Pi.

Pi is a decisioning engine developed by Paytm One97, Inc. that aims to personalize and manage risk for individual users. According to the vendor, Pi allows businesses of various sizes to confidently take on more risk and grow. It is used in professions and industries such as financial services, e-commerce, online marketplaces, fintech, and payment service providers.

Key Features

Personalized: Pi enables businesses to personalize user experiences based on their risk profiles and activity. The vendor claims that businesses can dynamically manage and personalize users by creating personalized risk scores and serving them unique experiences.

ML-Driven: According to the vendor, Pi utilizes machine learning models to identify patterns and outliers, continuously monitoring user activity, adjusting risk scores, recommending rule sets, and shifting thresholds in real-time. The machine learning studio allows businesses to test rules, anticipate changes, and proactively adjust defenses based on real-time data.

Lightning Fast: Pi provides quick risk assessments with decisioning typically taking less than 50ms, the vendor claims. Its scalable engine ingests data from the entire ecosystem, allowing businesses to create complex rule sets without worrying about response time.

Ready: Pi offers ready-to-use templates that businesses can customize and tune to their specific needs. The vendor states that leveraging a decade of firsthand learnings, businesses can potentially lower operation costs, reduce fraud loss, and transform their organization.

Data Ingestion: According to the vendor, Pi securely aggregates proprietary and third-party data into a comprehensive view of the business' ecosystem, making decisions 2x faster than the industry average.

Dynamic Risk Scoring: Pi calculates hyper-specific risk tiers for each user, evaluating and shifting risk scores dynamically throughout their lifecycle. The vendor claims that personalized user experiences are served based on configurable settings and scorecards.

Smart Simulations: According to the vendor, Pi allows users to test rules in a simulated environment, providing insights into anticipated changes and results before implementation.

Anomaly Detection: Pi offers built-in anomaly detection to automatically identify patterns and outliers, including unknown unknowns. The vendor states that real-time alerts are sent when activity shifts are detected.

Machine Learning Studio: According to the vendor, Pi's unsupervised machine learning auto-adjusts user risk tiering, rule sets, and thresholds in real-time, avoiding traditional hard-coded limits.

Case Management: Pi provides a collaborative case management system for fraud operations teams, with detailed reason codes, reports, and a full audit trail available in one place, according to the vendor.