Companies Can Now Cap AI Agent Spending by Team
Organisations using Devin can distribute budgets across business units, and engineers who approach their limits receive alerts and can request more within the product.
Updated on Nov. 4, 2026 in Budgeting
Cognition has introduced hierarchical budgets for its Devin AI coding agent, letting companies assign spending limits to teams and track usage across business units. The feature targets organisations deploying Devin beyond single teams, where spending controls must be distributed across departments.
Why it matters
As AI coding agents move into broader enterprise deployments, companies need ways to control and forecast consumption costs across many groups rather than one team at a time. The feature emerged from customer conversations, and early single-team controls did not scale when Devin moved into wider organisational use.
The vendor describes budget allocation nesting up to three levels deep, from overall budget to business units to teams and projects. No pricing, cost-per-unit figures or adoption numbers were disclosed in the available information.
The players
Cognition
The AI software company behind Devin, whose enterprise tools can add usage-based costs for organisations deploying its coding agent.
The details
An administrator starts with an overall budget and distributes it among business units, which can allocate funds to teams and projects up to three levels deep. Separate spending controls apply to Devin running in the cloud and on engineers' desktops. Engineers see homepage alerts when they are on track to exceed a limit and can request more Agent Compute Units from the budget owner inside the product. Administrators can enable efficiency-based approvals that automatically raise an engineer's limit up to a set ceiling, while central IT can set limits without approving each request. If a budget runs out mid-task, the Devin session stops.
Timeline
October: Cognition introduced its Efficiency Score, which estimates productive engineering hours relative to Devin consumption.
4 November 2026: Article published on hierarchical budgets for Devin.
Money Landscape
Consumption-based pricing for AI agents means software costs can rise with usage rather than with seats, so budgeting tools are becoming part of how organisations forecast tech spending. Cognition's efficiency measure and nested budgets reflect this shift toward tracking the cost of each unit of work.
Organisations using usage-based AI tools may need to review budgeting and approval processes so that agent consumption stays within planned technology spend. Teams that run out of budget mid-task will see sessions stop, so budget owners should track limits before projects are underway.
The takeaway
Spending controls for AI agents are becoming as important as the tools themselves, especially when usage is billed by consumption. Organisations deploying such tools should check how budgets are allocated across teams before expanding use.
Further reading
For broader context on setting and tracking spending limits, see the Budgeting section.
Source note: This article includes information reported by ETCIO.
