Episode 29: AI, RPA, and MSPs
Listen to the Podcast
Read the Transcript
Erick and Rich discuss ConnectWise’s stealthy step into the XDR market and how to set strong, effective Q3 business growth goals. Then Rich is joined by Gerwai Todd, CEO of Pia, to explore the present and future of AI as an operational efficiency tool for MSPs. And finally, one last thing about a thrift store shopper who bought a 2,000-year-old Mayan vase for a mere $3.99.
Discussed in this episode:
Of ConnectWise, Security, and Ducks
$3.99 thrift store find turns out to be nearly 2,000-year-old Mayan vase
Transcript:
Rich: [00:00:00] And three, two, one, blast off, ladies and gentlemen, welcome to another episode of the MSP chat podcast, your weekly visit with two talking heads talking with you about the services, strategies and success tips you need. To make it big and manage services. My name is Rich Freeman. I’m chief content officer and channel analyst at Channelmaster, the organization responsible for this podcast.
I am joined this week as I am every week by your other cohost, Erick Simpson, our chief strategist at Channelmaster. Erick, how you doing?
Erick: Doing well, Rich, doing well. How are you doing? You’re you’re globetrotting again this week as we shoot another episode of the MSP chat podcast.
Rich: Yeah, I am. I’m coming to you folks from sunny Orlando, Florida where I have been attending the ConnectWise IT Nation secure event this week.
That will have been in the past by the time people are enjoying this episode. But for reasons we’ll get into in a moment, Erick.
Pretty awful this week. So Florida was a better place to be.
Erick: Good time to travel.
Rich: Let’s dive into our story of the week then. And it concerns some product news that ConnectWise announced at the IT Nation Secure event this week. There were actually a number of different product announcements they made, including an MDR service for Microsoft 365, which is interesting, but the big new thing they introduced was a product that they’re calling Security 360.
And they refer to it as a centralized command center for all things security. I’ve heard people refer to it as a dashboard. All of that is accurate up to a point. Essentially what they’ve done is create a a single pane of glass for everything cyber. So you’ve got your RMM tool, your PSA tool and ConnectWise hopes in the future, if you’re a ConnectWise partner.
Everything security related you’re doing in this one place security 360. So you’ll see all your security related tickets. Vulnerability management information will be there. EDR email security. There’s a security score that they calculate for you, like the Microsoft secure score.
They’ll offer you a prioritized list. Of measures you can implement to raise the score. And then with a click, you can actually and automatically implement those measures. So there, there is a lot of consolidated dashboard like functionality there. The thing that jumped out to me right away about it though, is that it’s not just ConnectWise consolidating data and management control over its own products.
They are also pulling in as many ecosystem partner products, third party security vendor products as possible. And so you’re going to be not just having the single pane of glass to control your connect wise security experience, but also a lot of these third party solutions, no more juggling. Multiple consoles, but also on the back end, all of the data that this tool is aggregating is going to be in one place, which means ConnectWise is going to be able to correlate it in a lot of interesting ways.
And, if you’re like me, you’re thinking multiple vendors aggregating and correlating data. Doesn’t that sound just a little bit like an XDR solution? I had an opportunity to interview Amir Kareem, who runs cybersecurity for ConnectWise yesterday. And I asked him that question and he broke out in this ear to ear grin, Erick, and said it, it quacks like one walks like one.
It isn’t necessarily an XDR solution or at least right now. But. Pretty clearly, pretty explicitly, that’s where they are headed with it. They want to create a tool that is not only your one place your single interface for anything security management related that you need to do.
But also your XDR solution which is interesting. ConnectWise has a lot of different security partners. So unlike some of the XDR vendors that they’ll be competing with they’re going to have access to a whole lot of different security companies, assuming those companies get on board.
But I’ll tell you, Erick off the record, I spoke to a bunch of different security vendors exhibiting at the show here, somewhat to my surprise, not one of them expressed any hesitation about being a part of security 360. From their standpoint, it’s it’s basically an A distribution opportunity, another route to market.
And it doesn’t preclude them from doing anything else that they want to do on their own. So an interesting, very ecosystem forward, very characteristically connect wise XDR play in security. 360.
Erick: Wow. Rich, I can think of three things that has going for it right off the top of my head. Number one, it’s got.
The term 360 in the title, right? That seems to be out of this [00:05:00] new movement. No, just kidding. Number two, it’s, it solves for, or it’s attempting to solve for what I would. You and I always talk about, platform sprawl, vendor sprawl, things like that. It creates this overlay and experience so that, while you may have a bunch of different security vendors in your stack that you’re leveraging to deliver services and solutions to your end clients as an MSP, typically we’re, we’ve got all this platform sprawl, right?
All these different dashboards and things like that. So I think it’s aiming to solve for that. So that I like that about it a lot. And then thirdly, the integration. With, ConnectWise, I would assume, Rich, guide me here, but, would make tracking consumption and billing and license management and invoicing a lot simpler as well, so it’s tagging in those two big pains other than the 360.
Not I mentioned the two big pains for MSPs, the platform vendor sprawl and then the ease of invoicing and billing and things like that. What are your thoughts
Rich: now? I think you’re spot on. Particularly in terms of the vendor and tool sprawl aspect of it, but also from an integration standpoint, it’s worth pointing out that security 360 is built on ASIO which is this next generation platform beneath pretty much everything ConnectWise is doing or will be doing going forward.
So if you are today. with ConnectWise their RMM system or their PSA system, there is no additional integration. You’re really going to need to do to connect a security 360 it with ConnectWise. It’s a one, one time and you’re done kind of integration deal. So if you’re already integrated with the RMM PSA, you are, as soon as you.
