Aesopians E11: Pricing And Packaging AI Services For MSPs

I’ve been having more conversations with MSPs lately about AI services, and something has definitely shifted.

I’m seeing more prospective customers ask about AI through MSP websites. I’m hearing from clients that their existing customers are bringing it up. In one case, an MSP discovered its engineers were already working AI-related support tickets without management realizing how much demand was starting to develop.

They eventually had to tell the engineers to escalate those tickets because the company needed to figure out what it actually wanted to offer.

That would probably be one of the first things I’d do if I were running an MSP right now. Ask your engineers what customers are already requesting. You might discover you have demand for AI services before you’ve created a product to sell them.

Well, once you’ve established that demand exists, the harder question becomes: how should an MSP actually price and package AI services?

That’s what I wanted to dig into in this solo episode of the Aesopians Podcast.

Watch The Episode Now

📖 Related Resources:

🎧 Listen On Spotify, Apple & More:
https://creators.spotify.com/pod/profile/aesopians/

Microsoft Copilot Is The Obvious Starting Point For MSPs

Microsoft Copilot is probably the easiest entry point into AI services for most MSPs.

You’re already selling Microsoft products. Your customers are already using Microsoft 365. Copilot has a familiar per-user monthly pricing model, and Microsoft would obviously like its partners to sell more of it.

From a pricing and packaging standpoint, it fits pretty neatly into the traditional MSP model.

You add the license. You add some markup. The customer pays a recurring monthly fee. If someone has a problem using Copilot with Excel, PowerPoint, Outlook or another Microsoft product, they call the help desk.

I expect a lot of MSPs will go this route.

Personally, I’ve been less impressed with Copilot as an end user. I added it to my Microsoft 365 subscription, experimented with it for a while and eventually canceled it because I wasn’t getting enough value compared with how persistently we’re using ChatGPT.

That doesn’t mean Copilot won’t improve. AI products are probably about as bad today as they’re ever going to be. Microsoft has a lot of room to make the product better.

But even if the product gets significantly better, I’m not sure simply reselling Copilot creates the most interesting AI opportunity for MSPs.

It still leaves you in a familiar position.

Something breaks. The customer calls. You fix it.

More Advanced Customers Are Going To Want More Than Copilot

The customers spending the most time thinking about AI aren’t necessarily trying to decide whether to buy one AI product.

They’re experimenting with ChatGPT, Claude, Gemini, Perplexity and whatever comes next.

Based on the inbound activity I’ve been seeing through client websites, ChatGPT and Claude seem to be generating a lot of the interest from companies looking for help implementing AI. They’re not simply looking for access to a chatbot. They’re trying to figure out how these products can fit into their businesses.

That creates a different opportunity for MSPs.

Instead of picking one AI model and telling every customer to use it, MSPs may want to look at platforms that can provide controlled access to multiple models through a more secure environment.

That could give customers access to several AI platforms while creating more of the security and governance structure MSPs are accustomed to managing.

The underlying AI usage may then become a pass-through expense with some markup rather than a simple flat monthly software license.

That’s already a different pricing model than Copilot.

But I think the bigger change happens when customers start asking you to help them actually build things.

The Bigger Opportunity Is Business Process Automation

Basic AI adoption usually starts with prompting.

• Write this email.
• Summarize this document.
• Research this company.
• Analyze this spreadsheet.

That’s useful, but you eventually hit a point where the user wants AI to do something with the information it creates.

They want to take information from one system and put it into another. They want an incoming document to trigger a process. They want ChatGPT connected to company information. They want an AI agent to take an action inside another application.

Once you start doing that, the conversation quickly expands beyond ChatGPT or Claude.

Now you’re talking about APIs, MCP connections, Zapier, Make, n8n and other low-code or no-code integration tools that bridge the gaps between different business applications.

That’s why I’ve been thinking about the service as something closer to Business Process Automation Plus, or BPA+.

You can call it whatever makes sense for your MSP.

The important distinction is that I wouldn’t necessarily try to cram all of this into your normal managed IT agreement.

There will still be plenty of customers who just want the traditional services.

They want the help desk. They want cybersecurity. They want Microsoft 365 managed properly. They don’t particularly care about building AI workflows.

That’s fine.

But the more technology-forward customers may want something substantially more proactive.

Instead of:

Something broke. Can you fix it?

The conversation becomes:

We have this workflow inside the business. Can you help us make it better?

That’s a very different relationship.

A Premium AI Service Desk Could Be The Product

One way I’d think about packaging this is through a premium AI service desk.

Your normal service desk continues supporting the customer’s everyday technology needs.

The AI service desk is staffed by technicians with a different skill set.

These employees need to understand AI platforms, but they also need to understand tools such as Zapier, Make and n8n. They need to be comfortable with APIs, MCP connections and broader low-code or no-code systems integration.

A customer might come to that team and say:

We receive this type of document every day, someone manually reviews it, puts the information into another system and then sends an email. Can we automate some of that?

That’s where the service becomes more interesting.

You’re no longer waiting around for a laptop to stop working.

You’re helping design and implement a better business process.

