Microsoft is pairing its per-user Copilot license with usage-based billing for agentic and frontier-model work. The move gives buyers a predictable floor for broad deployment but shifts the hardest forecasting problem, the cost of long-running agents, onto enterprise FinOps teams.
What Is Covered in This Article:
- Microsoft’s September 25, 2026 update to the Copilot pricing model, which splits “everyday AI” and “advanced AI” into separate commercial structures.
- The FinOps and admin controls Microsoft is rolling out to govern Copilot Credit consumption.
- What the hybrid structure means for enterprise software buyers budgeting for agentic AI.
The News: In late September, Microsoft outlined several changes to the Copilot pricing model. “Everyday AI,” covering tasks such as summarizing documents or turning meeting transcripts into bullet points, stays under the Copilot user subscription license (USL), priced per user per month. Model selection is part of the USL, with GPT-5.6 and Sonnet 5 included and new models such as Opus 5 included with limits. Microsoft says fair-use limits are set so “the vast majority” of users never reach them; users who do receive a warning and can switch to Auto at no added cost or move to credits.
“Advanced AI,” which Microsoft defines as long-running agents and unsupervised workflows, including Cowork, Code, Autopilot, new advanced experiences in SharePoint, and frontier models such as Fable and Astra, runs on usage-based billing (UBB) with Copilot Credits. UBB requires the USL. For enterprise customers, UBB services stay off until an admin creates a spending policy in the Microsoft 365 admin center.
Microsoft also announced FinOps additions: cost management for Code and Managed Runtime alongside Cowork and the Work IQ APIs (with Copilot Studio agents coming in October), group-level model access controls in Cowork, Cowork consumption insights, Microsoft Graph API management of spending policies, and end-user visibility into credit usage. Microsoft said some capabilities will start on UBB and move into the USL as the economics allow.
Microsoft Splits the Copilot Pricing Model Into Everyday and Advanced AI
Analyst Take: Microsoft is formalizing what enterprise AI vendors have been circling for two years: a seat for breadth, a meter for depth. The more interesting part of the announcement is the justification. Microsoft’s argument rests on what it calls saturation, the point at which a more capable model no longer changes the result because the task never needed it.
That is a defensible line to draw. Summarizing a teammate’s document does not get meaningfully better with a frontier model running at high reasoning effort, and pricing it as if it did would push companies to ration access. The catch is that Microsoft decides where the saturation line sits, and fair-use limits give it a lever to move that line over time.
The Copilot Pricing Model Takes Aim at Metered Rivals
The post takes a direct swipe at “some vendors” that meter everyday AI work, arguing that metering leads companies to ration access. Microsoft backs this with an internal cost comparison modeling Copilot against Claude and ChatGPT across 40 prompts and two user profiles.
Buyers should treat that comparison as directional. It is vendor-modeled, uses list pricing and API rates, excludes negotiated discounts, and compares Copilot Cowork at variable rates against competitors’ public pricing. Notably, though, Copilot’s USL and UBB both run on OpenAI and Anthropic models, which makes Microsoft a reseller and competitor of the labs whose products it is benchmarking against.
Predictable Floor, Variable Ceiling
For CIOs and procurement teams, the USL keeps the budgeting model they already know for the broad workforce. The harder problem sits on the UBB side. Microsoft’s own September 2026 Copilot Credits Guide lists pay-as-you-go credits at $0.01 each, with pre-purchase discounts of up to 20%, but notes that Cowork consumption varies with model, runtime, context, and tools.
In practice, finance teams can cap spend but cannot easily forecast what a given agentic task will cost before it runs. A multi-step financial model “running all afternoon,” to use Microsoft’s example, could consume very different credit volumes depending on configuration.
Microsoft deserves credit for making UBB default-off for enterprise tenants and for letting admins map consumption to Azure subscriptions and resource groups. That puts AI spend inside existing chargeback structures rather than in a separate line item nobody owns. Graph API access to spending policies is also a practical addition for organizations that already run FinOps tooling.
Usage-Based Is Not Outcome-Based
The Copilot pricing model is still priced on effort, not results. Credits track compute, model choice, and runtime, meaning an agent that spends four hours producing a flawed output costs the same as one that nails the task.
The new Cowork consumption insights, which show what scenarios drive credit use and “what they produced,” are a step toward connecting spend with value. But the pricing itself does not reflect outcomes, and that places the burden of proving ROI squarely on the customer. As outcome-based pricing gains traction across enterprise software, Microsoft’s hybrid could look conservative within 12 to 18 months.
What Buyers Should Do Now
The sensible starting point is to focus on desired business outcomes, and then identify which workflows justify advanced AI, measure incremental gains such as hours saved or cycle time reduced, and only then set spending policies around those workflows.
Microsoft’s stated plan to move capabilities from UBB into the USL over time is also a negotiating point. Buyers signing multi-year agreements should seek commitments on how and when that migration happens, rather than leaving it to Microsoft’s discretion.
What to Watch:
- Whether Microsoft tightens USL fair-use limits or narrows model access, which would push more heavy users toward credits and erode the predictability that anchors the everyday AI pitch.
- How OpenAI, Anthropic, and Google respond on enterprise pricing, particularly whether any competitor adopts a flat-rate tier for routine work to neutralize Microsoft’s “no meter” argument.
- Whether Microsoft links Copilot Credits to measurable outcomes or publishes per-task cost benchmarks once Copilot Studio agents come under cost management in October, which would signal a shift from effort-based to value-based pricing.
For more information, see the press release or article on the vendor’s website.
Sources
1. MCN_Microsoft_Copilot_Pricing_Model_2026-09-29.md
Declaration of generative AI and AI-assisted technologies in the writing process: This content has been generated with the support of artificial intelligence technologies. Due to the fast pace of content creation and the continuous evolution of data and information, The Futurum Group and its analysts strive to ensure the accuracy and factual integrity of the information presented. However, the opinions and interpretations expressed in this content reflect those of the individual author/analyst. The Futurum Group makes no guarantees regarding the completeness, accuracy, or reliability of any information contained herein. Readers are encouraged to verify facts independently and consult relevant sources for further clarification.
Disclosure: Futurum is a research and advisory firm that engages or has engaged in research, analysis, and advisory services with many technology companies, including those mentioned in this article. The author does not hold any equity positions with any company mentioned in this article.
Analysis and opinions expressed herein are specific to the analyst individually and data and other information that might have been provided for validation, not those of Futurum as a whole.
Read the full Futurum Group Disclosure.
Other Insights From Futurum:
Autonomy Over Analytics: The Read-Write Decree Rewiring Enterprise Data Platforms
Can Salesforce’s Agentic AI Close Field Service’s Talent and ROI Gap?
Microsoft Copilot Hits 30 Million Seats: Is AI Transformation Finally Real?
Author Information
Keith Kirkpatrick is VP & Research Director, Enterprise Software & Digital Workflows for The Futurum Group. Keith has over 25 years of experience in research, marketing, and consulting-based fields.
He has authored in-depth reports and market forecast studies covering artificial intelligence, biometrics, data analytics, robotics, high performance computing, and quantum computing, with a specific focus on the use of these technologies within large enterprise organizations and SMBs. He has also established strong working relationships with the international technology vendor community and is a frequent speaker at industry conferences and events.
In his career as a financial and technology journalist he has written for national and trade publications, including BusinessWeek, CNBC.com, Investment Dealers’ Digest, The Red Herring, The Communications of the ACM, and Mobile Computing & Communications, among others.
He is a member of the Association of Independent Information Professionals (AIIP).
Keith holds dual Bachelor of Arts degrees in Magazine Journalism and Sociology from Syracuse University.