Flip the switch, integrated with security 360 as well and vice versa, because as they go out, which they will to a lot of security vendors who are maybe not part of the ConnectWise ecosystem today, part of the value proposition is when you get integrated into our XDR solution or however they participate position it, you are also going to be integrated with our RMM, our PSA and Erick.
Our marketplace because ConnectWise absolutely intends for people to be able to buy, not just manage, but buy third party security solutions through security 360. So it will be a pretty compelling value proposition for the vendors.
Erick: Yeah, absolutely. And yeah, we’ve been watching this marketplace, focus and evolution and.
And kind of the revenue streams that ConnectWise is generating from, it’s traditional platforms and things like that, but also from creating this marketplace and these relationships with these other vendors and making them available to MSPs as well. One last thought Rich that I have is this is also.
It sounds like a fast path for a vendor that doesn’t yet have an MSP friendly billing model for their channel partners. So by, by participating in this initiative and partnering with ConnectWise, that one integration that basically leads to all of these benefits. For the vendor and the MSP, they will have the ability now to leverage that more mature billing and tracking and reporting that ConnectWise brings just by being a partner with ConnectWise, it sounds
Rich: no, again, exactly right. The last IT Nation event that I went to was IT Nation Connect back in the fall last year. And one of the most interesting takeaways for me from that event was the degree to which ASIO, that platform I was referring to before is according to ConnectWise, a platform as a service offering, in addition to a foundation for their products.
Basically, a third party vendor, without even really needing permission or assistance from ConnectWise, can utilize that ASIO platform to build their own products. And in doing so, of course, they’re immediately integrated with the RMM and the PSA and marketplace, the billing mechanism, all of that stuff, all those those barriers to entry to the MSP channel that a lot of startups in particular but more mature companies as well deal with a lot of that goes away when you build on top of of ASIO.
And this just underscores how strategically important ASIO is. For ConnectWise, I, the name comes up a lot, ASIO, but I don’t know that people fully appreciate how big a deal it is for that company and, it took them a few years to make it but it really positions them to do some interesting things in the market.
Erick: They’ve really been building on it for the last several years. That was an acquisition they made and then they began integrating it and adding to it. Is that correct?
Rich: Yeah, I believe the work, I don’t know if work on ASIO predated or postdated the continuum acquisition, but it was right around that time, certainly, that they they started doing it.
Yeah,
Erick: And it’s, it is interesting that, we’ve been seeing this kind of parallel Almost like a business model, revenue [00:10:00] model, revenue stream, the platform as a service, the marketplace, these other things that, you know a little bit different than Neckwise’s competitors out there. So it’s a very different vision and approach.
So it’s very interesting to watch, how they’re enhancing those capabilities and growing Partners, businesses, along with ConnectWise’s core
Rich: business. This is actually something I pointed out in the column I wrote back after IT Nation Connect which is that in, in regards to the things you were just talking about, Erick, ConnectWise looks more like AWS or Azure than it does like Kaseya or Enable, right?
They’ve got this public cloud infrastructure that third parties can tap into and build on all, at this point, they can just do it at some point down the road. All they’ll have to do is pay connect wise some money to have access to that platform. And then there’s a marketplace as well.
It’s, it is a very different approach than than you see from the other big names, top names in the managed services space. Yeah.
Erick: And is, so is it broadly available now? Is there a launch date? When is this, when can partners participate?
Rich: There is an early access program opening up in July.
It was a little unclear to me how Broad, that early access program will be, if anyone will be able to get into it or if it’s invitation only, but they’re going to start making it available to partners next month. And then my guess is maybe right around the time of it nation connect again is when they’ll be ready to just roll it out to everybody.
And if, I got hints from Kareem that there, there will be a lot more functionality rolled into the tool between now and then too. Great reporting. Thanks Rich. You bet. Let’s move on to your tip of the week Erick. And it’s early in June as we’re recording this year. Can we really be talking about Q3?
Erick: It’s Q3 and tip of the week is establishing your Q3 business goals. I rewrote it rich because I said let’s talk about your second half 2024 business goals, because yes. In a couple of weeks as we record this it will be q3 and then very quickly after that q4 And you know what happens in q4, which is a lot of holidays a lot of time off So the tip of the week is hey Let’s get our q3 and for extra credit kind of rolling into q4 business goals From an operational, financial, marketing, sales, service delivery, solutions, platforms.
There’s a lot to talk about here. And as, as you and I always say, Rich, it’s hard keeping up with how fast the industry moves and all these developments. There have been massive developments from the, the top platform vendors we’ve been covering some of the top other vendors that we’ve talking about technology chat, GPT, there’s all kinds of things.
That are in motion right now. It’s probably I don’t know. Rich, would you say or agree that this is probably one of the, largest kind of newsworthy change industry change channel change periods that we’ve seen recently?
Rich: Yeah. Oh, absolutely. And in, you’re reminding me of conversations I’ve had at the conference here where people have said, this is this, and they’re alluding to AI, this is akin to what the cloud did to this industry a decade ago.
It’s one of those moments.
Erick: So if you’re an MSP and you’re going through you set goals at the beginning of the year, we always talk about that. And now you’re at this mid year point, this is let’s take the temperature, Of the strategies that are working for us and the strategies that not working so well.
And think about pivoting, right? Where, a business plan is not an immovable anchor rich that says, no matter what, we are gonna execute this business plan. A business plan. And a growth strategy is, it’s basically a framework, a guidepost guidelines. It’s not meant to be taken literally. If something isn’t working, we shouldn’t just keep chasing.
Something that’s not going to pay us returns and dividends. And as we just talked about, there are things that are changing that we may not have any control over other than how we respond to them. So are there opportunities out there that we can take advantage of? We’ve talked a lot about AI. We’ve talked a lot about, changing our our client ideal client profile.
