Microsoft made some pricing announcements today at their Inspire conference that included a $30 charge per person per month for M365 Copilot (AI) licensing. First impression? That’s pretty steep… But let’s dig in a bit.
Not everyone is going to need it (at first).
It’s pretty easy to justify when you know your value per hour and can calculate how much time you’ll save using the new features of Copilot. In many cases it’ll be easy to justify $30 a month cost when you can measure savings directly: “Hey, that thing takes 15 minutes, I do it 10 times a month”… Done.
Other use cases likely won’t be that straightforward, but we’ll see how they manifest.
With this, I’ve already seen it come up in multiple conversations and threads… Know your users and review your use cases before licensing. At this price point, most companies aren’t going to license everyone – at least right away. And there will likely be other added licenses in the mix as well. Don’t expect the Microsoft licensing topic to get any less confusing in the near future (unless maybe we can use AI to help sort out your licensing use cases… Copilot for Licensing?).
SaaS pricing is hard
A little background. SaaS pricing is a challenge. Dig in a little bit and you’ll see plenty of threads talking about how hard it is for companies (especially startups) to price their products. M365, Power Platform, and the rest of the cloud services out there ultimately are SaaS offerings (Software as a Service) – though it may be a bit of a simplification. Microsoft and other big players in the space at least had some historical product pricing to start from that (may or may not directly) lead to subscription pricing used for SaaS.
The nature of cloud-based apps – as we’ve been talking about for as long as Microsoft has been in the cloud – is that we get regular, rapid, incremental updates. We get new features and changes basically weekly – for good or bad. (“Bad” only really for trainers and folks trying to keep up with it all…) This replaces the old 3-year-cycle of product releases from Microsoft where it was much easier to justify a price increase. You could easily put together the list of new and improved features vs. the last product (now 3 years old). Now, things move so fast it’s challenging to keep up and get a high-level view of increased value over time – it’s almost an expectation.
Stepping back a bit, as initial moves to the cloud resolved, it became fairly easy for most organizations to rationalize M365 pricing. Now, years later, not much (relatively speaking) has changed in the pricing, but much has changed with the offering and capabilities. Some organizations have moved up from one level to another, but many have also stayed with the same licensing since they first adopted M365.
Microsoft needs (or at least wants) new revenue streams (oh look, a stock price jump at the time I published this) and it’s harder to justify price increases when you’re incrementally updating your software – harder for end users to see. But incremental (smaller, but more often) updates are the norm with cloud-based software. Vendors then are forced to rebrand and offer bigger bangs to justify step ups in cost. Enter Copilot and today’s licensing announcements. Viva is in this same boat. And don’t forget the recent “Entra” branding for authentication and ID services – just wait.
Power Platform (as part of the Dynamics/D365 product group) has been working on this challenge for a while now. Initial feature offerings were heavily in the M365 integration space – leveraging tens of millions of users. They got the traction they were looking for and at the same time were building a ravenous (#PowerAddicts) community around it to pull the movement ahead even more. But they’ve been seemingly challenged to get a mass shift to newly licensed products and offerings. It does seem to be shifting, but more slowly than (I expect) they’d like. As Power Apps and its sibling products grow and Dataverse finds its space in the storage area – more folks are realizing the business value and making the move to additional licensing.
Cost of ownership
There’s more than the $X/month/user. There’s also the cost of managing and training – after you’ve spent time and money evaluating the technology. Don’t overlook that. Give plenty of time for architecture folks to review functionality, security folks to review risk, your tenant admins time to ramp up, and trainers time to produce the materials you’ll need. It wouldn’t shock me if early on we see a “you need to take this training before we’ll buy this feature for you” sort of approach.
Finally, $30 a month is a baseline. Many large enterprises are able to negotiate pricing, so keep that in mind. Sorry SMB. 🙁
Copilot is a consistent brand, but it’s not one single cost. There will be a collection of products across the Microsoft stack and licensing will be different and separate for each. Copilot for Business (from Microsoft-owned GitHub) is a nice example since it’s so different from M365 offerings. It’s still “Copilot” but a very different offering for developers rather than business users (though… there’s the “maker” argument for another day. 🙂 )
Bottom line is, start thinking about Copilot licensing across the stack. There will likely be additional offerings. We already know about several in the flurry of announcements.
Not everything falls under “Copilot” either. Teams Premium, while feeling like a bit of an overlap with Copilot, has its own licensing and capabilities. You may need one, or the other, or both – hence the importance of evaluating your environment and users before diving in.
Microsoft is all over the security concerns. ChatGPT and others have highlighted the risk of playing with AI in the public domain and its collection of user data to learn and train itself.
Microsoft is working overtime to let folks how data is used, how their version of AI is being built, and how common components (LLM) are kept separate from organizational data (Microsoft Graph). There’s obviously going to be some risk with data in the cloud, but efforts and intention are in place. Each organization will need to evaluate the risks for themselves. As with other technologies, certain areas will be slower to adopt due to risk. We’ll keep hearing more from Microsoft and the community on this.
As the capabilities of software increase – so will the costs. This isn’t a bad thing. Microsoft and their industry peers are adding new capabilities far faster than they have in the past and many of them add great value for their users.
The challenge to us as consumers again falls into keeping up with the changes. As much as these new capabilities benefit us, we need to invest time and effort to learn how to best use them and reap the benefits.
It’s easy to get sucked into the hype. There’s a lot of cool stuff here. Just temper your expectations a bit to not get too frustrated or disappointed. We’re early in the AI application cycle. It’s going to take time to mature and polish implementation. Exciting times.