Reference
Frequently asked questions
Answers to common questions about Performance Blocks across roles, plans, Henry, integrations, security, billing, and developers.
Quick answers to the questions we hear most often. For deeper detail, follow the link at the end of each answer.
Getting started
What is Performance Blocks for?
Performance Blocks is a continuous performance management platform built around discrete pieces of evidence ("observations" or "blocks"). Managers capture observations as they happen; the product aggregates those into conversations, summaries, and review cycles. The Agentic plan adds Henry, an AI agent that helps capture, draft, and analyze content. See the getting started overview.
How long does setup take?
A small org (under 25 people) can be live in a day. A larger org with HRIS sync, SSO, custom attributes, and a phased rollout typically plans 2–4 weeks. The biggest variable is how much organizational alignment you want on attributes and review cycles before launching. See getting started.
Do we need to migrate data from another tool?
Not necessarily — many orgs start fresh because observations are most valuable when captured live. If you want to import historical data, the Agentic plan's API keys can be used for a one-time bulk import. Contact support if you would like a guided migration.
For managers
Can an employee see their own observations?
No, not by default. Observations are private to the manager. Employees see content only when the manager explicitly shares a summary or conversation. This separation is intentional — it lets managers capture candid evidence without performance-review pressure on every entry. See observations privacy.
How many observations should I capture per direct report?
There is no hard rule. Healthy patterns we see: 1–2 observations per direct report per week, with a mix of strengths and opportunities. The point is recency, not volume — aim for "if I had to write a summary today, I'd have something to draw on." See observation cadence.
How does the AI draft my summary?
When you click Draft with AI, the product passes your selected observations (filtered by date range and attributes) to the LLM with a prompt asking for a structured summary. You always see and edit the result before saving or sharing. The AI never autonomously sends content to your direct report. See AI-drafted summaries.
Can multiple people edit a summary?
A single summary has one author at a time. To collaborate, share a draft externally (export as PDF or copy text), gather edits, and apply them in one author's session. We are evaluating real-time collaboration based on customer feedback. See summaries.
How do I delete an observation permanently?
From the observation detail view, click Archive. Archived observations are excluded from summaries, conversations, and AI drafts. To purge an archived observation permanently (for compliance or accidental capture of sensitive info), an org admin can run Settings → Data → Purge within 30 days; after 30 days, archived observations are auto-purged. See data retention.
For employees
What can my manager see that I cannot?
Your manager sees their own observations about you, the conversations they have authored, and any summaries they have drafted. You see only what they have explicitly shared with you in your employee portal. You can always see the audit log entries that involve you. See employee portal.
Can I respond to a shared summary?
If your org has the employeeFeedback flag enabled, yes — you can submit feedback on any shared summary or conversation. Your feedback is visible to your manager and recorded as a separate entry; it does not modify the original summary. See employee feedback.
Can I see who completed my 360 feedback?
If feedback was assigned via feedback assignment, the assigned reviewers' identities are visible to your manager but not necessarily to you — your org admin sets the visibility policy. Ad-hoc feedback your manager solicits is also subject to that policy.
Can I author my own objectives?
If your org has employeeObjectives enabled (which requires both objectives and employeePortal), yes. You can draft objectives in your portal; depending on your org's policy, they may need manager approval before going live. See employee objectives.
For admins
Can I change roles after invite?
Yes. Open Settings → Employees, select the employee, and edit their roles. Roles are stored as an array — an employee can be employee, employee + manager, or employee + manager + org_admin. Changes take effect on the employee's next page load. See employees.
What happens to observations when an employee leaves?
Archive the employee in Settings → Employees. Their observation history, conversations, and summaries are retained. They no longer appear in lists, no longer count as a seat, and cannot sign in. To permanently delete their data (for example, on a verified deletion request under GDPR/CCPA), use Settings → Data → Purge employee. See employee lifecycle.
How do I export org data?
Org admins can export from Settings → Data → Export. Available formats: CSV (per-entity files in a ZIP) and JSON (full snapshot). Exports include all observations, conversations, summaries, objectives, attributes, and audit logs scoped to your tenant. Large exports run asynchronously and are emailed when ready. See data export.
Can different teams in my org be on different plans?
No. Plan is set at the tenant (organization) level. If you genuinely need different feature sets per team, set up separate tenants. See plans and pricing.
How do I audit who toggled a feature flag?
Open Settings → Audit log and filter by event type feature.flag.toggle. Each entry shows actor, timestamp, the flag name, the previous state, and the new state. Audit log entries are immutable. See audit logs.
Henry Agent
Can I use Performance Blocks without Henry?
Yes — the Team plan deliberately ships without the Henry agent. The Team plan still includes AI-drafted summaries and conversations as in-app actions; they just are not exposed through a chat interface. See plans and pricing.