We did a show. A couple of episodes ago, Rich, where it was identifying and adjusting your target, ideal client profile. Things have changed. Today’s buyers are not the buyers that we were working with pre pandemic, or maybe even five years ago or 10 years ago. These platform changes, the internet changing, the cloud changing, security, chat GPT.
It changes everything. Now it’s time to take stock, take some quiet time, take a few days and just read the tea leaves. Where are the gaps in your business strategy now? This isn’t rip and replace stuff. This isn’t revolution. This is slight evolution. If necessary, if you’re, if you’ve already planned for this for this year and things are going the way you want, then I would say [00:15:00] continue executing on that strategy.
Don’t add new things. If you haven’t yet realized the full ROI and the goals that you’ve set for yourself for this year, a lot of us, rich as entrepreneurs, we get excited and we see some bright, new, shiny object where you think, Oh, everything is working. We can take something more on, but, being an EOS shop at channel mastered rich, we know the danger of trying to do too much and adding too many rocks and goals and milestones to a plan.
Let’s make sure we can execute the plan successfully before we begin adding things. And we’re going to talk about issues that come up so that we can solve and actually achieve the results that we had planned out when we first developed our strategy. But like I said sometimes you gotta, take a pause look at what’s going on in the industry.
And we say this a lot too, borrowing from the great Wayne Gretzky, right? Skate to where the puck is going to be, not where it is today. So where do you see your business? In Q4 of this year, where do you see it in Q1 of 2025 and how can you begin planning internally for that outcome? Are there things that you have been pushing off because they’re outside your comfort zone?
Yeah, I know we need to do that, but I’m just I’ll keep pushing it to the bottom of my list. Now it’s time to start reprioritizing. If you have things that you’ve got to get done in order to set yourself up for success, In q1 2025 and finish off q4 of 2024 successfully How’s the time to do it?
Rich: The other thing that comes to mind, it’s a little bit more near term than some of what you were talking about there, but here at the conference this week, I had a chance to interview Peter Kujawa, who runs the service leadership arm of ConnectWise and service leadership has this giant database of MSP benchmarking data.
And one of the things that they’ve pretty consistently found Best in class MSPs from everyone else is the best in class ones are really disciplined about budgeting, meaning sales budgeting, they go into the year with a sales plan. They track it every month. If they’re falling behind plan, they spot that very quickly and they make the adjustments needed.
And as we head into the second half of the year and people are taking the pulse of where they are and and what they need to focus on in the second half of the year, if you are not one of those MSPs, who’s really good about sales planning and budgeting this is a good time to see how you’re doing compare how you’re doing to how you expected or hope to do, and then think about what you’ve got to do differently in the second half of the year, both in terms, maybe of, Of driving some more marketing and sales activity, but also maybe cutting back expenses.
If, you want to keep your profitability where it was, and sales have been a problem for you in the first half of the year, then maybe spending less will make sense as well as earning more. But that, that’s one of the many good things to think about as you think ahead to the second half of the year.
Erick: Yeah. Agreed, Rich. There’s some big big, Decisions that need to be made sometimes. And I recommend making the decision quickly, go with your gut instinct, analysis paralysis is not going to help you re restructure or re strategize, but just be honest with yourself and be candid.
And, sometimes you’re going to have to make a tough decision. You’re going to have to say, yeah, we’re not doing that anymore. Or we’re going to, release this particular client back to the industry because they are just sucking too much of and we’re not, hitting our margins or whatever that may be, so it’s a mid year.
Time to big stock.
Rich: We’ve touched on AI and co pilot and chat GPT in the conversation here and obviously most MSPs. Are already thinking hard about how they can utilize AI to make their practice more efficient. A lot of people are doing that already. RPA is is and has been a big point of conversation in the managed services world for a while.
We’re going to take a quick break here, Erick, but when we come back on the other side, we’re going to be joined by Gerwai Todd. He is the recently appointed CEO Pia one of the leading makers of RPA and AI software specifically for MSPs, he’s going to help us sort out the differences and similarities between RPA and AI some of the low hanging fruit opportunities to apply those today, and then also paint us a picture of where both of those technologies will be going down the road.
So interesting stuff, stick around for it. We will be right back
and welcome back to part two of this episode of the MSP chat podcast, our spotlight interview segment, where we are pleased and privileged to be joined by the CEO of Pia. His name is Gerwai Todd and Gerwai, thanks so much for joining us.
Gerwai: Yeah, absolutely. Great to see you and be with you.
Rich: I want to spend this time talking with you about two topics that are red hot, obviously in the world of managed services right now AI and RPA.
But before we get into it, just tell me a little bit about you and your background in the industry, and then maybe a little bit about PIA.
Gerwai: Okay. Great. I appreciate it, Rich. I [00:20:00] guess a little bit about me. Yeah, I’ve been in this business game for about 30 years. I’ve been 2008 and about 10 ish now, actually, how many years in was I at that point?
I was 14 years into my career. I got into the MSP space by being one of the early executive members at ConnectWise in 2008. So I joined the team, Artie brought me on board, Santo Canone was there at the time. As coo and I joined to head up our sales organization I did that for a couple years and then moved on to doing the m& a work that we Were really pursuing around 2010 and really putting a lot of energy and effort into so I managed through some of the transactions we did there causal Screen connect helped a little bit with the lab tech deal early on.
And yeah, I had a lot of fun chart tech too, along the way. So a lot of fun there and then exited in 2019, went to Mubarak came along and then started building my own companies in 2019. Founding Treafy and then also one of the co founders in TimeZest. And then now also in addition to running TimeZest, I’m also running PIA.
Rich: So you, as, as much as anyone we’ve probably had on the show you really know your way around the managed services world given the background you’re talking about there. And now you are running a company that does both RPA and AI. Those two. Product categories, in my experience, tend to get rolled together under the label hyperautomation sometimes, but as much as they’re related to one another, they are distinct.
