Originally published at Telefonica Tech
When Copilot first entered the enterprise market, the licensing conversation was relatively straightforward: identify the users who would benefit, purchase the appropriate licence and assign it to them. That conversation is now becoming more nuanced.
This article explains the difference between Microsoft Copilot licensing and consumption-based charging, including how Copilot Credits and pay-as-you-go models work, and what organisations should consider when managing Copilot usage at scale.
The key point is simple:
- Having a Microsoft Copilot licence does not mean every Copilot-related experience is automatically covered by that licence.
- Microsoft is increasingly combining subscription licensing with usage-based billing for certain AI experiences.
How does Microsoft Copilot Licensing and consumption-based charging work?
The easiest way to understand the change is to think about Copilot in two parts:
- User subscription licensing provides access to the core Copilot experience and capabilities included within the Microsoft 365 subscription.
- Usage-based billing applies to certain advanced AI experiences, where organisations pay according to the amount of AI work being performed.
Microsoft uses Copilot Credits as the common currency for eligible usage-based AI experiences. The number of credits consumed can vary depending on factors such as the models used, the context retrieved, the runtime required and the tools involved.
This creates an important distinction:
- Licence = access
- Consumption = payment for certain AI work
That distinction becomes particularly important as organisations move from individual Copilot users towards deploying AI agents across the business.
What are Copilot Credits?
Copilot Credits are the common currency used to measure eligible usage-based AI consumption across supported Microsoft experiences.
They can be used across services including Microsoft Copilot experiences, Copilot Studio, Power Platform workloads and Work IQ APIs. Microsoft supports different ways of funding this consumption, including prepaid capacity, pre-purchase plans and pay-as-you-go.
The amount of Copilot Credits consumed is linked to the work being performed rather than simply the number of users. This means two organisations with the same number of Copilot users could potentially generate very different levels of consumption depending on how they use agents and other advanced AI capabilities.
How do AI agents change Copilot consumption?
A traditional Copilot interaction might involve a user asking a question and receiving an answer. An agent can potentially do considerably more.
For example:
“Review our latest customer onboarding requests, identify anything missing, update the relevant records and notify the account team.”
The agent may need to retrieve information, reason over it, interact with other systems and perform actions. As the amount and complexity of work increases, consumption-based billing becomes more relevant. This also means that the number of Copilot licences alone no longer tells the whole story.
Organisations need to understand not just who has access to Copilot, but what AI workloads are being performed and how those workloads are being funded.
What does this mean for IT and procurement?
This shift introduces a new dimension to Microsoft licensing management. Historically, organisations could forecast much of their Microsoft 365 spend by looking at the number of licences assigned to users. With consumption-based AI, organisations also need visibility of what AI is doing and how much it is consuming.
Before deploying agents at scale, organisations should:
- Understand their Copilot estate: identify which Copilot products, agents and experiences are being used or planned.
- Identify the charging model: determine which capabilities are included within existing licences and which can generate additional consumption.
- Establish spending controls: put appropriate limits, policies and alerts in place before consumption scales.
- Monitor consumption: review usage by users, agents and workloads to understand where consumption is being generated.
- Measure business value: don't just ask how many Copilot Credits have been consumed, ask what the organisation has achieved as a result.
How can organisations govern Copilot consumption?
Understanding consumption is only part of the challenge. Organisations also need to establish appropriate controls around who can access usage-based Copilot experiences and how much they can consume.
Microsoft provides a Cost Management dashboard within the Microsoft 365 admin centre to help administrators govern usage-based billing. It allows organisations to govern access, establish spending controls, manage credit requests and monitor consumption.
Administrators can establish spending limits and alerts, while also managing how users and groups access supported usage-based services. Credit request processes can also be configured so organisations can determine how requests for additional access or spending are handled.
The dashboard provides visibility into consumption across areas such as users, groups, agents, services, spending policies and funding sources.
For organisations adopting Copilot at scale, this provides an important governance layer: control access, establish spending limits, manage credit requests and continuously monitor consumption, rather than simply reacting to usage after the bill arrives.
What should Microsoft customers do now?
If your organisation is already using Copilot, or planning a wider rollout, now is a good time to establish a clear view of your licensing and consumption model.
Ask:
- Which Copilot licences do we currently have?
- Which Copilot experiences are our users accessing?
- Which agents are being deployed?
- Which workloads can generate consumption?
- How will that consumption be funded?
- Is pay-as-you-go appropriate for our requirements?
- Who is responsible for managing AI spend?
- Do we have appropriate spending controls?
- Can we demonstrate the value we're getting from our AI investment?
The future of Copilot isn’t simply about who has a licence. It’s about what AI is doing, how much it is consuming, how that consumption is charged and what value it delivers.
Telefónica Tech can help organisations assess their Microsoft licensing and Copilot environment, understand potential consumption-based costs and develop an approach to managing Microsoft AI at scale.
Talk to our Microsoft licensing specialists and understand what the changes to copilot could mean for your organisation →
