Microsoft AI licensing splits across three commercial vehicles. Copilot for Microsoft 365 at the user level. GitHub Copilot at the developer level. Azure OpenAI Service at the per token level. Each vehicle has its own buyer side levers. This article is the 2026 reference.
Microsoft sells artificial intelligence two ways at once, per user for Copilot and per transaction for Azure OpenAI, and the budget mistakes come from mixing the two models up.
Microsoft uses two pricing engines. Copilot products are per user per month. Platform services such as Azure OpenAI are consumption based and metered per token or per unit of reserved capacity. The two rarely appear on the same quote, which is exactly why budgets miss one of them.
Microsoft AI products, pricing model at a glance
| Product | Pricing model | Unit | Prerequisite |
|---|---|---|---|
| Microsoft 365 Copilot | Per user | 30 dollars per user per month | Qualifying M365 or O365 base |
| GitHub Copilot | Per user | 19 to 39 dollars per user per month | GitHub organization |
| Azure OpenAI Service | Consumption | Per 1,000 tokens or per PTU | Azure subscription |
| Copilot Studio | Consumption | Per message pack or pay as you go | Power Platform tenant |
The Copilot add on cannot stand alone. Each seat needs a qualifying base such as Microsoft 365 E3 or E5, Business Standard or Premium, or Office 365 E3 or E5. Microsoft sets the eligible bases on the Microsoft 365 Copilot page and the detail on the Copilot licensing documentation.
GitHub Copilot is a separate per user product with Business and Enterprise tiers. The tiers differ on policy controls and knowledge features. Current pricing sits on the GitHub Copilot plans page.
Azure OpenAI charges per 1,000 tokens, split into input and output rates, with the output rate usually higher. Microsoft lists the rates on the Azure OpenAI pricing page. Heavy steady traffic can move to reserved capacity instead.
Provisioned Throughput Units reserve dedicated capacity for predictable latency at a fixed price. Microsoft explains the model in the provisioned throughput documentation. They beat pay as you go only once token volume is steady and large.
Copilot Studio bills on messages. A message pack covers a fixed monthly volume, and a pay as you go meter handles overflow. Generative answers and multi turn agents consume more messages than a simple bot, so volume estimates drive the budget.
A pack suits predictable volume and a steady run rate. Pay as you go suits early pilots where volume is unknown. Most enterprises start on the meter, then move the stable baseline onto a pack once usage settles.
The standard reseller pitch is to license Microsoft 365 Copilot broadly on day one so the whole organization can adopt it together. We disagree. Across the engagements we ran, 2 to 5 times more seats were bought than reached steady weekly use inside the first quarter, which stranded real money on the annual commitment. The buyer side move is to license a measured pilot, instrument adoption, and expand against proven weekly active use. Treat the 30 dollar seat as a metered cost that must earn its place, not a flat rollout to everyone with a mailbox.
Source: Redress Compliance advisory engagement file, 2024 to 2025.
We cut the Copilot order from twelve hundred seats to four hundred, instrumented weekly active use, and expanded only against proven adoption. The first year cost fell by more than half.
Head of Digital Workplace · European industrial group
Microsoft 365 Copilot is 30 dollars per user per month on an annual commitment. It is an add on, so each user also needs a qualifying base license. Confirm the current price and eligible bases on the Microsoft 365 Copilot page before you budget.
Microsoft 365 Copilot requires a qualifying base such as Microsoft 365 E3 or E5, Business Standard or Premium, or Office 365 E3 or E5. Without an eligible base, the Copilot add on cannot be assigned. The base license is a separate line in the budget.
Azure OpenAI Service is priced per 1,000 tokens for pay as you go, split into input and output rates. Reserved capacity is sold as Provisioned Throughput Units with monthly or annual terms. The token model suits variable use, while provisioned throughput suits steady high volume traffic.
Provisioned Throughput Units reserve dedicated model capacity for predictable latency and a fixed price. You commit to a number of units for a monthly or annual term rather than paying per token. They make sense once token volume is steady and large enough to beat the pay as you go rate.
Copilot Studio bills on a message basis. A message pack covers a fixed volume of messages per month, and a pay as you go meter is available for variable use. Custom agents and generative actions consume messages, so estimate volume before you commit to a pack.
GitHub Copilot is sold per user per month, with Business and Enterprise tiers. Business sits around 19 dollars per user per month and Enterprise around 39 dollars per user per month. The tiers differ on policy controls, knowledge base features, and administrative reach.
Yes, and most enterprises do. Microsoft 365 Copilot and GitHub Copilot are per user, while Azure OpenAI and Copilot Studio are consumption based. The buyer side task is to forecast usage so the consumption meters do not drift past the cost of a flat per user plan.
Redress maps the AI use cases to the cheapest compliant licensing model, forecasts the Azure OpenAI and Copilot Studio consumption, and pressure tests the Copilot add on count against real adoption. Engagements run buyer side, never Microsoft paid, and feed the wider Microsoft program.
A buyer side reference on the Microsoft Enterprise Agreement renewal, including the Copilot line, the GitHub Copilot tier choice, the Azure OpenAI PTU evaluation, and the level shift mechanics.
Independent. Buyer side. Written for CFOs, CIOs, and procurement leaders carrying Microsoft renewals. No Microsoft influence. No sales kickback.
Open the white paper in your browser. Corporate email only.
Open the Paper →The Copilot deployment dropped from four thousand seats to twenty four hundred seats after the active usage audit. The freed budget funded the Azure OpenAI PTU commitment, which delivered the same downstream productivity at a quarter of the Copilot cost. The EA renewal closed nineteen percent below the Microsoft opening.
We have run 500+ enterprise clients across 11 publishers. Every engagement starts with one conversation.
Copilot patterns, GitHub Copilot tier signals, Azure OpenAI PTU lessons, EA renewal wins, and the wider Microsoft commercial leverage signals across every program we run.
Once a month. Audit patterns, renewal benchmarks, vendor commercial signals across Oracle, Microsoft, SAP, Salesforce, IBM, Broadcom, AWS, Google Cloud, ServiceNow, Workday, Cisco, and the GenAI vendors. No follow up sales pressure.
Free providers (Gmail, Yahoo, Outlook) cannot subscribe. Work email only. Unsubscribe in one click.