So talk a little bit about that difference, the distinction between AI and RPA from a managed service provider’s point of view.
Gerwai: Yeah, it’s great. Great question there and an important thing for us to all frame out I think RPA can really be summed up pretty easily. It’s, it is automation. It is automation generally triggered by things that are variables that we have access to, or triggers and events that we’ve actually been doing for quite a long time.
So if you think about, PSAs, we’ve had workflow rules in those PSAs for a long time, you could really look at that as an early form of RPA. You look at RMM technologies and triggering scripts off of certain events. That’s a form of RPA also. What I try to think about is the two differences are RPA is triggering, maybe a little bit more modern triggering or advanced triggering is happening in RPA these days.
But the reality is with AI, you’re effectively adding on an additional brain into that environment. And so with AI, we’re able to do some things that are quite interesting these days. So NLU engines, natural language understanding engines, can do some really phenomenal work. The way I like to try to frame that one, and we’re talking to technology people, so everyone gets it.
But I don’t, I’m really not sure if they’re clear on how we go about building the technology side of this. So what I like to think about in the example I have for people is imagine your home automation system. And you think about Alexa or Siri or anything that you might be using. Those AI models, okay, is effectively what it is.
It’s an NLU model that has been trained on one thing, how to lock your door, open the garage, close the blinds, turn on a light. That’s all it really knows. Knows how to tell you the weather. Actually probably knows a bunch of things that we don’t even know how to command it to do, right? Because it’s a discovery process of figuring that out.
But effectively, what NLU models allow us to do It’s to take an environment like Alexa, wipe it of the brain that it has around home automation and program it and train it with a brain that’s specific to a particular use case in the I. T. Space and specifically for M. S. P. S. Where that really comes through is when you look at service tickets when you think about a service ticket and all the various types of issues that can come in.
With that NLU engine and the model built right, it can identify a printing problem being different from a VPN problem and do things to help that service issue along with the JIRT, whether it be, setting types of type and item on the ticket, whether it be trying to do some triage, which is what we were working out at Triify that I founded back in 2019 or at PIA where it’s doing identifying the issue.
And then running an RPA process in the background to automate the resolution of that particular problem. Those are all the great things you can do with AI because it’s been able to take natural words that we would use. English words, crunch them, and then come to a conclusion about what is the intent of that particular phrase, or ticket description, or content of a ticket.
Rich: Before we go a little deeper into AI, I want to talk a bit about RPA just to help people understand where the [00:25:00] opportunity is there. What’s the low hanging fruit for a typical MSP around RPA if they’re just getting started?
Gerwai: That’s a good question. There are a few tools out in the market that are working to solve that particular challenge.
We are also seeing big PSA companies like SayIt and ConnectWise looking to try to solve that challenge. You can start at your PSA provider, see what they’re doing to stitch together their applications. There’s some other products on the market that, that are trying to solve this problem also.
And I think, just, there’s plenty of solutions out there, quite frankly, that are out there solving the RPA challenge.
Rich: Are there any particular barriers to adoption or issues, challenges around RPA that are maybe inhibiting the degree to which MSPs take advantage of it?
Gerwai: No, I wouldn’t say so.
It’s much like what we’ve had to deal with the evolution of RMM. And that you effectively have to have engineering teams that can write the scripts, that can build the processes. And some of the applications out there are trying to help along by building, components and code that can be reused to some degree.
But in the end, you do have to stitch some things together. But it’s been work that, technology teams have been doing for a long time. And there’s really no major barrier in that regard.
Rich: So I, turning back to AI I, I take it as a given MSPs will be doing a lot more with artificial intelligence a year, two years, three years from now than they can today.
What is the state of the art today in terms of how an MSP can apply AI in their business?
Gerwai: Oh, that’s a really good question there, Rich. I’ve been working on taking AI and trying to relate it to MSP challenges for over five years now, and what I would tell you is a big reason why I’m at PIA is because I feel like the team at PIA has done a pretty phenomenal job with solving this challenge.
Because usually what the challenge is that what AI has been able to be the AI engines that we’ve seen built to date. Are generally doing some rudimentary kind of work. So let’s use an example. There’s companies out there that have been able to solve Reading a ticket and putting the types of type in item on right?
But that’s relatively lightweight kind of work and really just In my opinion, it doesn’t take someone all the way through the journey of what’s necessary. And so when I chatted with the PIA team and I really looked at the technology and looked at what they were doing behind the scenes, it was quite interesting to me and compelling around the following, which is they have built their own AI model to identify ticket issues.
And then they had built the automation behind it. To fix the problem in a lot of ways, it’s like the documentation that we’ve all worked on over all the years to add process to our business and ensure that when we set up a new user, as an example, that we follow a 10 step guide and that there’s multiple things that we do along the way to ensure that user gets set up properly, that didn’t usually exist in our documentation solutions back in the day, it used to be in word and SharePoint and all the kinds of places, right?
One note and whatnot. Yeah. But today, if you think about it, you can take that process document and put it into an automation flow and have it consistently execute that process consistently with every single tech that comes into the business. So if you look at long term techs that think that they know every single step that needs to be done in a process, right?
Sometimes they’re not always looking at the documentation when they should be looking at the documentation or something’s changing the documentation and they should be incorporating it in their process and they miss that. So the consistency that this drive, the example that we have is that we have had business owners and service managers come to us and say, Hey, I’ve got great techs.
They’re phenomenal at what they do, but they all follow this one process just a little bit differently. They’re doing fine. They’re not missing anything major. But the little things that we would want them to check sometimes get missed and is a little bit inconsistent. So I like to think of Michael Gerber and E Myth around the focus.
