Fraud Detection with AI SEON: Signals Rise from 900 to Over 1,100

SEON has significantly expanded its proprietary data infrastructure, increasing the number of available signals from over 900 to more than 1,100 directly sourced data points. This is a technical leap designed to address a threat that is growing faster than traditional defenses: AI-generated fake identities. The fraud detection with AI SEON now focuses on a much broader base of signals, capable of cross-referencing addresses, devices, session behavior, and phone history to unmask profiles constructed in minutes.

Summary

• Key Points
• SEON expands proprietary signals for fraud detection
  • New data on addresses, devices, phone, and session behavior
• The challenges of AI-generated fraud and industry alerts
  • Speed of AI tools and regulatory warnings
• How the new detection signals work
  • Digital history and phone data
  • Shared infrastructure between addresses and devices
  • Real-time behavior monitoring
• SEON's vision on AI fraud prevention
• FAQ
  • What new data has SEON added to its anti-fraud platform?
  • How do AI-generated fake identities complicate fraud detection?
  • What does the Financial Action Task Force say about AI-generated deepfakes?
  • How does SEON's signal intelligence help unmask fraudulent networks?

Key Points

• SEON has extended the proprietary signals of the platform from 900+ to over 1,100 data points.
• The expansion covers address intelligence, session behavior, phone and operator data, digital footprint, and device signals.
• According to the Financial Action Task Force (FATF), anyone with a smartphone can generate credible deepfakes in the time it takes to create a social profile.
• ACAMS reports that 75% of financial crime prevention professionals consider the misuse of GenAI the main emerging risk for the third consecutive year.
• The new signals flow into SEON's AI Command Center, usable in rules, alerts, reviews, and investigations of fraudulent networks.

SEON Expands Proprietary Signals for Fraud Detection

The expansion announced by SEON concerns five distinct areas: address intelligence, session behavior, phone and operator data, deeper digital footprint, and device signals. The stated goal is to provide risk teams with independent evidence on every level of a customer's identity, without adding extra steps to the user journey.

New Data on Addresses, Devices, Phone, and Session Behavior

Each individual signal, taken alone, may appear legitimate: an email passes validation, a device seems clean, an address appears correct. It is the intersection of these elements that reveals the anomaly. Therefore, the expansion of SEON's signal intelligence works on three parallel dimensions: the identity history over time, the shared infrastructure between linked accounts, and live behavior during the session. All this context is then made available for decisions made by automated rules, human analysts, and AI agents, within the AI Command Center of the platform, where the new signals integrate into rules, alerts, customer reviews, and network investigations. Each signal is also linkable to the AI tools chosen by the investigator via SEON's Model Context Protocol (MCP) server, allowing human analysts and automated agents to work on the same informational basis, from detection to action.

The Challenges of AI-Generated Fraud and Industry Alerts

Building a fake identity used to require time, patience, and almost artisanal work: gathering documents, filling out forms, managing accounts one by one. Today, all it takes is an internet connection and a prompt. GenAI tools can produce credible profiles and coherent histories on devices in just a few minutes, radically changing the scale at which fraudsters operate.

Speed of AI Tools and Regulatory Alerts

According to the Financial Action Task Force (FATF), anyone with a smartphone can generate a convincing deepfake in the time it takes to open a profile on a social network. Confirming the extent of the problem is an industry statistic: ACAMS reports that 75% of financial crime prevention professionals rank the misuse of GenAI as the most urgent emerging risk for the third consecutive year. The fraudster continues to choose identities and targets, but it is the AI agent that manages production at industrial speed: this is where AI fraud prevention becomes a race against time for those who need to defend themselves.

How New Detection Signals Work

Digital History and Phone Data

Checks on digital footprints verify where an email or phone number has appeared over time across different services. Coverage has expanded to AI developer platforms, job portals, real estate sites, and dating apps. To complete the picture, Phone Intelligence adds SIM-swap history and number portability. Together, these signals allow teams to understand whether an identity and phone number truly existed before that specific moment or were assembled on the fly.

Shared Infrastructure Among Addresses and Devices

Address Intelligence transforms chaotic address strings into a risk signal, verifying and standardizing addresses in over 240 countries and assigning consistent identifiers to the same building or housing unit. This allows for the discovery of when seemingly unrelated accounts revolve around the same apartment number or variations in formatting of the same location. On the device front, expanded signals identify AI agent activity, compromised iOS devices, discrepancies between Android eSIMs and network country, as well as VPN masking that alters the visible IP. Together, these elements help identify accounts linked to recycled infrastructures and assess how much trust can be placed in the device environment.

Real-Time Behavior Monitoring

Session Monitoring tracks customer behavior from onboarding to login, account recovery, checkout, and payment. Teams can detect automation, remote access, off-screen activity, and active calls while the session is ongoing, intervening before suspicious activity turns into an account takeover.

SEON's Vision on AI Fraud Prevention

Tamas Kadar, CEO and co-founder of SEON, commented on the logic behind the expansion: "AI has made producing a credible identity economical. What fraudsters cannot easily do at scale is build a coherent story for each account without reusing the same infrastructure. That’s where our signal base makes a difference. The more dimensions an anti-fraud team can check simultaneously, the harder it becomes to hide an identity that doesn’t hold the accounts together."

The statement summarizes well the strategic sense of the operation: while creating a fake digital face costs less and less, maintaining its credibility on dozens of different levels remains costly and complicated. It is precisely on this imbalance that the entire expansion of signal intelligence signed by SEON focuses, and more generally the evolution of the sector towards AI fraud detection capable of reading patterns invisible to the human eye but evident when data is cross-referenced on a large scale.

FAQ

What new data has SEON added to its anti-fraud platform?

SEON has expanded its signals by including address intelligence, session behavior, phone and operator data, digital footprint, and device signals.

How do AI-generated fake identities complicate fraud detection?

GenAI tools can create credible profiles and consistent device histories in just a few minutes, making it much harder to identify fake identities using traditional methods.

What does the Financial Action Task Force say about AI-generated deepfakes?

The FATF states that anyone with a smartphone can quickly generate a convincing deepfake in the time it takes to create a profile on a social network.

How does SEON's signal intelligence help uncover fraudulent networks?

It exposes inconsistencies in identity and reused infrastructures among linked accounts, verifying multiple levels of data such as addresses, devices, and phone histories.

Content created with the assistance of artificial intelligence and human editorial review.