Ask David Nicholls of Solve Business Services, and he’ll tell you AI is the next major shift in IT, and MSPs need hands-on experience if they are going to guide their customers through it.
“AI is one of those inevitable technologies, like cloud computing,” David says. “If you don’t know about it and don’t use it yourself, you can’t be there as an authority for your customers.”
For Solve Business Services, that means focusing primarily on Microsoft Copilot and learning where it can deliver productivity improvements for small businesses – while also helping customers understand the cost, complexity and risks that can come with more ambitious AI projects.
Start small
David’s advice for businesses excited about agents and automation is simple: don’t start by trying to build something enormous.
“Start with a little sandcastle in your backyard. Don’t build the whole Pyramids of Giza as your first thing.”
He has seen customers arrive with ambitious ideas involving multiple systems, APIs and external data sources, without necessarily appreciating the development work, governance or ongoing costs involved. Part of the MSP’s role, he believes, is to help customers understand what they are really talking about taking on. He refers to the “Spiderman clause”: with great power comes great responsibility.
From four hours to eight minutes
The potential payoff, however, can be significant.
One Solve Business Services customer, a financial planning business, developed a Copilot skill to help produce client facing work product. Previously, gathering recommendations, meeting notes and supporting documents could take three to four hours of actual work and several days elapsed time.
Using Copilot, the process was reduced to around seven or eight minutes, with a person reviewing the output before it went to the client.
For David, this is where AI becomes compelling: not necessarily replacing people, but removing repetitive process-driven work.
“They are getting more consistent output, happier staff because they’re spending less time on boring tasks,” he says. That frees employees to spend more time on more interesting work.
Productivity inside the MSP
David is applying the same principle within Solve Business Services. He uses AI to compare and summarise documents, review emails and calendars, help structure blog content and assist with scripting.
Using GitHub Copilot, David has created PowerShell scripts that automate previously manual administration and reporting. One example cited produces a formatted report showing mailbox access and delegations across an entire customer environment in seconds – a job that could otherwise require manually checking dozens of mailboxes.
“It’s helping your support be more efficient, absolutely. And more professional.”
Keep an eye on the bill
AI productivity also introduces a new consideration for MSPs: consumption costs.
One small customer embraced Microsoft Cowork while it was in preview (Frontier). Once it became Generally Available it became billable, Solve Business Services reviewed the usage and projected an annual cost of around $40,000. The solution was delivering value, but the cost prompted a rethink, including moving some functions into Copilot Studio agents.
David sees cost controls as an important part of responsible AI implementation. Spending limits should be configured for users and groups so businesses can embrace AI without fear of unexpected costs.
For MSPs, this creates a new responsibility: understanding not just what AI can do, but what different approaches are likely to cost.
Know where your expertise begins and ends
Solve Business Services isn’t trying to become an AI software development company. Instead, David has chosen to specialise in the Microsoft environment, understand Copilot’s capabilities and limitations, and know when to bring in someone else.
At this stage, Solve Business Services isn’t charging separately for most of this AI support. David sees helping customers enable and navigate Copilot as an extension of the Microsoft 365 service he already provides… and an opportunity to build his own expertise alongside his customers. More substantial work, such as a full AI readiness assessment, would sit outside the existing service agreements but will be assessed on case by case basis.
Security still comes first
For small businesses already operating in Microsoft 365, David sees another advantage to Copilot: it sits within an environment the MSP is already managing and offers Microsoft’s Enterprise Data Protection.
That doesn’t remove the need for AI preparation. Permissions and sensitive information still need consideration, and David suggests staged rollouts can be useful in larger or less well-understood environments to establish what different users can access before deploying Copilot more broadly.
He is particularly cautious about connecting third-party AI services directly to company data before understanding exactly what is happening to that information.
“Once that thing slurps up all your data, then it’s gone. Can’t undo it. You can’t take the dye out of the swimming pool after you’ve thrown it in.”
The first step? Use it yourself
David’s advice to MSPs that are still watching AI from the sidelines is straightforward: get a Copilot licence and start using it.
Begin with everyday tasks – summarising emails, researching a customer support question, improving a prompt or rewriting a difficult email – and gradually build an understanding of where the technology is most useful.
For an MSP working in the Microsoft ecosystem, David believes not having that firsthand experience is “crazy.”
His experience suggests that MSPs don’t need to become AI developers overnight. But they do need enough practical knowledge to help customers turn an exciting idea into a useful, secure and commercially sensible solution.