do on their own. Yeah, I'm not real optimistic.

I mean, I've been, before I had my MSP and, and True methods for 30 years, I sold software, and what you just described is exactly what I was trying to sell people, you know, 35 years ago. Um, it's, it's a, it's, it's elusive, right? Trying to get all this data and present it to people in a, in a meaningful way. Uh, yeah, definitely. Whoever figures that, it's gonna be pretty well, Steve, do, do, do you see it internally in your SOC or some of your larger customers?

Are they using it to look for patterns and threat protection or, or in false positives to, you know, quiet noise? Are you starting to maybe talk? We, we are, we ourselves are already using it to quiet noise, uh, really understand when, uh, we've got false positives coming in, look for patterns and behavior, see what we can do and apply elsewhere.

We're, we're already seeing that, uh, for, for our customer base, sort of like a few outlier, really large businesses we're, we're not seeing a lot yet. But the, you know, the majority of interest again, comes from, uh, people's use of those, those chat bots and, and how they can be leveraged to help take steps out of what they're doing every day. Um, and, and again, just, just getting meaningful information out of their mass amounts of data.

Uh, I'm not sure if I had a connectivity issue a, a second ago, but the, you know, the, the SMBs that I'm used to, to dealing with, they, they haven't been able to afford having a data warehouse and a full-time business intelligence person or two, or three, or four or five. And to be able to have that come down to the small business without needing all of those resources anymore is very exciting to me. And I think something that's gonna move quickly. Yeah.

So do you see that, so on one side of it, you, you already touched on it, which is using this to, you know, uh, increase, you know, help with security right? Can be a positive for security. To me, that's what Quieting Noise does. Uh, it, it allows you, um, to be able to sort through, you know, more data to find the things that need to be acted on, hopefully make you as an MSP right. More efficient.

Um, because we see noise levels in general going up over the past couple years with the admin of all the, the tools being implemented. What about on the other side of these applications? You're starting to touch on it now about how MSPs can generate actual recurring revenue in some of the ways that you're, that you're talking about, which providing some business value that MSPs didn't have in the past. Right.

And, and look, I mean the, we, we try to be as vendor agnostic as we can be, but you know, I'm, I'm looking at some of this chat here, and obviously a lot of what we're gonna talk about today revolves around co-pilot, which, you know, I've, I've, I've been a beta tester of, but I ivett even have it rolled out across my organization yet, even though we meet all of the minimum criteria, et cetera, right? Like, it's gonna take some time, uh, to, to roll all that out.

But each of these add-ons become new recurring revenue opportunities for the MSP and opportunities to create more value and stickiness for the customer if you're teaching them how to use it to be a competitive advantage before it becomes a competitive necessity. And that's something that like we see as a huge opportunity. So whether it's the additional Microsoft 365 licensing, whether it's a package where you include, uh, some consulting services that go into that, like we're very excited.

And that's, that's easy, right? Like that's, hey, you should get the Bing Enterprise addon instead of, uh, you know, uh, chat bots on the side because we can have data loss prevention in input it in there. That's gonna be x number of dollars per user per month. And included in that, we're gonna provide you a policy manual for your employees to review and execute that support the proper use of this tool.

You know, those are, those are some very easy examples that you could see going into 10 user companies, a hundred user companies and, and doing at, at scale. Um, so we're, we're very kind Of prepackaged. So it's good for MSPs. Yep, exactly. And then, and then there's, there's more custom solutions that, you know, we've been looking at too.

So, as an example, uh, we, we spend a lot of energy on, on API automation, and anytime we, we can figure out that, oh, this customer in manufacturing uses the vision ERP, and they are connecting it to HubSpot for their CRM, and they're doing all this double data entry in a manual way, like we'll go and we'll do a custom development project for them to tie all of that together.

And when we do it, we'll wind up with a library module, oftentimes leveraging artificial intelligence that then can be used again for people that have vision and HubSpot. And so we are very intentional about creating more of those modules and then using, um, some level of artificial intelligence in our database to quickly look, oh, we support 1800 clients. How many of them are in manufacturing? Oh, 70. How many of them use Vision? Oh, 40. How many of them use Vision and a HubSpot? Oh, 22. Okay.

Communicate to their account managers and tell them, Hey, we can solve this problem for you for $499 a month that used to take hundreds of thousands of dollars of custom development. So that's kind of the, the next extension. Yeah. Uh, that, that we're doing for, you know, a smaller subset of the customer base. Yeah. So I have one more question for you, but I, I just, you know, I can't help thinking, uh, just, this is just another example of, um, how scale makes such a difference today.

And, you know, I built my first MSP at a time when scale didn't really have a lot of advantages. It had more disadvantages, uh, than it did. But now we're seeing in all areas, this is just another, another one of those things. So the average MSP really has to think about this. Um, what do you think some of the business, the biggest risks are in terms of the average MSP trying to jump into some of these AI services? Like where are the main pitfalls?

