M365 Copilot: ROI and Cost in Real Numbers, Jan 2024 Availability

 Microsoft 365 Copilot is finally available to everyone, assuming you can afford it. Let's start with what's new. Firstly, all the limits have been removed. There is no more 300-seat requirement to purchase Copilot licenses. You can start with just one. Compared to ChatGPT's business plan, which requires at least two users to be subscribed to, that's a definite advantage if you only need one license.

Logically, it has also been brought into the CSP program, so you are no longer required to have an Enterprise Agreement.

And the cherry on the pie is the eligible plans. In addition to Microsoft E3 and E5, Copilot can now be added to Office 365, E3 and E5, and Microsoft Business Standard and Microsoft Business Premium plans. And how could I forget Microsoft 365, A3 and A5? But what does it actually mean?

It means that you cannot buy Copilot alone. You must have one of these licenses assigned to the user for whom you plan to buy Copilot. It also means that it's not included in any of the plans. It comes at an additional cost. And, speaking about the price, it is 30 per user per month, regardless of your industry and size, so education gets it at the same price and your Enterprise Agreement level.

Microsoft isn't going to give any Enterprise Agreement level discounts. If you have an Enterprise Agreement, you should be aware that the price list has four different predefined pricing levels. Depending on the size of your organisation. And that doesn't apply to Copilot. So the price stays the same.

Can you get a discount? Maybe. You will probably not get, as it stands, a discount on Copilot itself at the moment. Still, you may obviously use it as a negotiation leverage, asking for a discount on something else. We will get back to that. The most interesting question is, Do you want to implement Copilot?

The problem is not the price of Copilot itself. The problem is the additional budget required to implement it in an organisation. We don't believe that everybody will jump in and implement it org-wide, but if they did, here is the uplift compared to existing plans. If you have users with Microsoft 365 E5, adding Copilot licenses for these users will cost you an additional 53%.

If you add that license on top of E3, the increase is 83%. For Business Plans, it gets even better. Adding Copilot to Microsoft 365 Business Premium will cost you an additional 136 per cent of what you're paying right now, multiplying what you're paying by 2.4. And for the users with Microsoft Business Standard, the uplift is 240%.

The other thing to note is that frontline workers with F-plans are not getting Copilot. So, they would have to be switched to something like Microsoft 365 E3. And if you do that, the online subscription cost per such user will be 8.25 times higher than what you're paying now. The good thing, however, is that you don't have to buy it for every user in your organisation.

You choose who gets it. It could be top managers, it could be data analysts. It could be number crunchers, AI enthusiasts, or those who already have experience working with ChatGPT; you can have a gradual rollout. So, it doesn't mean that your budget will suddenly increase by 83 per cent for all the Microsoft licenses for all the users in your org.

But it's obvious that you must seek an additional budget. And when it comes to budgets, the thing that we need to present to our CFOs is the ROI. What return on investment are we expecting from Copilot? Is the price of 30 per user per month justified? Let me start with a personal example.

We've been using ChatGPT since its inception since they rolled out the commercial plan. In the first half of 2023, we calculated the ROI on usage of ChatGPT in SAMexpert, and the number was 30 to one. That is how much time ChatGPT saved us doing research, drafting articles, summarising video transcripts, working with subtitles, replying to emails, drafting documents, and many other things. And that's ChatGPT, which has much more limited functionality compared to Copilot. And by the end of 2023, the ROI went up to a hundred to one. So, in the hands of an experienced user who knows and wants to use AI in their day-to-day work, it's a no-brainer. $30 is the cost of one hour for a mid-level employee in the West.

Don't you think an AI cannot save at least one hour a month? However, we should not forget those users who probably won't even touch it. And thus, there'll be no return on investment in that particular user base. This is why we suggest a gradual rollout and training your employees before jumping into a company-wide subscription, which may cost you dearly.

Going back to the price, will it definitely be 30 per user per month for your organisation? Can you negotiate a discount? As I said in the beginning, it's very difficult at the moment. Microsoft is extremely confident in this offering. Look at what they did with the Enterprise Agreement discounts. They just don't exist.

There is no level D price for Copilot. It's 30 per user per month regardless. There's almost no movement on Copilot. Can it be used as a leverage? Cleverly, yes. What may also happen is the price can go up or down at the end of the year, maybe even earlier. It depends on the adoption.

It depends on the market's competition: whether there will be any significant competition from, say, Google. Also, Microsoft is working on continuous improvement of the back end and, therefore, decreasing costs for the data centres. It's not a secret that AI is very resource-hungry. Multiple experts quoted that numerous times as the reason why Microsoft, for a while, kept that 300-user minimum limit.

They didn't want it to be adopted by everybody to keep the load of the data centres under control. There's also a historical precedent when Microsoft cut the prices for Power Apps in half after the early adoption stage. However, looking at the ROI, I would not bet my money on that.

One of the other possible scenarios is Microsoft may introduce Microsoft 365 E7. Rumours and speculations about E7 have been around for almost two years. The reason is Microsoft 365 E5 was introduced as an all-inclusive plan. One after another, Microsoft started introducing premium add-ons that you have to purchase on top of E5. Copilot is one of those, but also Teams Premium and Intune Plan 2, which are pretty compelling for many enterprises.

So the question remains whether Microsoft will bring another all-inclusive plan this year and then begin adding another set of premium add-ons on top of E7. Nevertheless, there are, however, two arguments against it. If E7 is introduced alongside the current E3 and E5 plans precisely in the same framework, Microsoft would have to give Enterprise Agreement discounts. But would they? Look at Copilot. It is available in the Enterprise Agreement channel, but no discounts exist. It's a flat price. It's not even on the price list. So, E7 could potentially be positioned exactly the same.

The other argument is that E5 is not yet widely adopted. It's still a flagship product, so that goal has not yet been reached; so why introduce a new plan when the old one has not yet been adopted? We are even seeing rollbacks from E5 to E3. But watch that space closely. If that actually happens, and we're going to have E7 with Copilot included at a flat price, regardless of the discount level, that means the end of the Enterprise Agreement.

Rumours about it started circulating in December last year. It's not even an inside information. Multiple experts have already announced that Microsoft will move in that direction. When that will happen is yet to be seen. But the most important thing we will lose with the demise of the Enterprise Agreement is programmatic levels.

Suppose Microsoft moves everybody into CSP. If Microsoft moves enterprises into an equivalent of CSP, every discount will have to be negotiated carefully. And that's not the only problem because CSP, even with the latest changes, still differs from the Enterprise Agreement on so many levels.