Guests: Prasanna Venkatesh | Kartik Kumar
Company: Egen
Show Name: An Eye on AI
Topic: Agentic AI

Fraud detection remains one of the most complex challenges faced by public agencies today. Investigators handle millions of claims every year, and identifying which ones are fraudulent requires time, precision, and context. Despite having access to data and technology, the process is still largely manual—until now.

In this conversation, Swapnil Bhartiya talks with Prasanna Venkatesh, Sr. Engineering Manager, and Kartik Kumar, Sr. Data Engineer at Egen, about how artificial intelligence is helping investigators uncover fraud faster and more efficiently. The discussion dives deep into how AI and analytics are changing the investigative process, offering new ways to prioritize suspicious claims and detect emerging patterns across large data sets.

Venkatesh explains that traditional fraud checks—such as verifying claimant identities, cross-referencing employers, and checking wage records—are mostly human-driven and time-consuming. Egen’s approach introduces AI-driven scoring and document intelligence, helping investigators focus on the most suspicious claims first. This prioritization lets legitimate claims move quickly through the system while keeping potentially fraudulent ones under review.

Kumar adds that AI’s role isn’t to replace investigators but to empower them. These tools act as an “adjudicator assistant,” surfacing insights that humans might miss when processing data manually. Dashboards and pattern-recognition models allow agencies to see correlations across claims—for example, a sudden surge in submissions from the same IP address or similar contact details—which could signal coordinated fraud attempts.

Egen’s Document Intelligence tool is a centerpiece of this transformation. It automates document validation by cross-checking information across IDs, wage statements, and utility bills, detecting forged or altered records, and verifying them against official databases like DMV and Social Security. The system also flags if the same document is used in multiple claims, even with minor alterations such as different photos.

Both guests emphasize the importance of collaboration between AI and investigators. Fraud detection, they say, is evolving into an “arms race.” While fraudsters use generative AI to create fake documents and synthetic identities, agencies are leveraging AI to analyze and counter these tactics at scale. The result is faster detection, smarter decisions, and fewer false positives.

What excites the Egen team most is how the technology shifts agencies from being reactive to proactive. With generative and agentic AI, investigators can anticipate new fraud patterns before they spread, turning investigation into prevention. By working closely with agencies, Egen ensures that its solutions are adaptable, customized, and continuously refined based on real-world feedback.

Here is the edited Q&A of the interview:

Swapnil Bhartiya: Fraud detection and dealing with fraud is one of the most challenging jobs. Public agencies process millions of claims every year, and a growing number turn out to be fraudulent. What does the process actually look like today, and why is it still so challenging?

Prasanna Venkatesh: Fraudsters are constantly innovating, using new technologies to launch fresh schemes and exploit weaknesses. State agency investigators must stay one step ahead by verifying every detail—from claimant identities to employer and wage data—often across multiple databases. Many of these checks are manual, making the process slow and difficult to scale.

Kartik Kumar: Exactly. Most fraud investigations today involve reviewing one claim at a time, even though fraud often happens at scale across multiple claims. Investigators don’t have the tools to easily connect suspicious cases, so identifying patterns is time-consuming.

Swapnil Bhartiya: How is AI helping in these day-to-day investigations?

Prasanna Venkatesh: AI helps identify patterns in historical data and assign fraud scores to claims. This lets investigators focus on the most suspicious cases first. We also provide dashboards that reveal emerging fraud patterns and offer a 360-degree view of claims and claimants. Our Document Intelligence tool detects whether submitted documents, like driver’s licenses or social security cards, are authentic or altered.

Kartik Kumar: AI speeds up analysis by doing the manual data checks that investigators used to do one by one. It’s not replacing them; it’s helping them keep pace with fraudsters who are also using AI to scale attacks.

Swapnil Bhartiya: Fraudsters are evolving quickly. How do these tools help agencies spot new types of fraud before they grow?

Kartik Kumar: We use AI to recognize suspicious trends across thousands of claims—like a sudden spike in applications from the same IP or reused email patterns. These systems can flag such anomalies instantly. Investigators can then group related cases and act faster.

Prasanna Venkatesh: And by allowing investigators to drill down into data, link related claims, and view relationships visually, our tools make it easier to decide which cases are truly fraudulent.

Swapnil Bhartiya: Fraud is becoming an arms race. How do investigators compete as fraudsters adopt AI too?

Kartik Kumar: It’s a real arms race. Fraudsters use AI to generate fake identities and documents, but investigators use AI to prioritize workloads and detect subtle anomalies faster. The key is that humans remain in control—the AI just provides speed and scale.

Prasanna Venkatesh: Exactly. The human-in-the-loop approach ensures that investigators make final decisions, but AI gives them the data, insights, and automation to do it more efficiently.

Swapnil Bhartiya: Tell us more about Egen’s Document Intelligence.

Kartik Kumar: Document review is one of the hardest and most time-consuming tasks. Our tool automatically cross-checks details, detects forged elements, and validates formats. It produces a summary report so investigators can act faster.

Prasanna Venkatesh: It also compares documents across multiple claims and databases, spotting duplicates or altered IDs, which is nearly impossible manually.

Swapnil Bhartiya: What excites you most about the future of fraud detection?

Prasanna Venkatesh: The collaboration between human expertise and AI. Generative and agentic AI will help automate repetitive steps while enhancing decision-making.

Kartik Kumar: The shift from reactive to proactive detection. AI is helping agencies predict and prevent fraud before it happens.

Swapnil Bhartiya: Thank you both for sharing these insights.