Look, with, with, Lemme talk about me for a minute, right? So as this whole opportunity was coming up through chat, GPT coming out, becoming mainstream, uh, our team Atva really did some amazing stuff. And they figured out how to, um, use an Azure instance and put a private AI layer in between what, um, was happening in the engine and putting in data loss prevention right away, stripping all P-C-I-P-I-I, all the information that would cause risk.

And we're like, oh my goodness, we're gonna be first to market with this. We're gonna bring this to every customer. We're gonna do it for $20 per user per month. And while we're at it, we should throw Stripe on our website. 'cause anyone in the world could buy this thing from us, right? And so we're, we're, we're, we're running, we're running, we're running, and we're about to go live with this.

And like, boom, Bing Enterprise, uh, announcement comes out and Microsoft's got this, and it's gonna be integrated into M 365. So we just wasted, oh, I don't know, uh, 800 internal hours and 200 grand, right? Yeah. Now that's a risk, right? So I, I just wanna experience, share myself, right? It's, it's getting really excited. Um, and then, and then thinking you can move faster than one of the big technology companies that are out there, right?

Like, I, I, it's, it's always, it, it, it, it's just not the case, right? Like, they're there, So nothing that happens, they, they don't let us, they don't let an MSP, we can make money on some things for a little bit, but, you know, once there's enough money in it, they take it back. That that's, that's right. So, so really, you know, my, my cautionary tale here is let's make sure we're partnering with those people. We're, we're, we're early on in what they're bringing to the masses.

We believe in it, we understand it, we can be experts in it. We've been trained in it, so we can be one of the first movers to deliver it, but let's not necessarily think about trying to create that custom solution to bring into the market, right? So that's, that is one example. Uh, you know, the, the other, um, that, that, that I would just wanna raise awareness too, is just, just the, the security implications around all of this, right?

So if you're bringing this to the masses to your customers and saying, just go, go, go, go, go. And you haven't trained with policies, you haven't done the right, you know, CYA, make sure they understand the risks, you might be unintentionally taking on liability, uh, by, by pushing this into your customer base. Without that, that additional support. And that's, listen, if you're an MSP, you know, less than, you know, five, six, 7 million of annual revenue, really hard to do.

Like we see a lot of it going on collaboratively, um, inside our peer groups. So they try to share, you know, the r and d on it. But I would say the average, like the, the MSPs that are 3 million a year and below, they're probably gonna need to have it packaged by a vendor, like a lot of this. And whether it's Microsoft or some of the vendors in our space, well, I mean, you're working to them And, yeah, Exactly. Alright, so I'm gonna, um, hand the mic over to Phyllis. Great, thanks.

Um, so you mentioned that, um, you know, copilot, and you guys are pretty much almost at, that you're at the last mile, um, with actually using, um, or implementing copilot. But can you tell us why you as an MSP chose copilot and, um, have you had client interest in it? And if so, you know, what have they been asking for in particular? Yeah, so, so, you know, I wanna be cautious here, even to all my Microsoft friends that might hear it later. We are vendor agnostic, right?

So at the end of the day, our job is to make sure the right technology gets in the hands of our customers in a safe and secure way where they can have a competitive advantage, right? So, you know, whether, whether we believe that's copilot or Bard or whatever it might be, uh, you know, that that can and will change over time.

But right now, you know, we have had the ability to have the vast majority of our customers, I'd say, you know, 99% of our customer base is leveraging Microsoft 365 mm-Hmm. To some level. Um, you know, the, the few holdouts we had that were in, in Google Docs or whatever it might be, or G Suite, some component of it, um, have, have mostly migrated to this platform. So people are already in this.

And, and our job, again, is to bring the best technology to the customer that, that creates that competitive advantage in a secure way. And of course, we need to be thinking about cost too. So we've got clients who are almost all in this ecosystem with a vendor, Microsoft, who has historically done good with governance and policies and a focus on security. Does that mean it's gonna be perfect going forward? Absolutely not.

But they built a track record around it, they have trust around it, and that is a, uh, a focus of what they're doing. So understanding the fact, you know, direct answer to your question is most of our clients already have systems rolled out that this will layer right on top of very easily for them. And it's with a vendor who has historically shown key focus on both governance and security And who wins a lot. Yeah, yeah. As former, I totally get it.

Someone did ask, like, you know, um, I think someone asked in the, in the chat is, are there, um, specifics you can give about some implementations or, or no, like On, on, uh, like any on copilot specifically? Um, yeah, well, we can talk about copilot or just AI in general though. Yeah, look, I, I mean, I'll, I'll tell you.

