What Does AI Mean for MSPs and IT Service Management?

Last month, at N‑able’s Empower partner event in Prague, Chief Security Officer, Dave MacKinnon conducted a roundtable with executives from MSPs from the UK, Germany, Poland, and South Africa. The topic centered around artificial intelligence (AI) and initiated dialogue around whether they were planning to embrace AI and their concerns around the impact it could have on technology as a whole. After the event, I sat down with Dave to discuss some of the insights gained from the roundtable.

Q: What were some of the key concerns you picked up on?

Dave MacKinnon: I think everybody shares a certain level of concern around AI. It was very interesting, especially with me being American and having a chance to talk with folks who are European and South African to be able to gain an understanding of their views on not just privacy concerns, but also how, as these models are being trained and influenced, the geographies where this is happening. I think we all share the same concerns around what data is going in there and how it’s being secured and used.

Q: One of the comments that came up on the panel was around the challenge that we’re going to have in terms of who is going to take control of and have regulation around AI?

DM: While it’s still relatively new, that’s already started in Europe. I do expect that there will be rulings coming out in terms of what and how it can be regulated. The other great point that was brought up during the session was that you also don’t necessarily know the source of the data being used. When you’re getting information out of the AI tool, you don’t know what source was used to generate the information. So, you really don’t know the accuracy or potential bias of it. And that’s another big concern.

For example, when you do a Google topic or keyword search and the results come up, you can get details on the source, the website and when it was published. You don’t get that same level of information out of an AI tool. So, I think that transparency will be another place where you’ll likely see some regulation or some way to flag that you are receiving suspicious information.

Continued below

You can watch the full roundtable here: 

And Dave and Pete’s interview here: 

Q: How do MSPs and IT service providers go about building trust in AI?

DM: I think there are two aspects. One, you have to figure out whether you want to use something that is public versus private. And that’s an evolution. Some of our panelists used an AI to generate their bios, which I thought was a genius idea. Things like that are lower risk.

However, when you’re dealing with your own internal customer information, proprietary information, that’s where you’re going to need to figure out how to protect it and ensure the data is not for sale.

Q: How do you think IT service businesses will evolve as AI usage becomes more common?

DM: I think the regulation piece is a big part of it, but transparency around who is using it and how they are using it will also be a key factor.

The lawsuit aspect that was brought up in the panel discussion was something I hadn’t really thought about. But, if I or my company is taking somebody else’s intellectual property without realizing it, from a business perspective, especially somebody like N‑able, where we have all sorts of different intellectual property, if we found out we were using somebody else’s IP inside of our company, that would be a huge business risk.

Q: One subject that kept coming up was guardrails, why is this so important?

DM: That’s what you need. I don’t think it’s just specific to AI. It’s all things, all technologies. I mean, we were talking about the fact that self-driving cars have guardrails around ‘here’s where we operate, here’s where we don’t operate.’ We have a driver in place as a failsafe. That will continue to evolve. I fully expect within the next 10 years or so, we will have fully self-driving cars.

I think awareness is a huge piece of that. Businesses need to talk to their employees to make sure that they’re using AI efficiently and in a manner that doesn’t introduce risk to themselves or to the organization. That’s where you start to organically build those guardrails. It’s not the Wild West, but we’re still very early on in that cycle.

Q: We had a speaker that mentioned that they were the first platform to have Chat GPT built into it. Do you think we’re in danger of rushing in because vendors and suppliers want to be the first to implement GPT or AI generally?

DM: I think that’s where you need to review what and how your vendors are doing it, what data they’re sharing, and how they are using that data to enrich their offerings. By nature in this space, we are early adopters. So, we organically have that desire to get into those new pieces of technology. On the flip side, I think as a business, what you have to understand is the risk you may be introducing into your business by jumping in headfirst. I do think that if it’s done in a controlled manner where you have kind of analysed what the possible outcomes are and limited that risk it will be a great thing for businesses to adopt, you just need to be aware of the what, the how, and the why.

Q: An interesting question from the audience brought up how one MSP is seeing competitors targeting their customers and trying to undercut them by claiming that by using AI, they can dramatically reduce the cost of their service. As an honest MSP how do you keep up with that?

DM: The MSP isn’t just patching systems. They’re delivering a holistic set of solutions to protect a business and make sure it’s resilient. AI doesn’t do all those things.

I would almost question a customer who’s saying that they want the best solution for the least amount of money. Are their expectations realistic and do you really want them as customer? That’s a hard question to ask. But I think as a business you want a customer who is conscious of the value you’re adding, not whether a script can save them a few pennies. Is it worth fracturing a business relationship and the trust you’ve worked so hard to build? No AI can replace that trust.

Q: Should we be concerned about the idea of AI-to-AI decision making?

DM: I think at first with AI-to-AI decisions we’re going to see it in private areas, where it’s very controlled datasets interacting in such a manner. It’s not that different from what we do now where we’re linking APIs together with orchestration platforms to take data, complete a number of steps and pass it back. It will be very controlled. I don’t think you’re going to see 100% automation by a widespread number of businesses. There will be those specific use cases that really add tremendous value and it’s going to take time and expertise. The reality is, right now we have a platform, but inputting data and understanding how to manipulate it in such a way that it’s effectively used by the end user, that’s going to take time and expertise and training.

Pete Roythorne is Senior Brand Content Editor for N‑able