Please ensure Javascript is enabled for purposes of website accessibility
CUSTOMER SUCCESS

How a leading U.S. financial services company transformed fraud investigations with Isaac

A leading financial services organization with more than $50 billion in annual revenue used Isaac, an AI Agent for AML and fraud investigations, to reduce manual investigation work, improve consistency, and give investigators more time for deeper analysis.

Results:

  • Positive ROI in the first year
  • 3x IRR in three years
  • Production deployment within months
  • More consistent investigation decisioning and reporting
  • Token costs kept in check through purpose-built AI
  • Investigators freed to focus on complex analysis

From 1.5-day investigations to faster, more consistent decisions

The organization’s first-line AML and fraud investigations team was responsible for reviewing thousands of cases each month, including potential account takeover activity that could lead to suspicious activity reports.

Each investigation took approximately 1.5 days. Investigators spent much of that time gathering information from multiple systems and manually transferring it into Word templates. The process was slow, inconsistent, and difficult to scale.

An internal custom-build effort had also proven challenging. Its hard-coded approach became too complex to maintain as business processes changed, and a 12-month development timeline was longer than the team could accept.

Isaac brings governed automation to fraud investigations

The organization selected Isaac, an AI Agent for AML Transaction Monitoring and fraud investigations from WorkFusion, a UiPath company.

Isaac integrates with the team’s detection software and related systems to collect, organize, and present the information investigators need to determine whether a case warrants a suspicious activity report.

The team’s investigators remain responsible for the decision. Isaac handles the most time-consuming work around that decision, helping investigators move through the process faster while preserving the controls required in a regulated environment.

Every step Isaac takes is traceable and auditable. Investigators approve each disposition through a maker-checker model, supporting expectations from regulators including the OCC and FinCEN.

Isaac’s modular design also allows individual components to be updated as systems, processes, and stakeholders change—without requiring the organization to rebuild the entire solution.

More time for the investigations that matter most

Isaac was in production within months and delivered positive ROI in its first year.

By eliminating the most time-consuming parts of each investigation, Isaac significantly reduced handling time and variability. Investigators now spend less time collecting and documenting information and more time applying their expertise to complex cases, including ownership structures spanning trusts, holding companies, and other entities.

The organization also designed Isaac to apply AI where it adds the most value, keeping token costs in check at both design time and run time.

The result is a more consistent, adaptable investigation process—and a team better prepared for evolving suspicious activity report requirements.

Expanding the AI-powered financial crime operating model

Following its success with Isaac, the organization added Edward for Enhanced Due Diligence and is evaluating additional AI Agents to address rising PEP alerts and other screening false positives.

Its standard for every new AI initiative is clear: deliver a 3x internal rate of return within three years.

Ready to transform your investigation process?

See how AI Agents can help your investigators reduce manual work, improve consistency, and focus on higher-value analysis.

“”

Ultimately, a person still makes the decision. Regulators would be concerned otherwise. But with Isaac, the process is governed by automation 95% of the way through.

 Director, AI and Automation