Thirty dollars per user, per month, on top of a Microsoft 365 licence most organisations already pay for. For a 100-person company that lands at $36,000 a year. That is a serious line item, and it is usually approved or rejected on the strength of a 20-minute demo where Copilot summarised a Teams meeting beautifully and everyone nodded.
The demo is not the problem. The problem is that nobody in the room has worked out what those 30 dollars have to buy back. So here is a practical, numbers-first look at Microsoft 365 Copilot value for money, based on how the tool actually behaves once the novelty wears off.
Start with the break-even maths, not the feature list
Thirty dollars a month is $360 a year per person. If your fully loaded staff cost is around $60 an hour, that seat needs to return six hours of someone’s time across twelve months just to wash its face. Six hours a year is 30 minutes a month.
That framing changes the conversation. You are not asking “is Copilot impressive?” You are asking “will this person get half an hour a month back?” For an account manager drowning in email threads, the answer is probably yes. For a warehouse supervisor who touches Outlook twice a day, it is almost certainly no.
The other figure worth pinning on the wall is cost per working hour. At roughly 1,700 productive hours a year, that $360 seat costs about 21 cents an hour. Cheap if it works, expensive if it sits unused, which is exactly why seat selection matters more than seat count.
Where Copilot genuinely earns its $30
Meeting catch-up
Teams recaps are the strongest part of the product. Missing a 50-minute call and getting the decisions and action items in 90 seconds is a real recovery of time, and it is the one capability almost every user keeps using after month one. People who sit in four or more meetings a day tend to get their money back here alone.
Long email threads
A 40-message thread with three competing proposals is genuinely painful to reconstruct. Copilot does that well, and drafting a reply in the thread’s tone takes another two minutes off the job. Sales, support and project roles see this daily.
Spreadsheet work you half-remember how to do
Asking for a formula in plain English, or for a quick explanation of what a nested IF is actually doing, saves the ten-minute forum hunt. It is not a data analyst, and it will confidently misread a messy table. My notes from six weeks of scoring Copilot’s accuracy on real work put formula help in the reliable column and multi-sheet reasoning in the shaky one.
First drafts that need rewriting anyway
Copilot’s Word output is a decent skeleton and a poor finished document. Teams that treat it as a way to skip the blank page get value. Teams that expect publishable prose end up editing just as long as they would have spent writing.
Where the value quietly leaks away
Three things eat the return.
- Permissions debt. Copilot answers from anything the signed-in user can already open. If your SharePoint has a decade of over-sharing, it will surface documents nobody meant to circulate. Fixing that is a project, and it is not included in the $30.
- Adoption collapse. The usual pattern is 90% of seats active in week one and 25% active by week eight. Unused seats are pure cost, and they are the single biggest reason Copilot ROI disappoints.
- Task mismatch. Give it to someone whose job is mostly physical, client-facing or highly regulated, and you have bought a $360 calendar assistant.
There is also a hard ceiling on what the product does. My breakdown of what the $30 Microsoft 365 Copilot seat still can’t do covers the gaps in cross-file reasoning, long-document work and agent reliability, and it is worth reading before you sign a twelve-month commitment.
The licensing detail that catches finance teams out
Copilot is an add-on. It sits on top of Microsoft 365 Business Standard, Business Premium, E3 or E5, so the real comparison is not $30 against nothing. It is $30 against a base licence you may also need to upgrade.
The add-on is also sold on an annual commitment in most agreements. You cannot switch half your seats off in month four because the pilot flopped. Budget the full year for every seat you turn on, and pilot narrow rather than wide.
A 90-day test that produces a real answer
Skip the company-wide rollout. Pick 12 to 15 people across sales, finance, HR and operations, then measure before and after.
- Ask each person to name three recurring tasks and time themselves doing them for one week, unassisted.
- Give them Copilot for 30 days and have them time the same tasks again.
- Check licences active in week six. Anything below 50% is a warning you should act on, not a number to explain away.
- Run a permissions audit on the folders those 15 people can access before day one, not after.
- Score each participant honestly at day 90 on whether they would pay for it themselves.
That last question is blunt and remarkably accurate. People who would reach into their own pocket for a tool are the people who will actually use the company’s licence. On ease of use, the pattern from testing Copilot with 12 real users was that confidence in week one predicted usage in week twelve far better than job title did.
When Copilot Pro is the smarter spend
Not everyone needs the enterprise add-on. Freelancers, contractors, sole traders and small firms without a tenant full of documents can often get 80% of the practical benefit from Copilot Pro at $20 a month, bought month to month with no commitment. If that sounds like your situation, the breakdown of where Copilot Pro actually pays off is a better starting point than a sales call.
The dividing line is organisational data. If your value depends on Copilot reading your SharePoint, your mailboxes and your Teams history, you need the $30 seat. If it depends on it helping one person write and analyse faster, you probably do not.
Three numbers to take to your finance director
Forget feature comparisons. Walk in with these.
Daily active usage after 30 days. Anything above 60% supports a wider rollout. Below 35%, cut the seats to the people who are using them and revisit in six months.
Minutes saved per task, by task type. Not an average across the business. Meeting recaps and email triage carry the case; document drafting rarely does on its own.
Annual cost per seat against loaded hourly cost. $360 against $60 an hour means you need six hours back a year. If your pilot shows people recovering 20 minutes a week, the maths is not close, it is a clear yes, and you can defend it in a budget meeting with a straight face.
