Settings
Manage your account and platform preferences
Aditya Thakur
Senior Investigator
AdminCritical alerts
Immediate notification for critical-priority alerts
Case updates
When cases you are assigned to are updated
Daily digest
Morning summary of overnight activities
Low priority alerts
Notifications for low-priority items
Two-Factor Authentication
Add an extra layer of security
Session Timeout
Auto-logout after inactivity
What is Agentic AI?
PolitiTrace uses an Agentic AI system — an AI that doesn't just answer questions, but actively works alongside you. Unlike a simple chatbot that waits for your input, the Agentic AI continuously analyzes all entities, alerts, cases, and network connections in the background to proactively surface insights.
5
AI-powered features
Every
page has AI insights
You
stay in control
How do I know what's AI-generated?
Every AI-generated element is marked with a sparkle icon or labeled "AI". AI suggestions are recommendations — they help you prioritize and understand, but you make the final decisions. All analysis runs on platform data only; nothing is sent externally.
Quick Access: AI Copilot
Press Ctrl + J from any page to open the AI Copilot chat. Or click the floating button at the bottom-right.
Feature-by-Feature Guide
AI Copilot
Where: Floating button (bottom-right) or Ctrl+J
A chat assistant available on every page. It understands which page you're on and what case, entity, or alert you're looking at. Ask it questions in plain language like "analyze this case" or "which alerts should I escalate?" and it responds with data-driven analysis.
How does it work technically?
The Copilot reads the current context (page, selected case/entity) and queries the platform's data to generate contextual responses. It adapts its suggested prompts based on the page you're viewing.
AI Threat Briefing
Where: Top of Dashboard page
A daily threat assessment that summarizes everything happening on the platform. Shows the current threat level, highlights the most critical findings, recommends where to focus your time, and predicts emerging risks with confidence scores.
How does it work technically?
The AI aggregates all active alerts (by priority and category), all entity risk scores, and open case data. It calculates a threat level, identifies the dominant alert categories, and uses pattern analysis to generate predictions about emerging risks.
AI Investigation Intel
Where: Right sidebar when viewing a case in Investigations
Replaces the old static intel panel with dynamic, per-case analysis. Shows risk factors computed from the actual linked entities, AI-generated narrative analysis, detected corruption patterns, and prioritized recommendations for next steps.
How does it work technically?
When you select a case, the AI reads all linked entities' risk scores, examines the case tags (e.g., shell-company, procurement), counts evidence items, and generates custom risk factor breakdowns. Each recommendation is tailored to the specific case type.
AI Alert Triage
Where: Alerts page — insight cards above the table + triage column
Each alert gets a colored badge (Escalate / Investigate / Monitor / Auto-dismiss) with a confidence score. Hover over any badge to see the AI's reasoning. The insight cards above show detected alert clusters — groups of related alerts targeting the same entity or following the same pattern.
How does it work technically?
For each alert, the AI combines the alert's own risk score (60% weight) with the linked entity's risk score (40% weight) to compute a combined score. Based on thresholds, it assigns a triage recommendation. Clusters are detected by grouping alerts that share the same entity or category.
AI Network Pattern Detection
Where: Network Graph page — right panel when no node is selected
Scans the entire network graph to find patterns invisible to the naked eye: hub entities with the most connections, tightly connected clusters of high-risk nodes, chains between government officials and companies, and suspicious low-confidence inferred connections.
How does it work technically?
The AI computes degree centrality for every node to find hubs, then identifies subgraphs where high-risk nodes are interconnected. It looks for cross-type edges (official → company) to flag potential conflicts of interest, and flags inferred connections below 40% confidence for verification.
AI Feature Tour
Re-run the onboarding walkthrough that explains all AI features