Everyone's Using AI. Almost No One's Billing for It.

Walk into almost any MSP in the East Midlands and you'll find AI already at work. A copilot drafts the ticket responses. A tool summarises long email threads before an engineer picks them up. Someone on the service desk quietly runs client meeting notes through a chatbot to save twenty minutes. The technology has crept into daily operations without a single line item on an invoice.

That's the strange thing about AI in the MSP world right now. The adoption is real. The revenue isn't.

The numbers put a sharp edge on it. A recent industry survey found that 48% of MSPs rank AI as the top need their clients are asking about — the thing SME owners keep raising in review meetings. Yet only 13% generate meaningful revenue from it. Nearly half see the demand. One in eight has turned it into money.

That gap is the whole story. And closing it is more about packaging than it is about technology.

Two very different plays

The reason so many MSPs are stuck is that they've conflated two separate things. There's the AI you use to run your own business, and there's the AI you sell to clients. They feel similar. They are not the same play, and they don't make money the same way.

Play one: AI to protect your margin. This is the internal stuff. Faster triage, automated documentation, smarter alert filtering, copilots that help junior engineers punch above their experience. The payoff here is efficiency. If your team closes tickets faster and your senior people spend less time on grunt work, your cost to serve drops and your margin holds up even as wages rise. This is worth doing. Most MSPs have already started. But it's a cost play, not a revenue play — the client never sees it and never pays extra for it.

Play two: AI as a billable service. This is where you help the client adopt AI safely inside their business. You assess what they've got, you set guardrails, you pick the right tools, you train their staff, and you keep the whole thing from going off the rails over time. The client pays for this because it solves a problem they can't solve themselves.

Here's the mistake: because play one is quietly humming away in the background, a lot of MSPs feel like they're 'doing AI' and leave it there. They've improved their own operations and never built the client-facing offering. They're using AI to earn less from the same work instead of selling AI to earn more.

Why the demand is already there

You don't have to create appetite for this. Your clients are generating it on their own, usually in ways that should worry you.

Someone in accounts is pasting supplier contracts into a free chatbot to get a plain-English summary. A sales manager has signed up for an AI writing tool on a personal card and connected it to the company inbox. A director has read three articles about AI and now wants to 'do something with it' but has no idea what. We've seen a Nottingham manufacturing client where four different departments were each paying for a different AI subscription, none of them approved, none of them secure.

That's the real situation on the ground. Shadow AI, scattered subscriptions, and genuine anxiety about getting left behind or getting it wrong. SME owners know AI matters. They don't know what's safe, what's useful, or where to start. That uncertainty is exactly what they'll pay a trusted provider to resolve.

You are already the person they call about security and risk. Extending that to AI is a short step — but only if you've actually built something to sell.

A pricing structure that works

The cleanest way to close the gap is to split the offering in two: a one-off project to get the client ready, then an ongoing retainer to keep them safe. This mirrors how you probably already sell projects and managed services, so it slots into your existing commercial model without inventing anything exotic.

1. The AI Readiness Project (fixed fee)

This is a defined piece of work with a start and an end. You charge a fixed fee so the client knows exactly what they're committing to. For a typical 20 to 50 seat SME, something in the £2,500 to £6,000 range is reasonable, scaling with headcount and complexity.

What's inside:

• Discovery and audit. What AI tools are already in use across the business, approved or not? Where's data going? What are the obvious risks?
• Use-case mapping. Which tasks in this specific business would actually benefit — and which are hype? Naming two or three real wins is worth more than a glossy strategy deck.
• Policy and guardrails. A written acceptable-use policy, data handling rules, and a list of sanctioned tools. This alone is something most SMEs cannot produce on their own.
• Tool selection and setup. Choosing and configuring the right licensed, business-grade tools — Microsoft Copilot, for instance — rather than a free-tier free-for-all.
• Staff training. A couple of practical sessions so people use the tools well and know where the lines are.

The deliverable is a client who has gone from scattered and nervous to organised and confident. That outcome is easy to sell.

2. The AI Governance Retainer (monthly)

Readiness is a snapshot. AI moves too fast for a one-off fix to stay valid. New tools appear, vendors change their data policies overnight, staff find workarounds, and regulation keeps shifting. That ongoing change is your recurring revenue.

Price it per seat or as a flat monthly fee — think £8 to £15 per user per month, or a flat £300 to £800 for a small business depending on scope. The retainer covers:

• Monitoring sanctioned and shadow AI use across the estate
• Keeping the policy current as tools and law change
• Reviewing new tools the client wants to adopt
• Quarterly reviews on what's working and what to stop paying for
• Being the named point of contact when staff ask 'can I use this?'

This turns AI from a project you do once into a service you provide forever. It's governance, and governance never finishes.

Start before it's perfect

The temptation is to wait until you've built a polished practice before you charge for anything. Don't. The 13% aren't miles ahead on technology — they've simply decided to put a price on help they were half-giving away already.

Pick one client who keeps asking about AI. Scope a readiness project. Do it well, write it up, and use it as the template for the next ten. The demand is sitting in your inbox. The only real question is whether you keep absorbing it as free advice or start treating it as the service it clearly is.