Casap AI Agents Raise $56M to Fight Fraud and Disputes

Casap, an AI agent platform for financial services, has raised 56 million dollars in a Series B funding round to expand its work on fraud and payment disputes into wider bank operations, according to Finovate. The round was led by Base10 Partners, with strategic participation from Experian and Synchrony, and additional support from existing investors including Emergence Capital and Lightspeed Venture Partners.

The funding brings Casap AI agents to a new stage of growth. The California-based company, founded in 2023, now reports total funding of 90 million dollars, just over a year after a 25 million dollar Series A round, according to Finovate. The company plans to use the new capital to deepen its agentic AI platform and extend it beyond disputes into collections, claims, and other complex exceptions.

Casap AI agents are designed to work inside the systems financial institutions already use. According to the company, its specialized agents investigate cases, apply institution-specific policies, gather evidence, make decisions, and take action, while keeping auditable records and escalating cases that require human judgment. The pitch is outcome-focused rather than efficiency-only: the agents are meant to change results, not simply speed up existing workflows.

Casap AI Agents Take On First-Party Fraud

The core problem Casap addresses is first-party fraud, in which cardholders themselves dispute legitimate charges. According to reporting on the round, cardholders now account for nearly a third of fraudulent transactions, a trend fuelled by AI tools that make disputing charges easier and by social media groups sharing tips for obtaining refunds improperly.

Consumer protection rules complicate the picture for banks. According to Fortune reporting cited in coverage of the round, those rules protect consumers well against third-party card fraud but place a heavy burden on banks and card companies when they need to reject dishonest claims from their own cardholders. Investigations can be so costly that institutions sometimes concede low-value claims rather than fight them.

Casap says its approach changes that calculation. Its AI was trained on large models but uses a purpose-built harness designed especially for dispute resolution, according to Dealroom. The system weighs criteria such as the physical location of a card and spending patterns to flag likely dishonest claims, while keeping a human in the review loop. The company does not disclose its full set of fraud indicators.

The performance figures Casap reports are striking, though they are company-supplied claims. According to Finovate, customers have reduced first-party fraud losses by up to 51 percent and achieved a 96 percent chargeback win rate, compared with a 60 percent industry average. Resolution time has fallen from as long as 90 days under traditional methods to eight days on average, and 73 percent of dispute cases are resolved through agentic automation while retaining human oversight. Human processing time per case has fallen from about 45 minutes to one or two minutes, according to TipRanks.

From Disputes to Collections, Claims, and Complaints

Casap was founded by former Robinhood executive Shanthi Shanmugam and former Chime executive Saisi Peters, according to Finovate. The company launched with a focus on fraud and payment disputes, then expanded into other operational areas. With the Series B, it plans to push further into adjacent parts of bank and card company back offices, including customer complaints and collections services.

According to CityBiz, Casap now works with more than 100 financial institutions, including banks, credit unions, fintech companies, and payment processors. The company also reports a tenfold revenue increase since its Series A, exceeding its original growth target by more than two times, according to TipRanks. Strategic investor Experian is working with Casap on next-generation signals for distinguishing genuine fraud from first-party misuse, combining Experian identity and risk data with Casap transaction-level insights.

The broader context matters for anyone watching the agent economy. Casap AI agents join a wave of specialized agents moving from answering questions to completing regulated, auditable work inside enterprises. Similar momentum is visible elsewhere in agent infrastructure, from AI agents entering the analytical lab to AI agents automating corporate benefits work. In each case, the test is the same: can agents operate under real policies, produce records a regulator or auditor can inspect, and know when to hand off to a person?

That test is especially sharp in financial services, where a wrong dispute decision can harm a genuine customer or let fraud through. Casap says its agents escalate judgment-sensitive matters to human staff, and its reported 73 percent automation rate leaves more than a quarter of cases outside full automation. Whether that balance holds as the platform moves into collections and complaints, where customer harm risks are higher, will be the key question for its next phase.

For now, the funding signals strong investor confidence in agentic operations for finance. An oversubscribed Series B, strategic backing from Experian and Synchrony, and a customer base above 100 institutions give Casap room to prove that Casap AI agents can scale beyond disputes without losing the auditability and oversight that made the first use case work. Sources: Finovate, Dealroom, TipRanks, and CityBiz.