Are you familiar with what Michael Gerber has written and his philosophies around this?
Rich: Yeah, absolutely. And in fact this will be familiar to you that very first generation of MSPs 20 years ago are the people who introduced me to Gerber because he was so Influential to them.
Gerwai: Yeah, like already had him come speak At one of our IT nations, right? And so why don’t you share with the audience what it’s all about and what the philosophy is there.
Rich: The big issue and it’s a phrase that I imagine some people in the audience are familiar with without knowing why they’re familiar with it.
But there’s a distinction between [00:30:00] working on the business and in the business. If you are the owner, the founder, the creator of a managed services business, job number one should be. Building and growing this company and making it bigger and better as opposed to closing tickets.
Gerwai: Yeah. And so I think one of the key principles that Michael has pushed forward is you need your systems to know what the process is for your business. The more you can build into your systems, the better off you’ll be in onboarding new team members. The better off you’ll be about consistently delivering a service, right?
Rich: Yeah, absolutely. And and so the idea there, and this is really interesting because this is an issue that’s come up in conversations with a lot of different vendors over the years. And I’m thinking of BitTitan in particular. But this idea that, there is one ideal way to get something done.
That should be your company’s way of doing it, and every tech should be doing it exactly that way every time, but it’s very difficult to do.
Gerwai: Absolutely, and great techs, they’ll hit all the important things, but the process changes. Arnie had a great thing that he taught us, which is the moment you get done writing a process document, it starts rusting.
And at some point, you’re going to go back and rust proof that and knock the rust off of it. Meaning over time, our processes are always evolving and we’ll forget to look back at that process. We’ll forget to refresh them as much as we should. And before you know it, we’re out of sync with the way best practice would be for a given practice.
In the end, what we’ve been on a mission to do at PIA, really the team way longer than I’ve been here, It’s been on a mission to put that into an application that then consistently executes the same steps every single time, no matter who is executing the script. And effectively, if you think about it, that’s what RMMs have been doing for a long time.
That was with a mission that we had in the PSA with workflow rules is to consistently ensure that when a process needs to be applied, it’s getting applied consistently and potentially in real time when that event is happening.
Rich: So if we’re talking about processes getting rusty over time is one of the advantages of bringing AI into the work that technicians do is that it’s watching and learning and essentially rust proofing those processes to update them or optimize them over time.
Gerwai: I wish it was that smart.
Unfortunately one of the biggest things we have to do is we have to train these systems. So there’s the aspect of training the AI intelligence in the model, which, if you think about the models exist in this world, going out and leveraging chat GPT, as an example, isn’t a good answer for an MSP use case.
And the reason why is because it’s got loads and loads of information that’s unnecessary, right? So we go back to the example of Alexa. Alexa is a great example of this because Alexa really only knows home automation. Or how to help you get the weather or how to help you maybe, send in Add something to a note or a list or help you order an item Alexa doesn’t know how to tell you about everything out on the internet quite yet, right?
OpenAI does, and what they’ve done with ChatGPT does, but that’s overkill for the purpose of this use case. And quite frankly, it also has a lot of processing and compute that’s really problematic. When we go back to that scenario, what we are doing in this space is looking to wipe the brains, the easy AI models of everything that they know, and train them on one specific use case.
And in our case, we’re looking at the service ticket issue. We get a volume of service tickets into an MSP on a daily basis. Some of them can be guided along the way. And so it starts with identifying the problem, being really confident that we’ve got the right problem in front of us and the proper automation solution to it.
If we don’t, we won’t touch that ticket. We won’t effectuate the resolution of that ticket. But if we do have it right, Then that’s where our automation kicks in. And our automation has to be curated by team members. Our technology team is working to curate that automation script. We do allow MSPs to modify that script so that they can add in, their PowerShell script that they may want to run or extensions.
We’d like to call them extensions. To our automations that we will naturally run so that we can fit and be customized to their particular use cases, but effectively those automations that get triggered once we’ve identified the proper issue do need to be curated by a technology team that knows how to effectuate all the systems involved with making something happen.
When we have a printer [00:35:00] problem. We have to have the technology code written properly to go diagnose and triage that printer problem. And unfortunately AI is yet at the point in which you can just go tell it to do that. It doesn’t entirely know what systems it may need to effectuate, how to go about it.
So we guide and we build all those automations ourselves. Our goal is as a team is our mission is to make it so that when you implement our automations and our AI engine. It’s much like configuring your phone or configuring your PC. You go and change settings more than you have to go write code and write scripts to make things happen.
We have built a model where the AI model applies across MSPs. We’ve got all those use cases because they’re very similar, right? When one MSP is having a printing problem, it’s likely similar to the MSP down the street or across the country or across the world, and so we’ve been able to standardize that model.
And then when it comes to the automations, we’ve been able to standardize and curate those too.
Rich: The pace of innovation and advancement in AI just broadly has obviously been astonishing in the last year and a half. So you’re talking about what AI can do for MSPs today, as you look down the road, a year, two years, three years.
What is it that you are. Hoping or expecting to be able to do as AI grows more and more capable and sophisticated.
Gerwai: This is a great question. I always try to temper my expectations because we all, you’ve been around technology a long time, rich, as well as anyone it, it can, we can really sell a dream quite easily and we can really envision this amazing place, but it takes a lot of work to get there.
So I like to temper expectations. I think AI is going to continue to get better and better about sorting out one example of a challenge that we all face today is if you can imagine consuming a service ticket and trying to derive that there’s one particular issue in that ticket, that’s very, it’s a little challenging for AI to do that these days is we call it an utterance, which is what’s being provided.
To the AI model and the engine. And then out of it comes intense, like what we have this particular service issue. What is the intent of this service problem and what do we need to do about it? And how many steps could potentially be in that? It doesn’t do a great job today in sorting out that client could email that they’ve got four issues in that particular ticket.
