It’s easy to hear that there’s a massive opportunity for artificial intelligence for MSPs in areas such as financial forecasting, process optimization or compliance monitoring. But doing it is another story.There are so many questions:
- - How do you get started?
- - What can you expect?
- - Where do you find the right vendors?
- - How do you make money?
- To name just a few.
What’s an MSP to do? The answer is definitely not nothing. There are plenty of ways that MSPs can start leveraging AI in short order with the resources and skills they have, according to West McDonald, member of the GTIA AI Advisory Council and founder and partner of GoWest.ai, an AI-focused services firm.
West has spent nearly two years diving deep into AI, which includes making mistakes, learning and pivoting as innovations continue. He sees that many MSPs are reticent to explore AI for a couple of reasons: They don’t see a path to success, or they think AI hasn’t really moved beyond hype.
“It’s not that they’re afraid, but they’re not willing to pivot from their existing business models. But there’s a lot of low-hanging AI fruit for MSPs,” McDonald said. “I’ve had MSPs tell me, ‘We’d like to [explore AI], but we’ll stick with the vendors we have.’ The problem is it’s not about the vendors—it’s about identifying workflow problems and finding a solution for them. Find the problems then fill it with the right solution.”
Related Content: AI Use Cases by Functional Area Infographic
Turn the Theoretical into Practical: Getting Started with AI
In some ways, AI is similar to cybersecurity, McDonald said. Nobody calls you to say, “I’d like to buy more cybersecurity.” They call to say, “I have a business problem” and cybersecurity can help solve that problem.
“If you have a cyber offering, do you offer cyber assessments? Of course. AI is the same way. You can’t understand cybersecurity if you don’t know what the problem is until you go through an assessment and then match that to what solutions there are,” McDonald said.
To help MSPs, McDonald recommends three AI areas where MSPs can start:
- 1. Create a standard operating procedure (SOP) document that details guidelines and objectives for both the company, employees and customers.
- 2. Develop a best-practices document for AI agents, such as automated notetakers for meetings.
- 3. Standardize on a chatbot/large language model (LLM) and write policies and procedures
McDonald helped GTIA develop two Best Practices for MSPs Quick Start Guides: How to Use a Custom LLM Writing Engine and Developing AI Usage Policies and Guidelines, both of which offer more details on how to get started.
“In our view, and what we’ve learned in the last year and a half is that these are things that everyone should be doing for their company. For example, develop an SOP manual with policies and guidelines and go through it point by point within your organization. That takes AI from theory to reality for everyone and gets them thinking about it.”
Second, AI-enabled notetakers on sales calls have become ubiquitous on almost any Zoom or Teams call. But companies often don’t understand all the functionality or use them incorrectly, according to McDonald.
“Many of them are set up to auto-join a call and then send the notes afterward to clients. No, no, no. When we started using AI agents, I got some notes back because an AI agent stayed on after someone left the call and people were talking about company benefits or their recent divorce. Never again,” he said. “You should have a set of instructions included in a best practices document. You should ask for permission, and have to manually start it,” he said.
Another tip: AI agents can automatically update your CRM and task- or project-management software with the notes. Give your customers a framework of best-use policies so they get more value from the tools.
“Clients can use those notes to train on customer service and sales calls to send feedback to managers to talk about how that call went. Where you scored well, areas of improvement. There’s a lot that people aren’t using.”
Third, standardizing on a single chat or LLM enables you to learn how to leverage it better.
“You’ll hear arguments that Claude is better than ChatGPT or ChatGPT is better Microsoft. Like everything else, it all comes down to how you use it and are all the features set properly,” McDonald said. “Pick a platform and start integrating with other apps so they are standardized across the board.”
But also make room for exceptions. If some employees or customers prefer a different tool for longer writing, that’s OK.
“All of those things can be used, expertly used, and become a standard part of business. Once MSPs experience the efficiencies and safety and security of using this information, it will inspire them to make more AI decisions that can help them,” McDonald said. “But if you don’t have a foundation down, it’s very difficult.”
Start Internally, Then Go to Customers
It also behooves MSPs to learn about and use AI before they start talking to customers about it, let alone selling it. In many cases, customers don’t understand the technology or its use cases.
“The worst call I ever have is a customer that says ‘We don’t need to talk about AI. We’re a Microsoft partner.’ That’s great but what are you trying to do? Which pieces are you using? What business problem is it solving? Never start with a vendor. Start with the business problem.”
Once MSPs use AI features like LLMs, they’ll know what the models are good for, and not good for.
“For example, blogs are a dead giveaway. AI tends to overuse hyphens instead of commas and periods and writes overly verbose statements. But you can train that out of it with a set of instructions to write in the tone and voice of your company, and you can make sure you reference citations for facts and figures in the articles,” he said. “It’s very practical but you have to know how to use it.
“The mistake we made and that I see MSPs make is that we demand clients be the customer first. MSPs are not using it internally yet. You have to do that, and the ‘holy trifecta’ is a great place to start,” McDonald said. “We had one client that started with an AI policies and procedures document. Once that happened, they said let’s go through an assessment and it turned into a $30,000 project to help them with a knowledge base. The client saw low-level value and benefits and once they see the benefit, you’re off.”