Does Henry train on our data?
No. Henry uses your data to answer questions and draft content for you within your tenant. We do not train models on customer data. Prompts and completions are not retained in training pipelines. For details, see our data and privacy commitments.
What languages does Henry support?
Henry's chat and drafting interfaces are available in the same languages the Performance Blocks UI supports (managed via Paraglide). Henry will respond in the language you write to it in for major European languages and several Asian languages. Knowledge base documents are processed regardless of language; cross-language retrieval works but with reduced precision. See Henry languages.
Why doesn't Henry know about something I just did?
Henry's RAG layer indexes new content within seconds for most surfaces, but documents added to the knowledge base may take a minute or two to be queryable. If Henry says it cannot find recent content, wait briefly and re-ask. If it persists, check Settings → Knowledge base for processing errors.
Can the API trigger Henry?
There is no public Henry chat API. Henry is designed for human-in-the-loop interaction. For programmatic content creation, use the Performance Blocks REST API to create observations, conversations, and summaries directly. The API endpoints accept the same inputs Henry would assemble. See API reference.
Can I see what Henry has been used for?
If you have the Agentic plan and henryAnalytics enabled, Settings → Henry → Analytics shows usage volumes, intent distribution, top users, and quality signals (thumbs up/down rates). Individual chat transcripts are not shown to admins by default for privacy. See Henry analytics.
Integrations
Which HRIS providers do you support?
We currently support Workday, BambooHR, Rippling, Gusto, HiBob, and Personio. Additional providers can be added on request. The integration syncs employees, reporting lines, departments, and active/inactive status. Compensation data is not synced. See HRIS sync.
Which SSO providers do you support?
Any SAML 2.0 or OIDC-compliant identity provider works. Tested examples: Okta, Microsoft Entra ID (Azure AD), Google Workspace, OneLogin, JumpCloud. See SSO setup.
Does Slack/Teams integration require admin approval?
Yes. Installing the Slack or Teams app requires the workspace admin's approval (standard Slack/Teams app installation flow). Once installed, individual users connect their Performance Blocks identity by signing in once. See Henry on Slack and Henry on Teams.
Security
Where is my data hosted?
Performance Blocks runs on Cloudflare Workers (compute), Neon PostgreSQL (database), and Cloudflare R2 (file storage). Data residency options exist for certain regions on enterprise contracts; contact sales. See data hosting.
Do you support 2FA?
Yes. Org admins can require two-factor authentication for all users or for specific roles (typically org_admin). 2FA uses TOTP (Google Authenticator, 1Password, Authy, etc.).
How long are sessions valid?
Default session length is 30 days, sliding (extended on each authenticated request). Org admins can shorten this in Settings → Security. Sessions are revoked on sign-out, password change, or admin-initiated revocation.
Are passwords stored?
Performance Blocks does not use passwords by default. Authentication is via SSO (when enforced) or email magic link. If you have configured a password fallback, passwords are hashed using industry-standard algorithms and never stored in plain text.
Billing
Is there a free trial?
No. Performance Blocks does not offer a self-serve free trial — the product benefits from configuration (employees, attributes, objectives) and we onboard customers through a guided demo instead. Book a demo at performance-blocks.com/demo. See plans and pricing.
What happens if I downgrade from Agentic to Team?
The downgrade applies at the end of your current billing period. Henry surfaces, knowledge base, Henry analytics, SSO, HRIS sync, API keys, and skip-level summaries become unavailable. All core data — employees, observations, conversations, summaries, objectives, attributes, audit logs — is fully preserved. Knowledge base documents are retained for 30 days in case you re-upgrade. See switching plans.
Can I get a refund if I cancel mid-period?
Monthly contracts are not refunded for unused days. Annual contracts may be eligible for pro-rated refunds at our discretion — contact your account manager. See plans and pricing.
Does my seat count include archived employees?
No. Only active (non-archived) employees count as seats. Archive an employee to release the seat at the end of the current billing cycle. See seat definition.
Developers
How do I get an API key?
API keys require the Agentic plan and the apiKeys feature flag. An org admin creates keys in Settings → Developers → API keys. Each key is scoped (read-only or read-write) and rate-limited. See API keys.
What's the API rate limit?
Standard plan limits are documented in API rate limits. If your integration needs higher throughput, contact sales — we can provision elevated limits for known production workloads.
Do you have webhooks?
Yes — Agentic plan, configured in Settings → Developers → Webhooks. Subscribe to events such as observation.created, summary.shared, employee.archived. Webhooks include a signature header you should verify. See webhooks.
Is there a CLI?
Not currently. The REST API plus shell tooling (curl, httpie) covers the use cases we have heard. If you would benefit from a CLI, let us know.