So our job today as technologists is to build it in a way where it can be co-pilot. Which is a great term Microsoft had got it right when they said co piloting, right? You have to have a human team member alongside these models, guiding them along the way, confirming that they’ve got it right, until we get confidence scores up significantly, and that confidence score is going to grow based upon the data we consume.
And over time, we’re going to get better at being confident we know a particular issue. And over time, what’s going to happen is we’re going to get much better at being able to parse out four or five issues from a service ticket versus really today, if you think about what AI can do today, it can effectively say, I know this is the most important service issue in this ticket, but the rest of them, technology team member, you’re going to guide me through which automation I should run next.
I figured out the most important one, you’re gonna have to go sort out the other two or three that are in here and you can force me into an additional automation executions to solve those additional IT challenges this client is having or potentially completely exit the AI and automation schema and do it manually like you used to do the old days.
So that’s where that, that human technology team member is so important is to guide the AI model or engine through what it can do and then taking over when it’s hit its limits or taking over when we need that cognitive ability from a team member to really guide it.
Rich: Open ai, Andro Google.
They’re all, a lot of what they’re doing these days is in that area of mul multimodality, basically combining audio, video and text and allowing the ais to both read and produce in any of those formats. I can’t help but wonder is that something that at some point could be useful to a, an AI tool for MSPs in terms of being able to maybe.
Look at an image of something that you know an end user is reporting or having trouble with and understand that or just consume a phone call potentially from an end user and respond to that.
Gerwai: Yeah, that’s a great question too. It will, I do believe it will get there.
The technology is growing so rapidly, but you do need technology teams to shape it. and curate it and train it properly and that is [00:40:00] The mission at hand at pia is that we ensure that we are shaping these things it’s very easy. I, I was a little early at Triafi and what we were working on triaging tickets in 2019, but the use case was working.
It, the technology was working in, in, in 2019 and we have come light years since then, so I’m confident we’re going to continue to knock out additional use cases, but the key is, it’s taken really several years for us to get to a point where we can apply it in a really meaningful way. Where the ROI is clear, where we’re really truly able to save technology team members true time and do things accurately.
It’s to do that in a production commercial environment is a challenge. And so all these ideas and capabilities, I think each one of them is going to have some degree of work. Quite a, some of them quite a bit of work to get them to fit into our particular use cases in the IT space.
Rich: I’ve written a lot about AI for MSP is obviously in the last year plus in my blog channel haul, like I’ve spoken with a lot, I’m mostly thinking here of RMM PSA vendors, but I’ve spoken with a lot of companies that are incorporating AI into their products and specifically with respect to tickets and process reading and processing and closing tickets, I would say very broadly, there are these two kind of schools of thought there are vendors out there right now.
Are ultimately driving towards. And some of them think they’re very close to that right now, where you just turn the AI loose on the ticket. And assuming it, it thinks it knows how to close the ticket without any assistance oversight from a technician, you just let it do that. And there are others I’ve definitely spoken to, folks like yourself, CEOs of companies in this management space who are like, we, no matter whether or not it’s possible, we will never let the AI handle a ticket start to finish without a technician.
Seeing what’s going on and at least clicking a button saying yes, go do what you’re proposing to do what’s your thinking about the pros and cons of autonomous AI in Dealing with managed services tickets versus keeping the technician in the loop
Gerwai: Yeah What I would say there is it’s very important that we’re thoughtful about everyone’s Risk tolerance around these kinds of decisions and that we architect our technologies to support that very level of risk willingness and what I would tell you is that if you haven’t been in ai and working with it, it can be quite daunting and concerning that Oh what if it goes haywire?
But there is a there’s an aspect to these models It’s called a confidence score. All of them have it. Okay, and when that confidence score hits certain thresholds, I can tell you confidently and I’ve seen it work my own eyes. It will get it just as right as when you have a boolean value in code doing a true false, right?
So what I mean by that is that when that confidence is there, it I’ve found AI will not hallucinate at a high level of confidence. It will consistently achieve the goal. And part of this, in my opinion, is being able to, again, co pilot, to watch, to see the data, to be able to, when we expose the confidence level of a given transaction, right?
So meaning a service ticket that was guided by a tech, our technology got It ran, it’s, we ran our AI model on it. We identified the service issue, the tech authorized it, said, yep, you got it right. AI bot. Then we ran our automations. When that gets to a very high degree of confidence, it doesn’t hallucinate.
It can hallucinate though, when that confidence level is mid level or lower. And, there’s various degrees of work that has to be done around raising that confidence. A lot of that comes to curating the AI model to be able to clearly parse, that a print, one printing problem that is a printer offline is different than a printer paper jam.
And that we can parse those apart or that if the end client hasn’t provided us enough information that we know how to manage through that and not presume it’s one or the other. So that is the mission at hand, and quite frankly, it takes a lot of work to ensure that our technology is not hallucinating.
That is the mission with AI, because we’ve all seen it, right? We’ve all seen how you tell it something, and before you know it, it’s gone. It gave you the weather for, Sri Lanka or something like that.
Rich: Yeah. Hallucination is probably the best known, most familiar concern about AI in managed services and [00:45:00] beyond certainly another.
Widespread concern about artificial intelligence is data privacy. One of the advantages of working with a company like PA or other companies that are doing AI for managed services is you’re not uploading everything. The model knows to open AI or. Whatever. But for the MSPs in the audience who are a little bit concerned about data privacy, data security other risks, maybe potentially associated with with AI and using it in their business what do you have to say to those concerns?
Gerwai: Yeah, look, at the end of the day if you think about AI, it’s actually still quite rudimentary as it relates to the models that we build in the environments, we build them in. They can’t go power off the PC. They can’t go do anything. You haven’t told it to do. You have, if you haven’t programmed it to take a task, it can’t teach itself to do another task.