So, so the, the, um, data lakes and anything you can do with, with getting all of your information into a large language model that's processing it all the time and then seeing what kind of correlations and data you can get is, is mind blowing, right? Mm-hmm. So, so like, the ability to just dump a ton of historical information into a system and draw out patterns from the history of your business is, is incredible and, and relatively low cost. So that's what we're seeing.

Um, you know, we've seen a, a couple of, in our larger customers, we ourselves are doing a custom development project right now where we are, uh, we're actually going live officially on December 1st. We've been in beta and we've been testing this, but we now have a workflow where, um, you know, again, I wanna be cautious for everyone's time, but you all are msp so you'll get this, um, atfa, one of the things we, we've done a really good job with is got very, very fast response time, right?

So you call in with two minutes or less, you're on the phone with a technician, and 72% of the time that person you're talking to can resolve your issue on first touch, which is good, but we'd like that second piece to be much better. And when it goes from that first touch to what happens next, like, who does it go to for escalation? When will it happen? Who will be, that's where things fall apart.

So no matter how good our processes are, no matter how good our communication is, there's always mismatch of expectations. And our goal to improve that is to simply put increase our first call resolution, first touch resolution. Okay, so what are we doing? So now when tickets are emailed in, they're actually being read by an an AI bot, right? And that's breaking it down into different segments of what the request is.

So if it was a lot of stuff, it breaks it down into individual issues, it prioritizes it based on the parameters that we have in place. It searches our database to say, who in our company has worked on something similar to this before? It searches our knowledge base and says, what were the resolutions to do that? It searches the external internet and says, what are the things that we could start with there?

And then on top of that, it does sentiment analysis, and we'll turn the ticket to green, uh, yellow or red, depending on the customer tone, which allows our left brain engineers who are trained on customer service, but don't necessarily know when to turn it on or off to actually know what the heck's going on.

So now what happens is this, this information that, that used to kind of bounce around, we can be laser focused with where it's gonna go, how we're gonna start, and again, get that first touch resolution up. So I'll have more to report to you in about a month, but I, I haven't been as excited about launching something like this, uh, as I have been in a long time. And, you know, it's only a matter of time before, you know, ConnectWise or Kaseya built it into their own platforms.

I know they've been talking about it, but we just, Bruce already has crates on this. Oh, awesome. Awesome. I think that's, um, really awesome that you're also, um, you know, trying to discern the sentiment because that, that is a big deal when it comes to customer Satisfaction for, for those, for those engineers. And then, you know, before, before I, I get off this topic, like, I just wanna go back to the chatbots for a second.

So one of the things that I was challenged with from day one was just use it for 15 minutes a day. Doesn't matter for what, whether, you know, I p****d off my wife and I need to send her an apology poem, or I have a press release that I need to create that I want to, to accelerate how I do it. You know, the more and more I did that every day and I just had it up and running with me, the more I pushed its limits and I'll just throw out another use case or another.

You know, you, you asked for some, some examples of what, what's been done. Um, it, I have been playing now for maybe four months, 15 minutes a day was my commitment. And I got this 31 page contract for an $8,000 a year service for antia. And I was like, considering our scale, that's relatively low, and this is a really long legal document. So I actually popped it into to chat GPT-4, and I said, please summarize this for me as the buyer of these services. And it took like a good 10 minutes.

It was like the longest I had to wait for this. And it came back with like a succinct, less than one page summary of it. Wow. And I challenged it and I said, as the buyer of these services, what are some items that should be called to my attention that create risk for me? And then a couple seconds later, I got about 14 bullets coming at me, and like nine of 'em made a lot of sense.

Then I said, basically, can you please craft me an email response to the sender of this contract to reduce my risk or make it mutual in all appropriate areas? And then it crafted a, a, a, a draft response. I of course edited it. It wasn't perfect, and then I sent it on the way, the whole thing took about an hour, but it taken me two weeks and thousands of dollars to do with an attorney.

Now, would I have done this with my, you know, my next Kaseya contract for a hundred thousand dollars a month and a five year lock-in? No, uh, you know, I, I would've gone to some more experts there, but given the low risk, like these are the things that, like, I just tried and I kept, you know, I, I'll keep doing this and I'd encourage all of you all to do that too. Phyllis, can I just jump in here? Gary, I, I want, wanna po posture this to you. One, this is for all MSPs out there.

You got a CEO of a 400, 500 employee company, lot of revenue. Look at where his passion was. Example number one was around maintaining existing customers. Because Steve, we always hear talk about, you know, why are you gonna, or how are you gonna get a customer interested in something? Let's talk about security. First use case you talked about was maintaining or retaining customers, making them happier. Second use case was, um, was around, um, risk. He literally said it, right.

