Summary
Microsoft just reported $76.4 billion in revenue, an 18% year-on-year increase. Cloud contributed $46.7 billion of that, with the most considerable growth driven by Microsoft 365 E5 and Copilot. They're claiming 100 million monthly active users for Copilot and 20 million for GitHub Copilot. It was Microsoft's strongest quarter for Copilot seat commitments.
But here's what those numbers mean for your budget: Microsoft is increasing partner incentives for Copilot by around 50%. But don't look at that as just new rebates. Instead, it is a comprehensive incentive structure for the channel to push Copilot: direct rebates, indirect incentives, pilots, and net new licences. Everything the channel does gets incentivised around Copilot. They're also providing investment funds of up to 20% for direct enterprise customer assistance: consultancy, planning, adoption, rollout, and education.
Microsoft incentives push Copilot to increase revenue per user, not overall adoption
Microsoft is maximising ARPU (average revenue per user). Copilot pushes that revenue per user through the roof. The focus is on maximising current users and adding revenue on top, rather than on additional users in your organisation.
The November Earthquake
From 1st November 2025, Microsoft is implementing what I consider the most significant, tectonic change to Enterprise Agreement pricing in 25 years. They're retiring the A to D price levels that have existed since EAs began. A was the lowest discount, and D was the highest. Microsoft is aligning prices so you'll have one price list regardless of whether you have 2,500 or 200,000 users.
EA price tiers vanish on November 1 — the biggest change in 25 years of Microsoft licensing
Discounts are dropping. Negotiation becomes more difficult. For a 25,000-employee organisation, if their renewal hits after 1st November, they're looking at an additional $2 million per year just from the pricing realignment. Real money for what was previously cheaper on the price list.
🖐 Secure better renewal terms. Explore: Microsoft Enterprise Agreement Negotiation.
ARPA: The New Revenue Model
Microsoft's licensing strategy is evolving beyond human users entirely. We're moving from ARPU to what I call ARPA (average revenue per AI agent). Each AI agent receives an Entra ID, treated as a complete identity. Agents are managed via Intune, Purview, and conditional access. Usage analytics are built into Copilot control systems.
We're predicting basic agent licences will cost $15 to $25 per month. Advanced agents $35 to $50 per month. Then you have autonomous Copilots with base fees plus consumption charges (tokens, API calls). Deploying agents requires the broader E5 stack (Entra governance, Intune, Purview). Nobody's prepared for the cost increase this entails. Organisations are just getting a handle on FinOps and cloud economics, now you've got AI agents and Copilot metrics coming out.
🖐 Achieve optimal Azure cost efficiency. Learn more: Microsoft Azure Cloud Cost Optimisation.
The $30 Myth
The real cost is much higher than $30 per month. If you're on E3, you're moving to E5. If you're on F1 or F3, you need additional security components just to be eligible for Copilot. Even frontline worker SKUs need security add-ons. It's Microsoft maximising that average revenue per user, adding layer over layer.
I had a conversation today with a customer (over 150,000 users). The 1st November announcement hit them entirely out of the blue. The procurement executive said to her team, "I need to bring this up to the executive board. We can't continue to increase our Microsoft cost by 15 to 25% every single year. It's out of control. We need alternatives."
You've got techies and business people wanting the best of everything, but there's still a budget. You're running an organisation with increased cost per user for Microsoft for something where the ROI is uncertain.
🖐 Avoid overspending on AI deployments. Discover: Pricing Research and Pricing Metrics.
The ROI Problem
Microsoft talks about productivity: hours saved per user, how productivity is growing, and users have more time for other activities. I heard identical claims when Microsoft introduced smart tagging in Word and Excel years ago. You could press a smart tag and get more information. Microsoft said it was worth the extra $5, "less than a cup of coffee per day per user," and would save time. These are exact marketing phrases we heard many years ago.
Return on investment requires measurable usage data. That data is limited today. You've got multiple sources of analytical data straight out of the box. The Microsoft 365 admin centre shows enabled versus active users, adoption trends, totals, and prompts. The Viva Insights Copilot dashboard adds adoption by group, sentiment analysis, and action breakdowns, but has limitations, including reporting delays, licensing thresholds, and privacy filters. Power BI advanced reporting enables deeper correlation with business outcomes, but requires more integration and more customisation.
Real ROI comes from measurable business outcomes — not Microsoft’s “active user” metrics
These tools measure analytics of what people are doing with the tool. Real ROI is increased customer satisfaction from your call centre because you provide quicker, more accurate responses using Copilot.
The problem isn't ROI on one licence of Copilot. At SAMexpert, we're keen AI users. We use it cleverly: editing after Copilot and ChatGPT, reviewing all answers, never trusting anything. AI produces many hallucinations and wrong outputs. If you don't know how to prompt (and we don't like saying this, but it's true), you need to learn how to prompt.
If you're a keen user and know what you're using it for, assuming the total cost is $300 per month (about a day of your work), if you can save a day using AI cleverly, you get one-to-one ROI. Save two days, you get two-to-one ROI. Use it daily, save hours here and there. If you have a method to measure it, that's how you prove it.
The problem centres on ROI for projects and implementations: buying a thousand licences with only 5% using them. There's no objective measurement for usage. If you click the Tab button in Outlook, you're an "active user." Summarise one Word document in a month, you're an "active user." Is that really active usage? Is that ROI?
