Factory released an analytics layer for teams deploying coding agents, surfacing usage, tool calls, activity, and productivity from tokens through pull requests. Use it if you need ROI, readiness, and cost visibility as agent adoption scales.
Factory is packaging agent telemetry into an engineering analytics product, not just a dashboard for model spend. In Factory's launch thread, the company frames the core problem as proving how agent activity turns into software outcomes, and the product post expands that into a pipeline spanning tokens, tool calls, sessions, files, commits, and pull requests.
The technical payload is in the dimensions it exposes. According to the product post, teams can break down token use by model, user, date, billable versus cached tokens, and input versus output volume; inspect tool usage across skills, slash commands, hooks, and MCP servers; and track adoption through active users, session counts, client breakdowns, and a "stickiness ratio." Factory also surfaces per-user metrics and an "agent readiness" layer meant to show where environments, repos, and workflows are prepared for more autonomous use.
The screenshot in Factory's launch materials suggests the product is built for organizations already operating agents at scale: the Tools view highlights "tool calls," skills, slash commands, and a 30-day usage trend rather than a single prompt log.
The timing makes sense because the measurement problem has changed. As one engineer's post puts it, "the bottleneck has so quickly moved from code generation to code review," which means raw acceptance rates or anecdotal demos no longer capture the real operational constraint.
Factory is effectively pitching a control plane for that next phase. Its launch thread says leadership needs visibility into "adoption, cost, and output" when thousands of engineers start using agents, and the product post adds metrics like autonomy ratio, session behavior, and PR output that can show whether usage is concentrated in experimentation or translating into shipped changes. That matters more when adoption numbers start to look like the Uber repost, which claims "1,800 code changes per week" are now written entirely by agents.
The broader caveat is that instrumentation does not solve organizational redesign by itself. In a related reaction, Ethan Mollick's thread argues AI application rollouts are "far less of a technical issue" than a question of structure and decision-making, a useful reminder that analytics can quantify agent-native development without answering how teams should govern it.
Claude can now drive macOS apps, browser tabs, the keyboard, and the mouse from Claude Cowork and Claude Code, with permission prompts when it needs direct screen access. That makes legacy desktop workflows automatable, and Anthropic is pairing the push with more background-task support for longer agent loops.
releaseOpenClaw shipped version 2026.3.22 with ClawHub, OpenShell plus SSH sandboxes, side-question flows, and more search and model options, then followed with a 2026.3.23 patch. Teams get a broader plugin surface, but should patch quickly and review plugin trust boundaries as the ecosystem grows.
releaseCursor shipped Instant Grep, a local regex index built from n-grams, inverted indexes, and Bloom filters that drops large-repo searches from seconds to milliseconds. Faster candidate retrieval shortens the coding-agent loop, especially when ripgrep-style scans become the bottleneck.
breakingChatGPT now saves uploaded and generated files into an account-level Library that can be reused across conversations from the web sidebar or recent-files picker. It removes repetitive re-uploading and makes past PDFs, spreadsheets, and images part of a persistent working context.
breakingEpoch AI says GPT-5.4 Pro elicited a publishable solution to one 2019 conjecture in its FrontierMath Open Problems set, with a formal writeup planned. Treat it as an early milestone worth reproducing, not blanket evidence that frontier models can already automate math research.
Today we’re launching Factory Analytics. Enterprise teams can now see exactly how AI agents translate into engineering outcomes: tokens → usage → commits → pull requests → shipped software. The missing layer for proving ROI in agent-native software development.
The bottleneck has so quickly moved from code generation to code review that it is actually a bit jarring. None of the current systems / norms are setup for this world yet.
