Factory Analytics shows engineering leaders where AI coding spend goes

Consumption broken down by model and user

Factory positions the update as a response to a budget question. In its words, agentic software engineering is "a growing line item for engineering leaders and CFOs", and platforms earn trust by showing where spend goes and keeping it in check, without making engineers manage their own usage.

The Consumption view reports credits over time, by model and by user. Factory says it reports consumption at the level where decisions get made: which model ran and who ran it. Several charts support this:

• Usage Over Time separates total consumption from the per-model split.
• Model Mix shows each model's share of usage.
• Cost Concentration plots cumulative consumption across the organization.

When consumption spikes, teams can identify the cause directly instead of asking the account team for a report, according to Factory. The company also says that in most deployments a smaller group of power users and automations accounts for the majority of consumption. The Cost Concentration chart lets leaders check that this concentration maps to the teams and workloads they expect.

Model efficiency and the Factory Router

Factory argues that controlling consumption should not require engineers to ration prompts or cut sessions short. Its answer is Factory Router, which selects an efficient model for each request automatically. According to Factory, the router reduces aggregate cost by 63% in production compared with frontier-model rates, while holding frontier-level performance in Factory evaluations.

Analytics reports the estimated savings for each organization. Enterprise Controls let leaders set model defaults and policy centrally.

Adoption data for rollout decisions

The Adoption view reports daily and monthly active users and sessions, showing how usage is spreading across an organization. Combined with the Consumption view, leaders can see both who uses Factory and what that usage consumes.

Factory says this helps leaders see who is driving usage and whether adoption is keeping pace, so enablement can be directed where it is needed. Budget owners also get the adoption data behind decisions to expand a rollout. Because consumption remains visible by user and by model, leaders can expand access without restricting how engineers work, the company says.

Availability and verification

Every card in Analytics includes a tooltip that defines the metric and explains how it is calculated. Enterprise organizations that report in their own BI tools can pull usage by model and user through the Analytics API.

Analytics is available today for organizations on Business and Enterprise plans. Owners and Managers can open it from the profile menu in Factory.

For engineering leaders and budget owners, the release is aimed at a practical need: tying AI coding spend to specific models, users and teams, with the data exportable to existing reporting tools. The 63% cost reduction and the performance claim are Factory's own figures, based on its production data and evaluations.