Copilot Chat and agents: licensing and billing

Copilot Chat access, agent creation and agent usage are separate licensing questions. This guide helps distinguish user entitlements from consumption-based charges and plan a controlled rollout.

Distinguish between chat availability, permission to work with company context and consumption billing of a specific feature. The Copilot screen name does not address these questions.

In the current documentation, Microsoft also switches to the names Microsoft Copilot and Microsoft Copilot Chat. Some licenses and interfaces still use Microsoft 365 Copilot. So please use the exact product when ordering, not just the app name.

What basic Copilot Chat solves

Copilot Chat provides conversational work with AI and web context. Options for working with organizational content vary by specific access and license. Uploading a file yourself cannot be exchanged for the same right to work with the entire company context.

The current Copilot Chat FAQ differentiates chat and paid experience with a work context. Details privacy and protections also describe selected methods of adding organizational content.

Therefore, when piloting, don't just ask "does Copilot work?". Test a specific task: working with an attached document, finding information in team documents or another required function.

Data protection is not the same issue as functional scope

Work login and enterprise data protection are important. But they do not mean that every available AI function has the same access to data or the same cost.

Microsoft provides enterprise data protection terms and related differences in control by subscription. Check with the company for work identity and intended use.

Prepare a clear rule for users: what account they use, what information they are allowed to enter and who checks the result. The license approval itself does not replace the rules of working with internal data.

Agents do not have one universal model

An agent can reply from public pages, read shared company resources, or perform other actions. Each case needs its own assessment.

In the agents documentation for Copilot Chat, Microsoft distinguishes between agents available at no additional cost and agents with consumption billing. Specific conditions must be reviewed for access to shared tenant data.

Do not automatically transfer the conditions of a simple internal agent to autonomous processing or an external audience. Also note the engine, extensions used and services outside of Copilot itself.

Decision table

The table is a design check. It is not a price comparison and does not make a claim without knowledge of the specific product.

User license and consumption can be complementary

Documentation updated in September 2026 differentiates between user subscription license and usage-based billing for certain AI scenarios. Overview of USL and UBB also explicitly describes the need for a billing policy for usage-based experiences.

This leads to a practical requirement: prepare the budget according to the intended task and specific function. Don't use the blanket statement "Paid Copilot covers everything".

In the calculation, distinguish between access license, consumption accounting and related services. Specify an owner for each row. Otherwise, the product team may be tracking one invoice while operational consumption appears elsewhere.

Model example: agent for internal guidelines

A hypothetical firm creates an agent that answers employee questions from shared guidelines. The benefit should be to find a specific rule faster.

Before deployment, the team selects approved resources, prepares common questions, and verifies permissions. It divides the tests according to job roles. The agent must not derive the correct answer from a document to which the employee is not supposed to have access.

At the same time, the licensing and billing model for the intended audience is recorded. Test authors with other licenses is not enough. The pilot must also use representative accounts of future beneficiaries.

How to verify expert agent responses

Prepare the right documents for known questions, conflicting documents, and out-of-scope questions. Test if the agent recognizes when it doesn't have a reliable answer.

For directives and licences, the output needs a date, a clear source and the right context. A naturally written answer may be wrong. Therefore, the verification must not depend on the impression of the fluency of the text.

For each error, determine whether it is a missing resource, inappropriately shared document, outdated rule, or incorrect use of information. Addressing the cause tends to be more helpful than another general "get it right" instruction.

What to measure in the pilot

• Completed work assignments and correctness of answers.
• The need for human corrections.
• Availability and quality of cited documents.
• Results under different user roles.
• Real consumption and its allocation.
• Behavior when resource is unavailable or conflicting.

The number of conversations alone does not demonstrate business benefit. What matters is whether people manage to complete a specific job correctly.

Prepare cost management before expansion

Determine the owner of the billing policy, the allowed scenarios and the method of evaluating the consumption. For cost alerts, verify whether this is information or a real limitation. Do not confuse these functions.

Separate the pilot budget from future operation. Testing, knowledge enhancement, and subsequent integrations may have a different profile than normal employee use.

Re-assess before any significant expansion. A new data source or autonomous trigger can also change the licensing and financial model.

What should be stated in the proposal

Exact task, specific product, users, data sources, permissions and billing model. With this basis, the company compares the benefits and costs without expectations that are based only on similar names.