Risk to the business. And, and this is what we talk about a lot, Gary, from a sales perspective and having command in these conversations. Can you, can you give us a perspective on that? Yeah. Well, let me tie two of the things he said together. I mean, which is, uh, focusing on that help desk efficiency and first, uh, call resolution as an MSP, they're tied together. We have finite resources.

And right now what we're seeing in the data quarter over quarter is that there's more tickets per endpoint per month being created and on support, less tickets closed on average per day. Mm-Hmm. So when that happens, we're more inefficient. And when more of our resources as a percentage go to help desk, 'cause Steven knows his math, you're, you're not as good at VCIO, you're not as good at security, you're not as good as anything. Right? It's a pie.

And the more pie that you, that you're having go into support. So that's kind of goes to the sense like, what a great place to start, Stephen, uh, there Yeah. Because it rings the bell in, in not just profitability, but in your ability, uh, to be able to take better care of the customer. Yeah. Look, absolutely. And like you, you, you touched on it here, but you, no one wants to be calling the help desk, right? Like, nobody wants to be doing that, right?

So, and, and by the way, for those service desk technicians, again, I, I mentioned this early on in the call, like, we exist to grow each other at antia. Like it is not the highest and best use of their time to be fixing the same printer every day for the customer, right?

So the more of that we can automate to elevate that technical talent to do that, which only they can do and help our technology or our, our clients use technology to grow, again, we create stickiness with the client and we have really good experiences for those employees, and they're able to grow. You know, so like when you look internally about how to use this, that's exactly what we wanna be thinking about.

You know, the more, the more we can, um, the more we can take the monotonous repeat tasks out of the business, the better off we'll be. And you know, Gary, to your point that the profit will absolutely followed, right? Like having these things staffed and, and, um, you know, just the, the, the more of that we can make efficient, the better off the customer will be and the better off our profit margin will be. Yeah.

And my recommendation to MSPs Andrew, is focus on those things that make you strategic right. To your customers. Because, you know, the day is coming next. Steven mentioned it, where first call resolution's gonna come from ai, and at that point, it, we've now commoditized that level, that first level of support. And so it's not gonna differentiate, uh, like what MS too many MSPs hang on to, which is their level of support is not, they're not gonna be able to use that at some point. So, Right.

And what I, I think that time is now, right? Like that's, that's already Yeah. At your level. I would agree. I I, yeah, I would agree. It's, that's already here. And, and by the way, I, I love this, this venue, and I'm, I'm gonna join more of these calls because it's so awesome to see all the chatter and all the questions. It's also distracting for me. So as I'm answering questions, I'm reading these things, so, so apologies to everyone. Yeah.

But you know, there's statements here about, um, you know, about, about this, meaning we need less people. Well, maybe we do, maybe we don't. Uh, but the reality is we're we're doing, we're doing a disservice. We're doing a disservice to our people by not allowing them to learn these tools now. Like we have to teach them, or they will, uh, not have a job in the future, right? It's the people that know how to use these job, uh, these, these tools that will have those jobs.

So now that we, we, um, you know, we're aware of it, it's, it's our duty to teach and our customers are going to expect that they're getting better results faster. And, and, you know, we could, we could pretend and say, well, no, each time you call, you're gonna talk to David, and David knows all about your family, and that's gonna be the secret sauce. Yeah, maybe. But at some point, the, the, the game's changed and it's not that anymore.

Or at least it's not that on the keep stuff up and running part of the business, right? The strategic part, how you're helping your client u or your client use the technology to grow. Like that's where your, your knowledge of the person, your knowledge of the person's business is really important. But fixing that same repeat printer issue that I used, uh, as an example before is not Different. Phil, us back to you, different scale.

But Steve, I mean, everybody call probably remembers early days and 2 0 3, 2 0 4, 2 0 5 about RMM similar. Like, is this gonna replace X, Y, Z? Right? Steve, you know, so us back to you. Yeah, Sure. Mm-Hmm. Um, so we talked about security a little bit, and even the EO of course addresses this. So Microsoft claims that copilot has a strong emphasis on, um, data protection and privacy with the use of strong encryption protocols.

And so, um, how do you as an MSP, um, how does it affect your, um, client policies? And then what security controls do you need to put in place as you, um, deploy AI throughout your business, business systems and enterprise? Yeah. Look, Phyllis, we're, we're all learning as we go right now, right? So, so like, we, we try to have a, so look, you're, you're learning a lot about the culture of Antia today, which, which was not intentional from my end.

Like inside of Antia, we obviously, we have some of the greatest people, uh, including great legal folks that keep us protected and reduce risk, right? And their job is to, to make some of these things hard to learn and hard to adapt until we really have all of the answers. But the answer right now is we don't have all of the answers. So what's the safe way to start using, uh, to, to really get going? Right?