That gives the MSP a chance to participate in how the customer improves and grows the company, instead of primarily being viewed as the cleanup crew.

Price AI Services Around Power Users

I don’t think you necessarily need to sell this premium service to every employee.

Imagine you have a 100-person customer.

Maybe 100 employees need normal managed IT services.

Maybe 50 of them eventually use some sort of AI product.

But there might only be five or ten people inside that company who are going to become serious AI power users.

Those are the employees I’d consider building the premium package around.

I don’t have a magic price point yet, and I don’t think anyone really does because the market is still developing.

But I’d probably test a higher monthly price against a smaller group of users.

• Maybe that’s $100 per power user per month.
• Maybe it’s $200.
• Maybe it’s $200 plus the customer’s AI usage.

The exact number needs testing.

The pricing structure is more important than pretending we’ve already discovered the perfect price.

You’re taking a relatively small number of employees who are actively trying to improve the business and giving them access to technical resources capable of helping them build workflows, integrations and automations.

That’s worth more than answering another Copilot support ticket.

And, at least while the market is still relatively new, I suspect many MSPs can charge more for this kind of service than they initially think.

A Two-Tier AI Offering Could Make Sense

I could see MSP AI services developing into two fairly distinct packages.

Basic AI Support

This looks more like the traditional MSP model.

You resell Microsoft Copilot or another standardized AI product. You help get users set up. You provide basic support when they have problems.

Maybe this adds $20 per user per month to an existing agreement, depending on how the licensing and support are structured.

The individual price is relatively low, but you may be able to sell it to a much larger percentage of the customer’s employees.

AI And Business Process Automation

This is the premium package.

A smaller number of power users get access to technicians who can help them build AI workflows, create integrations, connect business applications and solve more complicated automation problems.

Maybe that’s $100 or $200 per power user per month. Maybe there are additional usage charges.

The important thing is that you’ve separated basic AI support from deeper AI implementation work.

One is an extension of the help desk.

The other starts looking more like recurring technical consulting.

Recurring AI Services Fit The MSP Business Model

MSPs have spent years trying to move more of their businesses toward recurring revenue.

That’s another reason I think this type of AI offering is worth exploring.

You could certainly sell one-off AI consulting projects.

There’s nothing inherently wrong with that.

But if you can create an ongoing service where a customer has access to a technical team that continuously helps them identify, build and improve automations, you’re back in a model MSPs already understand.

• Monthly recurring revenue.
• Repeatable service delivery.
• Ongoing customer relationships.

The economics are potentially attractive too. If you can develop technicians who become good at this work and reuse the lessons they learn across multiple customers, you may be able to build a relatively high-margin service line.

Then you have to figure out utilization, scope, usage costs and exactly what is included.

Those are solvable service design problems.

AI Could Change How Customers Perceive Their MSP

I think this might be the biggest opportunity.

A lot of traditional MSP services are ultimately built around risk reduction.

• Keep the network running.
• Protect the company from cybersecurity threats.
• Back everything up.
• Fix technology when something goes wrong.

Those services are valuable, but they can also make the MSP feel like an insurance policy.

And businesses tend to look pretty aggressively at the price of insurance when nothing is currently going wrong.

AI and business process automation give MSPs a chance to get involved in a different side of the business.

You’re helping the customer create something.

You’re helping automate repetitive processes.

You’re helping employees become more productive.

You’re participating in research and development, systems integration and potentially the growth of the company.

I don’t think creating a basic Copilot package changes that perception very much.

If the offer is basically, “Call the help desk when you can’t get Copilot to work with Excel,” you’re still operating inside the traditional support relationship.

If you’re helping the customer design workflows, connect systems and deploy AI throughout the business, the relationship starts to look different.

You’re moving from primarily protecting the business to also helping improve it.

That creates more room for differentiation and potentially more room for premium pricing.

MSPs Should Start Experimenting With AI Services Now

It’s hard to predict exactly what this market will look like in another couple of years.

Competitive pressure could push prices down. AI platforms will change. Customers will become more sophisticated. Products that are impressive today may be commodities tomorrow.

But if I were running an MSP right now, I wouldn’t wait for everybody else to figure out the standardized package first.

I’d start experimenting.

Talk to your engineers and figure out what customers are already asking them to do.

Find a few customers that are actively experimenting with AI.

Identify the handful of employees inside those organizations who are likely to become power users.

Start building the processes, technical skills and service boundaries required to support them.

Then figure out what customers are willing to pay.

Stop thinking about AI only as another software license you can resell and start thinking about the service layer you can build around it.

That’s where I think the more interesting opportunity exists.

For some MSPs, that may become a simple AI support add-on.

For others, it could develop into a premium business process automation service that creates recurring revenue and gets the MSP involved much deeper inside the customer’s operations.

The market is going to tell us which models work.

But right now, there are already customers asking for help.

I’d rather start learning how to serve them today than find out in two years that somebody else already figured it out.

If you’re one of the MSPs starting to bring AI services to market, Tortoise and Hare Software is already working with MSP clients on content, SEO and advertising campaigns designed to reach businesses looking for AI help.

Ready to start marketing your MSP’s AI services? Book a free consultation.