The problem is also trusting too much. Copilot represents a different challenge entirely. Your own Copilot connected to your account can see events in your calendar. That's why it fails when massively deployed without proof of concept, understanding the why. Not just because everybody else has it, but understanding the why.
If you understand that, it can empower your negotiations with Microsoft when they tell you just to buy a bright future.
Lock-in Mechanics
If you think Microsoft 365 was lock-in for the long term, Copilot is a chain. Once agents are deployed and GitHub Copilot is working and everybody is using it, switching to a competitor becomes extraordinarily difficult. Microsoft knows this.
They're successful in pushing licences into the market. That is how Microsoft operates: first, they push aggressively through discounting and channel incentives. Once they get a core hold on the market, they turn to implementation, utilisation and building business cases. We're at the peak of the first adoption wave. The next stage is business value coming into play.
Negotiation Reality
Enterprises must treat Copilot as a core budget item, a strategic platform. If you don't understand its strategic and don't take control, your budget will be entirely out of balance.
Copilot is a core budget item — control the sequence and never allow Microsoft to anchor first-round negotiations
Control the sequence. Don't allow Copilot to be positioned as the anchor early in negotiation. Microsoft tries to anchor first-round negotiations with Copilot. Secure baseline discounts and terms initially, then open Copilot discussions in later rounds. Don't get Microsoft to anchor quantities, definitely not large amounts at the start.
Challenge prerequisites. Microsoft often ties Copilot eligibility to upgrading E3 to E5. Present the total cost of ownership analysis. There's so much in E5 you're not using or never will use. Verify there's value and push back.
Demand flexibility. Negotiate multi-year opt-in clauses, phased adoption rights, and price lock-ins. If you're a large organisation, don't commit 50,000 seats upfront. Commit 5,000. Ask Microsoft for boots on the ground, assist in deployment, and build rights to expand at the same unit price.
Link spend to existing commitments. Look at Azure MACC commitments. Copilot can potentially offset those commitments. You might be double-counting cloud spend.
Microsoft has a take-it-or-leave-it stance on Copilot pricing. Significant acquisitions, large commitments, and big customers still get discounts. These are smaller than the old EA discounts, mostly minimum discounts. The SKU might be fixed with a small discount, but value can be achieved in other areas: pilot funding, favourable true-up terms, rollout funding through third parties.
The RFP Question
You're introducing something potentially bigger than your ERP system into your organisation. Have you run an RFP? Have you done competitive analysis? We're seeing almost automatic evolution: Office to Office Standard to Office Professional to 365, E3 to E5. It is something much bigger.
It is as big as an ERP implementation for the long run when you look at the implications. Where's the RFI, then RFP? Where is the old school, prudent evaluation of requirements? You're building a competitive option, a BATNA (best alternative to negotiated agreement) with Microsoft. You might have leverage.
There are options: Google Gemini, Watson X from IBM, Amazon Q, and OpenAI. There are niche solutions. Basic thinking is absolutely in place.
Midterm Commitment Traps
Organisations overcommit mid-term. You're 6, 12, 18, 24 months into a three-to-five-year agreement. Microsoft aggressively pushes Copilot top-down (executive fear of missing out all the way down). Organisations acquire Copilot mid-term. Don't commit to a substantial Copilot seat expansion midterm without getting renewal advantages. I tell you this from real battle scars: don't commit without getting renewal advantages.
Midterm Copilot additions reset your baseline — killing future negotiation power
Any midterm addition resets the baseline. It erodes your negotiation leverage. Microsoft expects you to increase commitment from the new baseline, not where you were prior to adoption. They won't remember you didn't have Copilot. They look at the baseline (you're paying $2 million, $3 million per year). Now they want a 25-35% increase in adoption. You're not growing 25-35%? That discount you got 12 months ago is gone.
Lock in benefits for the long term, not only what you need today.
The Productivity Paradox
Copilot can potentially make people lazy. If your users save X hours per day because productivity increased, are they using those free hours productively to increase business productivity? Business productivity means increased sales and higher customer satisfaction.
You need to evaluate the specific metric: are end users using freed time to provide better service, higher value to end customers, or increase revenue? That's the only metric a business cares about. Are we making more money or saving from the bottom line to improve margins? Business economics.
The Revenue Reality
Microsoft's average revenue per user started around $200-300 twenty years ago. With Copilot and everything to do with 365 and security, we're reaching $1,500 per user per year. Have we increased our revenue in the same proportion as a business? Are we making that much more money?
That's the understanding you need to take back to your organisation. That puts everything in proportion.
What's Next
Microsoft tried something remarkable with Copilot: not introducing volume discounts at all for the first time. From 1st November, all organisations, regardless of size, pay full price for Microsoft online services.
We're on the edge of a licensing revolution: more productivity, potentially ROI, and increased cost. You need to manage it. Nobody's prepared for the price increase and what this entails, and I'm not even talking about governance and management. Organisations are starting to get a handle on FinOps, now you've got AI agents and Copilot metrics.
Stay strategic. Watch the 1st November price increase. If you have a renewal coming after 1st November, consider renewing early. Lock in prices.
The age of digital workers means digital licensing, digital revenue. Microsoft invented licensing 40 years ago. Now they are leading a revolution of licensing.