Product Intelligence
Where AI Suggestions check individual nodes for quality gaps, Product Intelligence analyses your entire branch for higher-order risk: compliance exposure, security holes, UX gaps, and weak analytics coverage. It runs whenever you seal a commit, maps findings to real frameworks, and rolls everything up into a single Product Health Score.
Like the proactive Spec Quality agents, Product Intelligence is event-driven and never consumes AI credits — it runs in the background for free.
The Intelligence inbox
Product Intelligence findings appear in their own Intelligence view, separate from the per-node AI Suggestions inbox. Each finding shows:
- A framework badge (e.g.
SOC2,OWASP,GDPR) and, where relevant, the specific control or article it maps to - A severity and confidence score
- A plain-language explanation of the gap and why it matters
- AI Fix — opens AI Chat pre-filled with the finding’s context, ready to address it
What it checks
Product Intelligence bundles several analysis areas. Each runs across the whole product model on the active branch.
Compliance & security
| Area | Example findings |
|---|---|
| SOC2 | Personas are defined but no access-control Rule governs them (maps to CC6.1 / CC6.6) |
| OWASP | A sensitive Flow has no rate-limiting Rule; a change opens a privilege-escalation or injection-class gap |
| PCI-DSS | Payment-handling Features lack rules for cardholder-data protection |
| HIPAA | Features touch health data (PHI) without the required safeguards |
| Privacy (GDPR / LGPD) | No consent capture for PII; no data-subject access or deletion Flow |
Experience & completeness
| Area | Example findings |
|---|---|
| UX | No login, signup, or password-reset Flow; a destructive action with no confirmation step; inconsistent terminology across the model |
| Completeness | Features with no onboarding path; Personas with no way to reach the product |
Analytics
Analytics checks use your live metric data where it’s connected:
- A Metric is tracking below (or above) its target
- A Metric has live data but no target set, or a target but stale data
- A Feature or Flow has no Metric at all
- Low overall analytics coverage; a signup Flow with no funnel Metric; no retention Metric
Feature ideas
Product Intelligence also surfaces proactive feature suggestions — gaps in your product that a comparable product would typically fill.
The Product Health Score
Every open finding — from both Product Intelligence and the Spec Quality agents — rolls up into a single 0–100 Product Health Score. It’s a defensible, at-a-glance read on whether your product model is coherent, complete, and low-risk. Watch it climb as your team closes findings.
Acting on findings
The workflow mirrors AI Suggestions:
- AI Fix to open a pre-filled chat and resolve the gap
- Snooze to defer a valid finding
- Reject / suppress to dismiss one that doesn’t apply, optionally suppressing that check going forward
Tips
- Seal commits regularly. Product Intelligence runs on commit seal, so frequent checkpoints mean earlier signal.
- Treat framework findings as a pre-audit. Closing
SOC2andOWASPfindings before a formal review turns audit prep into routine maintenance. - Connect your analytics. The Analytics checks are far more useful once Live Metrics are bound to your Metric nodes.