Claude Enterprise adds smart reports, an AI-generated analytics view of how teams use Claude

Anthropic has launched smart reports for Claude Enterprise, a beta analytics feature that turns raw usage data into a written report on how a team is actually using Claude — not just usage counts.

What's new

Anthropic describes the feature directly: "We've launched smart reports, which analyze how a team uses Claude and report on the work getting done, what it costs, where sessions run into friction, and which repeated patterns are worth packaging as shared skills." Rather than a dashboard of raw metrics, the output is a synthesized narrative report covering four things: what work is happening, what it's costing, where users are hitting friction, and which repeated request patterns are common enough to be worth turning into a reusable shared skill.

Access requires Owner or Primary Owner permissions to enable, via Organization settings > Capabilities > Analytics, where the "Smart reports (beta)" toggle is off by default. Once enabled, Primary Owners, Owners, Admins, and custom roles with analytics-view access can generate and read reports. During the beta, each organization is capped at 10 reports per month, resetting on the first of each calendar month. The feature is not available to organizations using customer-managed encryption keys, HIPAA configurations, or Access Transparency.

Context

This lands alongside a run of Claude Platform admin and governance tooling shipped this month — server-evaluated auto permission policies for Managed Agents, the ant apply infrastructure-as-code command, and per-message effort controls on Vertex AI — all aimed at giving enterprise administrators more visibility and control over how Claude is deployed inside their organizations, not just more raw capability in the model itself.

Why it matters

Usage dashboards are common; a report that identifies which recurring request patterns are worth converting into a shared, packaged skill is a more opinionated and more useful artifact for an IT or platform team trying to scale AI adoption without proliferating one-off prompts. It also gives Anthropic a lever most competitors don't have as sharply defined: a built-in feedback loop that nudges ad hoc usage toward the platform's own reusable-skill abstraction, reinforcing the skills/plugins ecosystem Anthropic has been building out across Claude Code and Claude Platform this month.