You can’t force itself to do another task. We are in control as technology teams of what it’s capable of doing and what it has access to. So at the end of the day, it isn’t a whole lot different than what we’ve been doing since the sixties of writing programming code, whether it’s COBOL or any other language that we’ve.
And so we’ve been working in over all those years. We control what it does and we control what it has access to. It can’t just grant itself access on its own. And so at the end of the day, what it is just an additional tool in our tool belt. Our tool belt needs to have the ability to write Java code, the ability to write, PowerShell scripts, the ability to access things remotely.
And at the end of the day, we need to find ways to leverage this technology to its fullest, to help bring that labor costs down that we’ve got in these MSPs. Unfortunately, 80 percent of the cost in MSP is generally, payroll on people and all the things that go into that. And the help desk is a good bit of that, right?
The help desk is a significant part of that. You can outsource it, but then you lose touch with your clients. And at the end of the day, your technology team members have very important tasks that they need to complete on a daily basis. They shouldn’t be in there trying to resolve a print problem, 10 times a day.
We need to help them get to those kinds of issues quicker and easier, so they can really focus on the really challenging problems, like not resetting a password manually, right? So those are the things that we think of on a daily basis. How can we help make those technology team members.
That we’ve got on board more effective, right? Maybe not have to hire another technology team member until we get to 300 endpoints versus 200 endpoints. What can we do to be just more efficient? So we built a healthier MSP at the end of the day.
Rich: Yeah. You’re getting at another issue that comes up around AI more broadly, and it’s economic impact more broadly, which is this question of will AI put people out of work or empower them to do more.
And it sounds like what you’re saying basically is from your point of view, AI Could delay the speed with which an MSP increases their staff and hires technicians. But the primary opportunity here isn’t necessarily to get rid of techs, but to free up time they can use for, more important activities.
Gerwai: Yeah, I think the classic challenge we’re always going to have and at least in the, into the foreseeable future, We don’t have enough technology people to do the work. That’s been a common problem. Have you had other people on your podcast that have shared how hard it is to find talented technology team members?
Rich: Yeah, absolutely. It’s been a huge issue in the industry for years already.
Gerwai: And so what have people turned to? They’ve looked to offshore team members. They’ve looked to outsource of help desk. There’s all kinds of solutions that people turn to, to try to solve this particular challenge.
What I believe is we’ve got to empower our great team members to be more efficient. Like I said before, we get a service issue, a ticket that can have a random number of issues in it, right? Because clients can email in or submit in whatever they want in a lot of the ways that they submit their issues.
And before you know it, we are needing someone or something to work through that. And AI is a good ways away from being able to parse through all of those various issues. What I’d say is that we need those technology team members freed up from the day to day really administrative stuff to work on the higher value stuff.
And quite frankly, grow in their career and do more important things than resetting passwords all day.
Rich: Now you’re going to have a particular point of view on this question. It is maybe a predictable point of view, but it’s also a very valid one. To the extent that ConnectWise and Kaseya and Enable, Atera, Ninja they’re all building AI into their management platforms.
That’s one degree or another right now. And so an MSP has a choice [00:50:00] basically between getting their AI from whoever they’re getting their RMM software from, or getting it from a company like PIA that maybe specializes specifically and exclusively in in AI based automation for MSPs. I, how would you advise an MSP?
to think through the pros and cons of those two alternatives.
Gerwai: That’s a good question, Rich. I think going back to 2008, back in 2008, there were about five or six RMM solutions. I think there’s plenty of opportunity for lots of companies to innovate around this arena. And I think that at the end of the day, there’s going to be some great solutions that come out from everyone.
We’re all going to have different ideas about how to solve this particular challenge. We are looking to be platform agnostic. We’re not so concerned about what you’re using for your PSA. What we’re focused on and laser focused on is the service desk use case. When service issues come in, how many of those can we help co pilot with attack?
How many of those can we help them get to resolution quicker, consistently? And, quite frankly, at this point in time, from what I see, We’re the only company laser focused on that one problem. And unfortunately, I don’t think we can expect all innovation to come from one company or the big companies.
Innovation comes from all these really interesting nooks and crannies of the industry. And at the end of the day, great teams focused on one challenge and only one challenge will do amazing things.
Rich: All right. Drew, I thank you so much for joining us on the show and exploring these issues with us.
For folks in the audience who would to maybe get in touch with you, learn more about you, or learn more about PIA where should they go?
Gerwai: Yeah you can actually email me, so Gerwai. todd. @ pia. ai. I’m happy to take those and then just check out our website at pia. ai.
Rich: All right, fantastic Once again, thank you very much dear way.
Todd ceo of pia folks, we’re going to take a quick break here when we come back on the other side I will be rejoined by Erick he and I will share some thoughts about this very interesting conversation here Have a little fun wrap up the show. So stick around We’ll be right back. All
Welcome back to part three of this episode of the MSP chat podcast. And thanks once again to Jerweh Todd of PIA for joining us to talk about AI and opportunities around AI for managed service providers. A couple of things Erick, that that sort of jump out at me. First of all it was refreshing and kudos to Jerweh.
There was that point in the conversation where he made a point of emphasizing that he likes to set reasonable expectations. He doesn’t like to over promise what AI can do today, where, what it’s going to be able to do a year or two from now. That is far from the norm, I would say, in the AI world right now.
There are very big promises quite regularly being made. So I congratulate him, for making it a policy at PIA. That they’re going to be straight and and realistic with you about what they can do now and what they can do going forward. And he had some very interesting stories to tell about where the improvement will come in the next few years.
