Introduction
Agents and workflows built on the GitHub Copilot harness in Microsoft Copilot Studio can perform complex, multistep tasks autonomously. That capability is metered through usage-based billing in Copilot Credits, where consumption scales with the complexity of the work an agent performs. When you understand that relationship, cost becomes something you shape deliberately as a maker rather than a number your organization reconciles after the fact.
The same decisions that make an agent capable also shape what it costs to run. The model that powers each task, how much organizational context the agent draws in, how many tools and actions it calls, and how complex its orchestration becomes all move consumption up or down.
In this module, you explore:
- How billing works across the GitHub Copilot, standard, and Copilot chat harnesses, and when consumption begins on each.
- The maker design decisions that drive runtime consumption on the GitHub Copilot harness.
- The build-time activities that consume Copilot Credits before you publish.
- How administrators fund and monitor capacity across the tenant.
By the end of this module, you're able to build agents on the GitHub Copilot harness with cost awareness, reasoning about what drives consumption and collaborating with administrators on capacity planning.