Everything you need to master Clarion

Once your signals are processed, go to What to Build? in the sidebar. Type a question like “What should we build next?” and Clarion returns ranked, evidence-backed recommendations with customer quotes, revenue impact, and confidence levels.

The complete flow from customer voice to shipped code:

• Ingest — Signals arrive via CSV, webhook, or integration sync.
• Process — Clarion AI automatically classifies, extracts themes, scores sentiment, computes importance, and indexes everything for semantic search.
• Synthesize — The AI engine clusters related insights into opportunities and scores them with PM frameworks.
• Decide— The “What to Build?” engine recommends priorities. The Knowledge Explorer lets you ask questions against your entire dataset.
• Specify — Spec Studio generates PRDs and machine specs with agent artifacts (.cursorrules, tasks.json, test scenarios, agents.md).
• Ship — Agent Bridge exposes specs via MCP, pushes artifacts to GitHub, and exports to Jira/Linear.
• Measure — Feature-opportunity linking tracks outcomes, the learning system calibrates future recommendations, and the executive dashboard reports ROI.

1. 1

   Welcome & Overview

   A quick product tour showing Clarion’s core value proposition — how signals become insights, opportunities, specifications, and shipped features. Includes an animated walkthrough of the complete data pipeline.
2. 2

   Connect Your Data

   Choose how to get signals into Clarion: CSV Upload (quickest — drag and drop a file), Zendesk/Intercom/Gong (live integrations), or API/Webhook (custom pipelines). You can connect multiple sources.
3. 3

   Set Up Product Hierarchy

   Define your product structure: create products, add feature areas, and map your team’s vocabulary. This hierarchy powers AI classification — signals auto-route to the right product area. You can also skip this step and let the AI suggest a hierarchy after processing your first signals.
4. 4

   Review Your Context (Owners & Admins)

   Once a source is connected, see the personas, decisions, goals, and features Clarion auto-extracted from your data — review and confirm them within minutes, not weeks. A “Try These Questions” panel drops you straight into chat to test it.
5. 5

   Interactive Tour

   A spotlight-based guided tour that highlights each section of the interface — sidebar navigation, signal feed, intelligence dashboard, and AI chat. Clickable hotspots explain each feature in context. Dismiss at any time; replay from Settings → General.

After the wizard, the Dashboard shows a persistent Getting Started Checklistwith progress tracking. Items include: connect your first integration, import 50+ signals, review AI insights, explore an opportunity, generate your first spec, and invite a teammate. Items auto-complete as you perform each action. Dismiss the checklist once you’re comfortable.

1. 1

   Paste Your PRD

   Navigate to /translate(or click “Spec Translator” on the landing page). Paste any product requirements document, user story, or feature brief into the editor. Supports plain text, markdown, and rich text.
2. 2

   AI Translation

   Click “Translate to Machine Spec”. The AI produces a SPEC.md— every vague requirement is quantified: “fast” becomes “< 200ms p95”, “responsive” becomes specific breakpoints, “secure” becomes OWASP compliance requirements.
3. 3

   Download & Use

   Copy the generated SPEC.md or download it. Drop it into your project directory and point Cursor, Claude Code, or Bolt at it — the coding agent knows exactly what to build, with testable acceptance criteria for every requirement.

Navigate to Spec Studio in the sidebar. The list page shows all specifications as a table or card view, with columns for title, linked opportunity, status, version, author, and last updated date. Click “Generate New Spec” to begin.

There are two ways to create a spec:

• From an Opportunity (Essentials, Pro, Enterprise) — Select an opportunity and Clarion generates a PRD pre-filled with evidence from linked signals — customer quotes, pain points, and quantitative data.
• Standalone (Free, Essentials, Pro, Enterprise) — Paste existing PRD text or describe a feature — Clarion structures it into the full PRD format. Free users get Clarion AI; Pro users get Clarion AI Pro for higher quality output.

Every PRD is editable with a rich text editor. The editor supports:

• Structured sections — Problem Statement, Goals, User Stories, Requirements, Success Metrics, Risks, Timeline, and more.
• Inline editing with formatting toolbar (bold, italic, lists, headings, links).
• Auto-save every 30 seconds with visual save indicator.
• Version history — revert to any previous version.

A generated PRD is a starting point, not the final word. Use the AI review loop to sharpen it before it goes to engineering:

• Critique — Clarion reviews the PRD like a senior PM and flags weaknesses: vague or untestable requirements, thin evidence, unaddressed risks, and weak success metrics — with a short assessment of each.
• Revise— apply the critique and Clarion rewrites the PRD in place to address it, preserving the spec’s ID and all its links. You can also regenerate the PRD from scratch.

Specs auto-save every 30 seconds. Click “Save Version” to create a named checkpoint with an optional change note. The version history panel shows all versions with timestamps — click any to see a diff against the current spec. You can restore any previous version.

