Introducing Agent Effectiveness. Measure Factory usage across cycle time, work p

Factory, which sells AI coding agents it calls Droids, built Agent Effectiveness to show engineering leaders what their spending on those agents produces. Its announcement post of August 13, 2026 says the feature measures Factory usage against three things: cycle time, meaning how long a piece of work takes from start to finish; work priorities; and shipped artifacts.

Factory's fuller write-up, also recorded as a separate Intel entry, explains how. The feature reads from the issue trackers and code hosts a team already uses, naming Jira, Linear, GitHub and GitLab, and links each agent session to the issues, projects and pull requests those systems record. One view shows how project, issue and pull request cycle times change alongside agent usage. A second classifies sessions by intent, such as feature work, maintenance, bug fixing or exploration, so the split of spending can be compared with the quarter's plan. A third traces output back to specific sessions and users.

The result ties an organization's agent bill to particular pieces of work, with one limit Factory itself states: attribution and the intent view cover only the systems a team connects, so work tracked elsewhere drops out. A second limit comes from what the write-up leaves out. It describes usage and cycle times shown side by side, and says leaders can see where Droids improve delivery, but it does not describe a way to separate the agents' effect from other changes in how a team works. A faster cycle time in the dashboard is therefore a reason to look closer rather than proof the agents caused it. Factory described the feature as being in private preview, available on request.