September 17, 2026: GitHub announced on September 16 that Copilot budget increase requests are generally available for Copilot Business and Enterprise under usage-based billing. Members who exhaust their available AI credits can request a higher budget when they reach the limit. Source: GitHub announcement.
What changed
Credit-consuming Copilot features remain blocked after the available credits run out. The new request flow gives members a way to seek additional budget, while administrators retain the decision to approve, adjust or deny it.
Requests follow the paying account: organization-owned budgets are handled in organization settings, and enterprise-owned budgets in enterprise settings. GitHub says approving an increase immediately restores access to AI credits.
Who handles approval?
Organization owners, enterprise owners and billing managers review pending items under Requests from members. They enter a new budget amount and select Approve and increase. The entered value becomes the member’s budget; it is not described as an extra amount added to the existing limit. Source: GitHub’s budget request guide.
Why this matters for QA engineers
For a tester using Copilot to investigate a regression or draft automation changes, exhausted credits can interrupt the investigation. This release provides an explicit escalation path, but does not promise automatic approval or completion of an interrupted task.
Our practical recommendation: identify the budget owner before a release-critical investigation and save the failing test, logs and current hypothesis in a shared work item. After approval, verify access and the investigation’s state before continuing. Restored credits alone are not evidence that a previous agent action finished or that a generated test passed.
Teams evaluating the workflow should check that requests reach the correct paying account and that the approved total matches the intended budget. These are QA rollout checks, not new testing capabilities announced by GitHub.