What Microsoft 365 Copilot actually costs an SMB (and when to skip it)

We get asked "is Copilot worth it?" almost as often as we get asked to turn it on. The two questions are related but the answer to the second depends entirely on the first, and the first has a more interesting answer than the price list suggests. Copilot can be very much worth it - for the right person, doing the right work. It can also be a per-seat line item that quietly does nothing. The difference is rarely the technology and almost always the fit.

The sticker price is the easy number

Microsoft 365 Copilot is an add-on on top of an eligible Microsoft 365 licence, billed per user per month, and at the time of writing it's an annual commitment. So the first part of the real cost is the prerequisite: Copilot assumes you're already on a qualifying business or enterprise plan, and if you're not, the true cost includes that uplift too. Check what you're actually licensed for before you price the add-on, because the gap between "we have Microsoft 365" and "we have a Copilot-eligible plan" catches people out.

The second part of the real cost is the readiness work. Copilot inherits your existing permissions, so switching it on safely means an oversharing and data-hygiene pass first - the same pre-work we cover in the Copilot rollout playbook. That's a one-off cost rather than a recurring one, but it's real, and a quote that pretends Copilot is purely a per-seat decision is leaving it out.

When Copilot is worth it

Copilot pays off fastest for people who spend their day producing or processing words and documents. Finance staff summarising and reconciling, anyone who lives in long email threads, people drafting proposals or reports from scattered notes, managers turning meetings into minutes and actions. For these roles the time-back is real and recurring, and the per-seat cost is easy to justify against an hour or two a week returned.

When to skip it, or buy fewer seats

Copilot does much less for people whose work isn't document-and-message shaped. A warehouse team, field technicians, anyone whose day is in a line-of-business application rather than Office, will mostly leave it idle. That's not a failure of the tool, it's a mismatch, and the right response is to not buy those seats. The expensive mistake is org-wide deployment because it felt fairer than picking. Buy seats where the workflows live, measure the time-back, and expand from evidence. A smaller deployment that's fully used beats a blanket one that's half ignored, every time.

Where to start

Map your roles against the workflows above before you count licences. If you want help sizing it honestly - including the readiness work and which seats will actually earn their keep - that's exactly what our Copilot Ready service scopes, and it sits on the tenant health that Managed IT keeps in shape. The goal is to spend on the seats that pay you back and not a dollar on the ones that won't.

Worth it is the wrong question at the company level and the right one at the role level. Ask it per person, buy accordingly, and Copilot stops being a leap of faith and becomes a line item you can defend.