And, and I don't, I don't have all those answers, but atva, we look at original mistakes, right? We are allowed to make original mistakes only. That's part of failing fast. It's part of failing forward, but we have to learn from them. So first and foremost, before we give access to our entire team on chat, GPT or an equivalent, we will absolutely put in place the right data loss prevention.

If we cannot turn that on and cannot centrally use it, it will not be proven inside of the company because we know that that exists and that can be done. In addition to that, we need a policy that everybody understands it's been given to them, they, they've read it, they understand what it can do, they understand the risks, and they've signed off on it. So, so the between, you know, that having at least DLP and a clear policy, that's a starting point. Okay?

Then, you know, what I would say is with, with anyone that you're deploying this for, you've really gotta get all of the stakeholders to understand what's going on and do the right upfront diligence. Hey, are you using a third party line of business application plugin that's supposed to take all of your data and give you more meaningful analysis out of it? Sounds great. Uh, you know, don't just buy it off that website. Call in your technology team.

Make sure the leader is aware of what's happening, make sure security's been tested, like all of the things that need to happen in advance, you know, make sure it is involved to do that. And the closer a relationship that you can have with your customer as the m where they're bringing you into those costs, uh, to those conversations before they make business decisions, the more, uh, valuable you are to that organization. And, and ultimately, the better you can help.

Phyllis, I don't feel like I gave you a fair and complete answer. No, I, I think that's a great answer. And, and I'll follow up with, you know, we do get asked, given ai, are you gonna update the critical security controls? What is it that you all are going to do? And my response is today, um, not, not much. Because what we see a ai AI doing is enabling automation, just like those use cases that you described, excuse me.

So the actual security controls that we have detailed in the critical security controls, stand stand strong, right? Yeah. Have, have those, you know, know what you have, know your environment, have access controls in place, et cetera, so on and so forth. The basics are still there right now. There's not that we know of. And like you said, we're all learning at the same time.

There isn't really anything unique, um, with AI or Emma, whatever you wanna call it, that we think we need to have, um, new safeguards or, or anything put in place at this time. So, so I, I think that's a, a great answer. So Was cloud why you brought in, you know, cloud Security Alliance and that piece singled out post pandemic or around the pandemic just because of the vast shift of work Mm-Hmm.

From on-prem to the cloud versus AI just being not demeaning it, but is was that the, the, trying to sense the dichotomy on why there was such a emphasis there? Yeah, I mean, we, we brought in cloud security alliance and other things because organizations were wholesale moving to the cloud during the pandemic and post pandemic. And so the question was, when you do move to cloud, how is it that we can address these controls in the cloud? Is there a difference?

And how is it that we make sure the way we word it, we're addressing it properly, so on and so forth. And so yes, that's why we did that. Got it. Um, you know, is AI a a different technology space where it causes different management where you're not necessarily in control of everything if you go to your cloud service provider there in perhaps, you know, the shared service, service model, et cetera? So, um, we're not really there yet, so we'll see. Yeah.

Um, you Know, Phyllis, can I, can I, I don't, I'm breaking rules. Uh, can I, can I throw something back at you? So, so when we, when we look at, um, you know, the, the, this goes back to the Biden order and just in general, what can be done with ai, but like the, the depth of these deep fake attacks, right? Like, so, so the, the now the phishing, the, the other, the other ways that like, you know, it's not just me in some crappy text going, Hey, do me a favor, I'm in a meeting.

Can you please send me eight, you know, gift cards to Applebee's? You know, it, it, it's, it's, it's getting more real. It can take, it can take five words of my voice and sound like me on a call. It can take a picture of me, make me into a live streaming video. Um, you know, what, what, uh, what are we doing, you know, short of the government, uh, executive order, which may or may not be effective to train our people on these new attacks and how to get validation, right?

So, you know, have you seen anything there? What, what are you putting in place? So I'll say that it is user training, and we do have kind of like high level sentences that say, you know, deploy security awareness training. I mean, it, it goes down to a little bit, like do anti phishing and things like that. So it could be updated to do like, you know, that deep fake training, right? And what we would expect to see or hope to see is those companies that provide training to up their game, right?

So just like the anti phishing training et cetera is getting better. There's, there are no typos, you know, perhaps the email address is gonna look better, you know, maybe the email address won't be, um, you know, Phyllis dot Lee at, you know, in a bunch of numbers.org or, you know, whatever. And so I think that we do need to up our training game because of that. So maybe it's, if my son calls me for money, maybe it's, well, don't just give him money.

Call your son back, make sure that he needs the money. Do, do, do you know what I mean? Like, Ask him about that, you know, that birthmark on his left butt cheek, Right? I mean, just, just like when, when you had little kids, you know, they taught you safety training. What's our code word when someone, you know, um, comes to pick you up at school or, you know, when someone knocks on the door and a stranger knocks on the door, what do you say? And, and, you know, all that kind of stuff. Yeah.

