How This Briefing Works
This report opens with key findings, then maps the gaps between what Perplexity discloses and what BLACKOUT observed at runtime. From there: what it means for your organization, what to do about it, and the detection data and evidence underneath.
Key Findings
Pre-Consent Activity
Perplexity was observed loading and executing before user consent was obtained on 100% of sites where it was detected.
Claims vs. Observed Behavior
pending
“Unknown”
Requires claims extraction via CDT
What This Means For You
What To Do About It
Role-specific actions based on observed behavior
If You Use Perplexity
- →Demand data processing addendum with explicit query retention and usage terms
- →Require consent framework integration that blocks tracking until user acceptance
- →Implement query data minimization to exclude PII from search logs
- →Configure search analytics to separate product intelligence from user profiling
- →Establish retention limits for query history and behavioral profiles
If You're Evaluating Perplexity
- →Request technical documentation on consent detection and tracking initialization
- →Verify whether query data is used for AI model training or competitive intelligence
- →Test persistence mechanisms to understand profile continuity after cookie deletion
- →Review data flows to third-party AI infrastructure and analytics platforms
- →Assess fingerprinting techniques and cross-session identity resolution logic
Negotiation Leverage
- →Perplexity deploys session recording and consent bypass that captures all user search behavior—demand explicit DPA terms covering query data processing and regulatory liability protection
- →Search query patterns reveal detailed competitive research and product evaluation behavior—negotiate contractual limits on secondary use of query data for vendor intelligence or AI training
- →Persistence mechanisms ensure profile continuity across sessions and cookie deletion—require transparency into storage architecture and user data deletion capabilities
- →Platform creates attribution distortion by over-crediting search touches in conversion paths—establish baseline measurement methodology before deployment
- →Legal tail risk of 100% cannot be fully mitigated through configuration—evaluate whether conversational search value justifies privacy exposure or consider privacy-respecting alternatives
Runtime Detections
BLACKOUT observed this vendor's JavaScript executing in a live browser and classified each hostile behavior using our BTI-C (Behavioral Threat Intelligence — Capability) taxonomy. These are not theoretical risks — each code below was triggered by something we watched this vendor's code actually do.
Evasion infrastructure, auditor bypass
Impact: Perplexity can detect security analysis environments and alter tracking behavior during privacy assessments, masking production data collection.
Keystroke/mouse tracking
Impact: Query velocity, interaction patterns, and answer consumption behavior create persistent user profiles across sessions.
Full session replay
Impact: Full session capture records all search queries, answer interactions, and navigation behavior for profile refinement.
Ignoring CMP signals
Impact: Tracking initialization occurs before consent capture, processing user queries and behavior regardless of privacy preferences.
Device identification
Impact: Device and browser fingerprinting creates persistent identifiers for user recognition across search sessions.
Long-lived identifiers
Impact: Multiple storage mechanisms (cookies, localStorage, indexedDB) ensure profile continuity even after users clear cookies.
IOC Manifest
Indicators of compromise across 4 categories. Use for detection rules, CSP policies, or Pi-hole blocklists.
Ecosystem & Supply Chain
Evidence Artifacts
Artifacts collected during analysis, available with evidence-tier access.
Complete network capture with all requests and responses
516 detection signatures across scripts, domains, cookies, and network endpoints