Claude Enterprise: Smart Reports, Surveys and Budgets
Claude Enterprise adds smart reports, in-app surveys and pooled budgets. See setup, 10-report limits, privacy rules and spending controls.
On October 28, 2026, Anthropic announced three generally available administration features for Claude Enterprise: smart reports, in-app surveys, and pooled group budgets. Together they address a common enterprise AI problem: knowing which work is happening, hearing whether employees find it useful, and controlling what the organization spends. The three tools measure different things, however. None by itself proves a productivity gain or guarantees that a workflow should be automated.
Three tools, three different decisions
Choose the right Claude Enterprise control
| Feature | What it reveals or controls | Important limit |
|---|---|---|
| Smart reports | Sampled Claude sessions, workstreams, deliverables, costs and friction | Up to 10 reports monthly at no charge; not for employee performance reviews |
| In-app surveys | Users' self-reported value, time saved and obstacles | Identifiable responses; HIPAA-configured organizations excluded |
| Pooled budgets | Shared monthly spending by group plus member guardrails | Usage credits and seat-included usage are treated differently |
Smart reports analyze a sample of transcripts from a selected team, time window and Claude products. They group work into categories, surface repeated problems such as missing connectors, and suggest skills or reusable workflows. Surveys ask users about their own experiences at natural breaks in chat, Cowork or Claude Code. Pooled budgets let a team draw from one shared allowance while retaining controls over individual spending. Treat the report as usage evidence, the survey as feedback, and the budget as a spending mechanism.
How to enable and scope smart reports
- Confirm the Enterprise configuration is eligible: smart reports exclude CMEK, HIPAA and hybrid inference deployments.
- Have a Primary Owner or Owner enable Smart reports under Organization settings > Capabilities > Analytics; the feature is off by default.
- Open Analytics > Smart reports and select a team, a time range within the past 28 days, and the Claude products to include.
- Review the sampled-session scope, choose questions or custom insights, then create the report and wait for processing.
- Use the results to identify a repeated workflow or connector problem, and verify it with the team before making changes.
Anthropic currently allows up to 10 smart reports per organization per calendar month at no charge. Queued and in-progress reports consume the allowance, while failed reports do not; the quota resets on the first day of the following month. Reports can take hours to generate. Owners can delegate scoped access to team leads without granting full organization administration. If Claude Code has zero data retention enabled, its sessions are excluded from these reports even when chat and Cowork remain available.
Anthropic explicitly says smart reports are for understanding adoption and investment, not evaluating individual employee performance or making employment decisions. Restricted-topic sessions receive aggregate-only handling, and administrators control whether member names can be revealed in eligible drilldowns. Sampling and model-generated categorization also make reports unsuitable as a complete productivity audit.
Surveys: useful feedback, but not anonymous
Enterprise admins can schedule a survey for all members or selected groups, choose chat, Cowork or Claude Code, and customize questions. A user sees a survey at a natural break rather than in the middle of a task. Results include response counts, charts and written answers, and can be exported as CSV. That export contains respondent emails and groups, so teams should not describe surveys as anonymous. Anthropic's current survey documentation supports CMEK while excluding HIPAA configurations; this differs from smart reports' eligibility and should not be conflated.
Pooled budgets: share spend without removing guardrails
A pooled budget is one monthly amount available to members of a group. It supplements rather than replaces member spending limits. For usage-based Enterprise plans, a request counts against the individual's limit and the applicable group pool. For Standard and Premium seat plans, pooled budgets require usage credits and do not count seat-included usage. When the pool is exhausted, users with no other available pool cannot consume additional pooled usage until an administrator raises the budget or the month resets. Budget managers receive threshold notifications, but the pool is not a substitute for monitoring usage.
A sensible pilot uses one team with a measurable workflow and a clear budget owner. Start by reviewing sampled report findings, ask the team whether the identified friction is real, then choose one fix such as enabling a missing connector or documenting a reusable skill. Compare adoption, request volume and spending after the change. A survey's reported time savings are subjective estimates, not an independently verified ROI number; pair them with observed output quality and actual operating costs.
A practical rollout checklist
- Verify plan, HIPAA, CMEK and hybrid deployment eligibility separately for each feature.
- Explain survey visibility and CSV identity fields before inviting employees to respond.
- Use team-level reporting and documented privacy controls; never treat sampled transcripts as a performance scorecard.
- Set a shared pool based on actual recent usage and keep a separate member limit as a guardrail.
- Review one recurring workflow, one friction point and the cost trend before expanding to more teams.
The October launch is relevant to operations, RevOps and AI enablement teams because it connects three otherwise disconnected decisions: what Claude is used for, what employees say is valuable, and what it costs. Actual feature eligibility still depends on organization configuration and billing plan. Teams should consult the current Claude Help Center before committing to a deployment or promising a particular financial outcome.
Questions people ask
Can every Claude Enterprise organization enable smart reports?
No. Smart reports are off by default and require an owner to enable them in Organization settings. Anthropic excludes Enterprise organizations using customer-managed encryption keys (CMEK), HIPAA configurations, or hybrid inference deployments. Organizations using zero data retention for Claude Code can still report on chat and Cowork, but Claude Code sessions are excluded. Check your configuration before planning a rollout.
How does Claude's 10-report monthly limit work?
Each eligible Claude Enterprise organization can run up to 10 smart reports per calendar month at no charge. Reports that are queued or still processing count toward that allowance; failed reports do not. The allowance resets on the first day of the next calendar month. If the organization needs more before the reset, Anthropic says to contact its account team rather than assuming unlimited free reporting.
Are Claude Enterprise survey responses anonymous, and can they be exported?
No. Admins can export survey results as CSV, and the export includes each respondent's email, group, submission time, and answers. Surveys support customer-managed encryption keys according to Anthropic's current survey help page, although HIPAA-configured organizations are excluded. Because responses are identifiable and can contain sensitive free text, administrators should disclose their use and avoid collecting regulated information.
Do pooled group budgets replace individual Claude spending limits?
No. A pooled budget is a shared monthly allowance layered on top of each member's own spend limit. On usage-based Enterprise plans, requests count against both controls, so either can stop additional usage. On Standard or Premium seat plans, pools apply to usage credits, not usage included with the seat. Budget administrators should monitor both group consumption and individual limits rather than dividing the pool into rigid equal shares.
Claude Enterprise smart reports, surveys and pooled budgets can improve visibility and spending discipline when used together. Start with eligible team-level reporting, collect informed feedback, and put shared budgets alongside individual guardrails. Avoid treating generated insights or survey estimates as proof of employee productivity.
Explore related Make Better resources
- Claude Frontier Academy enterprise AI deployment— Related adoption strategy, distinct intent
- Claude Dashboards and Motion business workflows— Related workflow reporting, distinct product
Sources & useful resources
- Anthropic: Smart reports, surveys and pooled group budgets (Oct 28, 2026)— Primary launch, three features and configuration caveats
- Claude Help: Get started with smart reports— Eligibility, quotas, reporting scope, privacy and roles
- Claude Help: Create surveys for your organization— CMEK support, HIPAA exclusion, CSV identity fields
- Claude Help: Manage pooled group budgets— Plan eligibility, shared and member limits, usage-credit rules