And I'm not sure everybody's thinking about that. You know, people have felt safer than they probably need to. Now's the opportunity to think again. Yeah. I mean, you know, the deep fakes are, are real, you know, and, and, and people have to be aware, you know, like with, um, the IRS, the IRS always says, we'll never call you and ask you for money. We'll never call you and tell you that you owe us money. We'll never do anything over the phone. Don't ever give us any information over the phone.

Um, and so I think there has to be some sorts of kind of policies like that, um, for some of these scams that are so prevalent where the agencies, you know, have to come out and say, Hey, 'cause you know, the tax IRS scam is pretty, is pretty common. And these other ones, well, It, it, it's exciting. And these are all opportunities, right? So these are, yeah, these are also opportunities for us as MSPs to educate our customers, right?

Anytime we can do that, right, we're, we're creating value, Right? I mean, maybe your Ms P would say, Hey, if you're laying on a bill, this is how we're gonna notify you. We're never gonna call you, or we're never gonna ask for this kind of information over the phone. Or I, you know, whatever. Yeah. Or, or, and over to Wes Wess, maybe a segue here, how do we, how do we do this, you know, to stay away from the next MGM in our small world, right? Our help desk would never do this.

It'll only do these verifications. So Yeah. And I was actually just typing this in. You saved me a whole bunch of words, Andrew. I, I also think that while training is important, we've all seen the sophistication. Like, for example, those of you've seen like the, the Tom Cruise deep fakes. I mean, they look just like Tom Cruise and those things are like three years old now. It's crazy.

Um, and my point is, they're, we're always gonna have difficulty standing up from a training perspective against these kinds of threats, right? They're gonna get more sophisticated. And, and so I think too, like it's not just training, it's also what kind of deployment mechanisms and preventative mechanisms we put in place.

So now if the deep fake is being leveraged to get a trick of human into doing X, Y, and Z, do we start putting in countermeasures around, well, that doesn't make sense why this user would go and do this thing, so we're just gonna prevent that altogether, or we're gonna put multiple stop gaps in place, right? Sometimes it just requires that because it's only gonna get worse and harder. And Steve, I'm curious your thoughts on that. Yeah, I'm with you. I agree.

You know, it's gonna be this forever cat and mouse game and, uh, you know, I'm just lucky to be playing in it. Yeah, for sure. So I'm gonna post a link, uh, in chat right here. Uh, you guys did a really cool video on copilot Microsoft 365 copilot. Um, I thought it was cool how you guys are very clearly and openly just throwing it out there. Here's what we know, here's the implications. Here's some things you might not have known. I learned a bunch watching, it was really cool.

I haven't finished it yet. Um, I'm curious, should more MSPs be doing this kind of thing? And has this been helpful for you guys, just in terms of, um, whether it's market outreach, lead generation, or just making sure that you understand and, and know cutting edge technology yourself? I'm, I'm curious what your thoughts were on creation of that video. Yeah, so look, we're, we're, we're grateful to have scale. And with that scale comes dedicated resources, right?

Like, I remember being, uh, like, like I'm sure many of you on this call where I was wearing 17 different hats at any given time, and everybody in my organization was wearing eight to 10. You know, now we're at a point where we have a full-time product manager, we have people that do this research. We have people that are designed to just do webinars and marketing teams that just focus on these things, right? So we're, we're in a situation where we've chosen to make those investments.

Now why did we choose to make those investments? So back to the idea of growing each other, right? That is, that is our ethos here at antia, right? So we need to learn these products ourselves to pretend like, oh, we know everything about something ambiguous that Microsoft is probably making up on the fly too. Uh, you know, it's, it's bs and, and we, we don't want that, right? So we want to be open and transparent about what we know we want to commit to our audience.

And our audience is of course, prospects, existing clients, and many of the other MSP community members, right? We're, we're believers in the Go-Giver model. And if we're doing this, we're, we're willing to share it. We're all willing to learn and grow together. And, um, you know, that's, that's what we're about. So understanding that's what we're about. We chose to make these now have these investments yielded good financial results. Yes, they have, right?

So by doing this, it's improved our client retention because again, clients are getting the strategic advice, and that's not always giving them the answer of, oh yeah, this is gonna be out on November 17th and it's gonna be this much money and we can roll it out for you by December 15th. Sometimes you just gotta be open and transparent and share what we know.

And, and so that's, that, that, that's helped with clients improving trust with us, which leads to, um, you know, improved client stickiness, uh, for, from a prospect perspective, it's, it's interesting, you know, putting out this content has certainly helped make us thought leaders, and it's, it's generated interest coming into us. And we've heard from many CIOs of companies that, you know, this was enlightening to them and they're grateful and they ask us to keep doing it.