Click “Generate Machine Spec” to translate the PRD into a structured, machine-verifiable specification. Progress is shown in real-time with a phase timeline. If generation is interrupted, it automatically resumes from where it left off.

The machine spec includes:

• Functional requirements — Detailed specifications with acceptance criteria in Given/When/Then format.
• API contracts — Endpoint definitions, request/response schemas, error codes, and authentication requirements.
• Data models — Database schema changes, field types, constraints, and relationships.
• Test scenarios — Comprehensive test cases covering happy paths, edge cases, error handling, and performance criteria.
• Non-functional requirements — Performance targets, security requirements, accessibility standards.

After generating a machine spec, you can generate five agent artifacts optimized for AI coding assistants:

Download individually or as a ZIP bundle. Preview each artifact before downloading.

From the spec editor, use the “Export for Agents” dropdown to generate and download:

Push generated artifacts directly to a GitHub repository. Clarion creates a new branch, commits all artifact files, and opens a pull request. When the PR is merged, Clarion automatically detects the shipped feature and updates the linked opportunity status.

Specs are validated for completeness and consistency. The validator checks for missing acceptance criteria, undefined API contracts, inconsistent data models, and untested edge cases. Validation results are shown as warnings that you can address before generating artifacts.

Clarion uses role-based access control (RBAC) to manage permissions within your organization.

Invite team members from Settings → Team. Each invitation specifies a role. For Pro plans, you can optionally assign a Pro seat to the invited member (giving them access to Pro features). Invitations are delivered via email and expire after 7 days.

You can also create shareable invite links for bulk onboarding. Invite links can be configured with a maximum number of uses and an expiration date.

Organization name, logo, default timezone, and global preferences. Configure the product hierarchy (Product → Areas → Features) that’s used by all AI operations for accurate signal classification.

Invite team members by email. Four role levels:

Connect and manage all third-party integrations from Settings → Integrations. Each integration shows its connection status, last sync time, and signal count.

Navigate to Settings → API Keys to generate keys for programmatic access and MCP server authentication. Each key is shown once on creation (copy immediately), with a name, creation date, last used timestamp, and a revoke button. Keys are scoped to your organization.

Define your product structure as a multi-level hierarchy — from top-level product areas down to individual features: Product → Areas → Features → Sub-features. The hierarchy uses a drag-and-drop tree editor. Each node displays its signal count badge, aggregated sentiment, and auto-mapped signals. This hierarchy feeds into every AI operation — classification maps signals to product areas, themes roll up to the hierarchy, and specs reference specific product areas for context. Reorder by dragging; add children by clicking the “+” button on any node. Clarion automatically maps incoming signals to 1–3 relevant product areas using AI, enabling per-area analytics and filtering.

• Tree management — Add, edit, reorder, and nest product areas up to 5 levels deep.
• Signal mapping — Each signal is automatically assigned to relevant product areas during processing. The AI reads signal content and matches it against your hierarchy.
• Aggregated metrics — View signal counts, sentiment averages, and opportunity counts per product area — with automatic rollup from children to parent nodes.
• Tree visualization — Interactive tree view with expand/collapse, drag-and-drop reordering, and color-coded health indicators.

Under Settings → Product Contextyou’ll find your product structure plus what Clarion knows about how your team builds — the detected tech stack (frameworks, languages, databases) and codebase conventions. This context makes generated PRDs and specs stack-correct even when no coding agent is connected.

• How it’s populated — run npx clarion-cli stack in your repo (uploads a summary, never your source), connect a GitHub / GitLab / Bitbucket repo (read-only, manifest-only, refreshed on push), or have your coding agent report it via MCP.
• Where it’s used — PRD & machine-spec generation, “What to Build?”, and drift checks all read it.

Clarion separates the company (the account that pays you — ARR, segment, plan, domain) from the contacts (the individual people who submit feedback) inside it. A signal attaches to the person who said it; their account is reachable through them, so ARR and segments roll up correctly without double-counting.

• Companies (Settings → Companies) — a searchable, ARR-sortable, segment-filterable list of accounts. Open one to edit segment/ARR/plan once and have every contact inherit it, and to see nested contacts.
• Customers / Contacts — individual contact pages with full signal history and linked opportunities.
• Bulk import — upload a CSV with a company column and contacts are grouped into accounts automatically (by domain or name).
• Auto-populated — integrations, Stripe (ARR), and CSV feed these records; emails are matched case-insensitively so the same person never duplicates.

Navigate to Settings → Segments to create and manage customer segments. Define segments with rule-based criteria:

• ARR range— e.g., Enterprise (>$100K ARR), Mid-market ($10K–$100K), SMB (<$10K).
• Company size — filter by employee count or organization tier.
• Custom attributes — any metadata from your signal sources (plan type, region, industry, etc.).