I also thought it was interesting, right at the end of the conversation there I had to bring it up that, that the ConnectWise’s and Kaseya’s and all these companies are getting into AI and management themselves these days, how does an MSP go about am I going to go with a specialist or something that’s just bundled into the products that I’m already paying for and, he made some good points.
If you go with a third party like Pia, there are platform agnostic, which for some MSPs can be an advantage. And then there is just the fact, ConnectWise as a, for example, they’re doing a lot around AI. Very deep pockets, a lot of money, a lot of resources. But they are covering a whole lot of waterfront.
They’ve got a lot of things going on at any given point in time. And all PIA does basically is AI for MSPs. And that might or might not, we’ll see over time, that might pay off in terms of the pace of innovation. I thought that was interesting too.
Erick: Yeah, I, me too, Rich. And I think that, that, especially that last That’s kind of part of the conversation where we’re talking about.
I would liken it to the, like that Swiss army knife versus the scalpel conversation. You can have something that does a lot of things. Okay. A platform that, that, like you said, there’s a lot going on, a lot of waterfront they’re trying to cover. Or you can go with something that is very purpose built for a particular application.
and it reminds me rich of when I’m working with partners and we’re determining how to bundle and price different tiers of solutions, that value ladder. And some partners will say in order to keep the costs down and attract as many prospects and clients as possible to sign up for my entry level managed service agreement, for instance, I’m going to use [00:55:00] the, antivirus and anti spam that comes bundled with my fill in the blank tool here, right?
It’s not the full featured product that you could get if you bought it, independently of what’s bundled in, but It’s probably good enough to get these part these clients and give them better than what they’ve got. And then give me an opportunity to upsell them over time. Then as we go up the value ladder, let’s say to the next tier or bundle of services, or maybe even the highest tier now we’re saying, okay we’re going to replace three or four of those kind of bundled services that come with something else that you’re selling in these entry level.
Agreements with something that is, more high level, right? More specific, maybe even from a different vendor, because, it brings that additional level of enterprise protection or ease of use or reporting or value, et cetera. So I look at it the same way. I think we’ll see And we are, talking a lot on the show, rich about what these other vendors and, marketplaces and platforms are adding AI into their platforms and solutions to do for MSPs.
And I look at it the same way. Maybe it gets you some of the way, but if you really want to get serious and value. The unique value proposition that the provider that is, specifically focused on creating that scalpel for the surgeon rather than a Swiss army knife like PIA, then you have that option.
Rich: Interesting theory. I like that. One of the many things to watch is this is this market evolves to see. If there is that kind of value play, if you will that differentiates those two options for getting your AI functionality. Good stuff. Folks, it leaves us with time for just one last thing on this episode.
Erick, If you’re like me and you’re on vacation or you’re tooling around town on the weekend, maybe you stop into a thrift store somewhere. I think everybody who does that occasionally has the fantasy where they find something that turns out to be really important and valuable. You buy it for a few bucks and it turns out to be, a Jackson Pollock or something crazy like that.
We’ve got a story that broke recently that may be the ultimate example of this. Somebody goes into a thrift store and Finds a vase on the shelf there that looks pretty cool, buys it for 3. 99. And long story how this came about, but they subsequently discovered this vase was a 2, 000 year old Mayan vase.
Forget about an expensive find in a thrift store, Erick. It is literally priceless. Now the person who made this find did the right thing, got in touch with the Mexican government. This is a priceless artifact in Mexico and they returned it even though they could have obviously sold it for an immense amount of money.
But it turns out that just wait last year, last December. And I don’t know all the technical details around this, but somebody else in a thrift store found a vase that they bought for 3. 99 and discovered it was worth 107, 000. Not a a national artifact. This person went out and sold the darn thing for 107, 000.
You
Erick: know, Rich you never know what you’re going to find in those thrift stores. Sometimes I watch some of those There’s some TV shows, right? They say, bring in your, bring in your heirloom or whatever and get it appraised and all that. And I, I know for a fact that there, some of that’s game, like they know that there’s some, hugely valuable thing and they okay, we know ahead of time that it’s going to be worth something.
They may not know what the owner bringing in, but we did our research. We know what it is. And looking for that big aha moment, that big reveal. I wish I had thrift stores around here that You know, had relics and stuff like that. I’m just happy finding, some, vinyl in good condition, to play on the turntable that, we are now finding new joy in, we’ve gone away from vinyl. We’re back to vinyl, but finding some of these old, classic albums that we like on vinyl is is a great find for us. But vases, I’m gonna have to keep my eye out, Rich. I’m gonna start eyeballing those now. And then when I go into some of these secondhand stores.
Well,
Rich: folks, we are out of time on this episode of the MSP chat podcast. Thank you so much for joining us, Erick. And I will be back with another episode in a week. Until then please remember we distribute this show, both in audio format and video format. If you are listening to the audio version of the show, but you are curious to check us out on video, you will find us on YouTube.
Under the name MSP Chat on YouTube and if you’re watching the YouTube version of the show, but you’re into audio podcast to go to wherever it is, you get your audio podcast look us up this MSP chat. Again, you’re gonna find us wherever you go. Whichever direction you go, however you choose to enjoy the show, please subscribe, rate, review.
It allows other people to find the show. More easily and enjoy it just as you do. This show is produced by the great Russ Johns, who is part of the team with us here at Channel Mastered. He [01:00:00] can do a show like this for you. You wanna learn more about Russ you wanna learn more about Channel Master and the many things that we can do for you.
If you’re a vendor in the MSP channel, go to www.channel master dot. Channel Master has a sister business called MSP Master that consults with MSPs. You wanna learn more about that business, you will find all the details you need at www.mspmaster.com. So once again, we thank you for joining us. We will see again on a week.
Folks. Until then, please remember, you can’t spell channel without MSP.