And then we hear from our, our competitors, you know, and, and we, we hear, oh, you know, thanks for putting this out. We used it to train our team, you know, so that, that feeds us, that fuels us eventually those competitors, you know, we we're friendly with them. We want them to grow awesome businesses, and one day join forces, uh, as the, the market continues to consolidate.

So to me, you know, we stay true to our, our focus and, and yes, it drives meaningful results for other metrics that we're trying to achieve as an organization. Any follow up thought I'm a much smaller MSP, maybe trying to get to my first 3 million. How, any words of advice for a smaller MSP thinking, well, that's great Steve, but I don't have those resources and that capability.

Where, where would you start in a smaller MSP when it comes to content and education around, um, AI and, and other things that I think are relevant for, um, for business? Yeah, look, I I, I am such a big believer in peer groups, you know, so if I, if I were to start anywhere, it would be there, um, through the years, and I continue to learn and grow more in peer groups with others than, than, um, probably anything else. Didn't get that. What did you say? Have you heard of these, Gary?

What's that? Have you, have you heard of these peer groups? Yes, yes. That's an awesome idea. Yeah. So, so, you know, I I say double down in that, especially if you're, you know, trying to get to 3 million, 3 million, whatever, the more of that you do, the better. Uh, you can learn and grow from each other.

And I'd say, you know, for years and years and years and years, I was only in Vistage and YPO and all these things where, where I would have these peer groups, but they wouldn't be with other MSPs. And there was concern of like, you know, competition and, and people knowing more than the next person, ah, get over it, get together. We're stronger together. There's enough business to go around. And, uh, you know, I'd say by far that's, that's the, the place to learn the most.

And then from there, people can talk about You're at a disadvantage right now. Without that, Steven, you're at a disadvantage because things are moving so much faster. Like if you're not moving forward, you are losing to the marketplace right now. And it, up until the last few years, it was never really like that. That's right. So I wanna shift a little towards role types, right?

Like, um, we've seen AI create for the very first time this new whole world of like, prompt engineers, and you see all these YouTube bros, they're like, yo, take my prompt engineering class and you can make 300 KA year. And, you know, and let's throw that aside and let's talk about like actual real, um, job roles. Like, I know that there's definitely engineering types of roles.

There's probably folks that have been heavy into like Microsoft bi probably have roles here where they can, they can be effective. Um, you look at some senior help desk, tier three folks that are heavy into automation, there's probably roles for them. But can you unpack what those AI types of roles look like for you guys, ATVA, and also maybe the future for you of what those roles might look like? You know, it, it, it goes down to using this every day and then adjusting, right?

So, you know, prompt engineering, just using that as an example, and this isn't like, I, I'm given a different answer to your question, but prompt engineering was probably a really cool thing for the first, like, I don't know, 45 days that the world was figuring, right? And today we're in a world where, you know, let's be clear, uh, you wanna speak to it in natural language, period. End of story. The more that you do, the more it learns, the more it responds in natural language.

You know, you don't need to engineer a prompt. You need to, you need to have a conversation, right? So I think that's no longer applicable. Right? Agree. So I can't, I can't predict the future five years out, but what, what I can say is, you know, anywhere this technology can be used is critical. And I'm not sure it's a new role, right?

So as an example, our digital transformation team, again, that's just a fancy word for application development, has been using GitHub with copilot built in now for a number of months before it would take them, you know, like several hours a day, basically repeating the duplicative code that does the same monotonous routine. But this is now going, oh, I think this is what you want. Is that right? Yep. And it's saving hours, right?

So, so it's not that we need a different developer, it's that we need these developers to know how to use the tool to support their task, right? So, and then, you know, we talked about business intelligence before. You know, I think the, the mid-market, the enterprise companies have full-time people that do this. And I think those full-time people might be the first AI czars for these organizations, for lack of a better term.

They're the ones that are gonna, you know, be the super user and figure out how to use it and take steps out of, or make sense out of all the information in a faster and more meaningful way, or at least the good ones will. Right? But then, you know, in, in an SMB space, like I mentioned before, I don't think most players have these roles.

So once this gets perfected or, or closer to perfect, uh, the, these other organizations are gonna be able to use it, hopefully with the support of an MSP who can help many clients get that going all at once. Um, so, you know, you know, Wes, I, I, I don't wanna, I don't wanna play too much of a futurist game and like guess what that job will be. Uh, but I think most jobs will have to have some basic understanding of ai.

Just like we expect people to know how to use a Word document and, and write an email Quick. Could I just ask a quick question here? We, yeah, yeah. Um, Gary, is this, is there a, you know, where does this buck, you know, with Steven's size and scale in delivery areas, is this a new one? Does this fit under centralized services? Like Probably fits under centralized services?