Each segment shows a live match count — the number of customers that meet the criteria. Segments power the Segment Comparison page, filter “What to Build?” recommendations, scope Knowledge Explorer queries, and shape every AI analysis across the platform.

Navigate to Settings → Brandingto customize your organization’s appearance in outbound communications. Upload your company logo and set a primary brand color. These appear on validation survey pages, digest emails, and shared reports — giving a professional, white-labeled experience to customers and stakeholders.

Navigate to Settings → Workflows to configure multi-stage approval processes (see the Approval Workflows section for full details). The visual drag-and-drop editor lets you define stages, assign roles, configure parallel branches, and set auto-escalation rules.

Navigate to Settings → Notifications to configure how you receive alerts. Two delivery channels: Email and Slack webhook. Toggle notifications for: new high-priority signals, anomaly alerts, spec status changes, approval requests, and team activity. Set a Slack webhook URL to receive notifications in your team channel.

Scheduled summary reports delivered via email or Slack. Digests keep stakeholders informed without requiring them to log in.

Navigate to Settings → Digests to configure automated summary emails. Each team member can set their own schedule and delivery method independently. Choose frequency (Daily, Weekly, or Monthly), set your timezone, and toggle which sections appear: Signal Summary, Highlights, Anomaly Alerts, New Opportunities, Sentiment Trends, and AI Recommendations. Use “Send Test” to preview the email. View digest history to see what was sent.

Configuration

• Frequency — Daily, weekly, or monthly.
• Delivery — Email or Slack.
• Timezone— Reports are generated at the configured time in the user’s timezone.
• Content — AI-generated highlights summarizing new signals, emerging themes, sentiment shifts, and top opportunities since the last digest.

Navigate to Settings → Login Activity to view your login history — timestamps, IP addresses, device info, and success/failure status. Filter by date range and status. Useful for security auditing and detecting unauthorized access attempts.

Navigate to Settings → Profile (or click your avatar in the top-right corner). You can update your display name, email address, and upload a profile avatar. Your avatar appears on signals you create, specs you author, opportunity cards you own, and in the team activity feed.

Change your password from the profile page. If you signed up with Google OAuth, you can optionally add a password for email/password login. You can also link additional OAuth providers to your account for flexible sign-in options.

If you belong to multiple organizations, click the organization name in the sidebar dropdown to switch between them. Each organization has its own data, team, plan, and settings — switching is instant with no logout required.

Clarion meters its generative AI features with AI credits. Every plan includes a monthly credit balance that refreshes at the start of each cycle. Unused credits do not roll over.

What spends credits(generative actions): generating a PRD or machine spec, agent artifacts, the “What to Build?” engine, Knowledge Explorer chat, research runs, persona extraction, drafting documents, and validation reports and emails.

What is always free: the background AI that keeps your workspace organized (signal classification, theme and entity extraction, embeddings, clustering, sentiment, and anomaly detection), plus everything non-generative (browsing, viewing, editing, dashboards, exports, and integrations).

Starter (free) includes 325 credits per month, roughly 5 PRDs plus 1 machine spec. Paid plans include a larger monthly allowance that scales with your plan, and with your seat count on team plans.

Top-ups. Need more before your next refresh? Owners and admins can buy a credit top-up anytime in Settings → AI Credits, added to your balance immediately.

Team distribution. On team plans, admins decide how the monthly pool is used: share one pool across everyone, or allocate a set amount per member, in Settings → AI Credits.

Check your balance, this cycle’s usage, and renewal date in the credit indicator at the top of the app (click it for details), or in Settings → AI Credits. Admins also get a per-member usage breakdown there.

Payments are processed securely. We accept all major credit/debit cards, UPI, and net banking. Billing is monthly or annual (annual saves 20%). Manage your subscription, download invoices, and update payment methods in Settings → Billing.

You can let your plan auto-renew each period, or choose Pay once at checkout for a single month or year with no auto-charge. With Pay once, re-pay before the date shown on Settings → Billing to keep your plan active. Paying early stacks onto your current date, so you never lose remaining days.

If a paid plan lapses (the subscription ends, or a one-time plan is not renewed in time), your workspace moves to the free Startertier. Your data stays intact and you keep the Starter monthly credit allowance. Re-subscribe anytime to restore full access. Your plan’s active-until date is visible to everyone on the team in Settings → Billing.

If you have a coupon code, enter it on the checkout page before completing payment. Coupons can provide a percentage discount or grant free access for a specific plan. Coupon codes are created by platform administrators and may have expiration dates or usage limits.

Enterprise plans include custom seat counts, dedicated support, SLA guarantees, custom integrations, and advanced security features. Contact sales to discuss your requirements.