Especially if you're looking at the, you know, at the economics or the peak economics of it, you're gonna make these investments. You're expecting at some point, one or two things have to happen. Um, you either have to get efficiencies that allow you to have the same, you know, ratio of revenue in people, or there's other benefits to it, and you're gonna have to cost it into your model one way or the other. Got it. Sorry, Wess, go Ahead. No, it's good. Yeah.

Uh, and I was just looking what Steven said in Chad about don't being focused just on LLMs. Maybe that'll be my, um, ad hoc, uh, next question for you, uh, Steven, if, and it's okay if this is not something you thought much about, but I was looking, it's funny he mentioned that, 'cause I was looking at Gartner on this, uh, two or three weeks ago, and of course, LLMs are at the very top of Gartner's hype cycle, right? Like, that's what everyone's talking about.

But there's so much more that's sort of like following up that I'm like, I haven't even heard of. Are there any areas of, uh, AI that you guys have looked at or you have curiosity in that's beyond just what most people can, like when they think ai, they think large language models typing something into a prompt and getting answers back, or generative, you know, artwork. Is there something more that you guys are aware of and looking at?

Look, I don't, I don't think I have anything specific to call out right now, and maybe somebody in the chat can throw it out there. But like, I wanna be clear, AI has been around since like the 1950s, right? And each little piece of improvement has, has layered onto each other. So like the large language model is what's layered onto everything else that's been put out there, right? Like, you know, in, in its, its most basic form.

You know, maybe 50 years ago, uh, our, our cars started to be able, or maybe 30 years ago, our cars started to be able to tell us that still take fuel, you know, how many miles are anticipated to be left in the tank based on what their average fuel consumption is and how many gallons are left in the tank, right? Like, that's artificial intelligence, right? And then layered on top of that are, uh, more and more tools that did more, right?

Like, and so, you know, today, I think the ability to interface with all of this, with natural language and get response in natural language is the most revolutionary. So everything else we're talking about is just more stuff you can do because of that. And now the people that can do it grow exponentially versus just a handful of geeks, uh, you know, in a, in a room that can figure this stuff out, right? Like, so it becomes accessible to everyone.

And with that, you can start asking questions and coming up with ideas. And I think, I remember reading in this chat a while ago, like just coming up with mini, uh, minimal viable products and things like that really quickly. Like you can do that now. Um, so that, that's, you know, Wes kind of my take on it. Um, does anyone, does anyone have something to share? Is there something specific, uh, that, that could be thrown out? You know, obviously lots of cool stuff.

You, you can, you, I've been playing with Convert GPT, right? And it, it, it's just the ability to literally take any file type and make it anything else, right? I made an MP four into a dock, right? Boom. Like, I didn't need to go hunt around on the internet and find a hundred different, uh, converters to make that happen. Yeah, it's pretty cool stuff, for sure. Andrew, back to you, my friend Steve. That was, um, awesome. Uh, really, really appreciate all your insight.

Um, it was, it was, I think, man, Gary, what are your thoughts? I, I think this was really helpful, at least to kind of ground the conversation of where MSPs need to start. Yeah. And, and again, like most things we talk about, think about where you are in your starting point. You know, we're doing a session at SCH n Fest called A Day in the Life and taking leaders of MSPs at, at different scale levels about how they spend their day. You get a good idea from Steven today, how he spends his day.

Um, and a lot of innovation. We're talking about innovation today. It, it starts at your maturity level. And so like, assess where you are from that standpoint and what you have to do. It's a great time to do it, right? Going into a new year about where you need to be and what your job needs to be a year from now. 'cause until you do that, you won't advance these other things. Other things. Yeah. Yeah. Really, really good point.

The, and like I said, the, I started off by saying this, and I wanna end by saying this. Um, a lot of MSPs, you know, you guys all have the capability and ability to, to do something great. You know, like I said, I, I've known Steve now, um, I haven't got reconnected with, I'm so grateful I have, but for almost 20 years. And it's, it's amazing what he's done. You know, don't ever say you can't get through it. We're different or right there. We're seeing it day in and day out.

Especially today with the market that we're in, we happen to have dropped into one of the most amazing, uh, tams, uh, in history when they look back on it. And it forgives a lot of sins. But everybody today, I do believe everybody has a chance to achieve whatever goal they set in this business. Yeah, Hey, you only gotta do two things. You gotta know what to do and then you have to be able to do it. Steve, any closing thoughts? Help.

You can even get help from artificial intelligence for some of those gaps. That's right. Steve, any closing thoughts? And again, thank you. No, I'm really grateful to have joined you all today. And you know, I just say, you know, stay hungry, stay curious, keep at it play, right? Yeah. Just take some calculated risk. Awesome, Wes. Thanks as always, bud. You bet. Thanks Everyone.