IntouchCX

Resources

Case Study: AI-Enabled Fraud Checks for Optimal Efficiency

Effective fraud prevention is crucial for safeguarding a company’s financial stability. However, distinguishing legitimate transactions from suspicious ones requires judgment calls. Manual fraud checks are slow, error-prone, and struggle with high transaction volumes, often missing subtle fraud patterns or flagging too many false positives. That’s why there’s a need for a digital solution that can address these challenges efficiently and accurately.

The Challenge

We collaborated with a leading grocery delivery service to optimize their fraud detection processes and drive significant improvements in operational efficiency prior to customer interaction. Previously, customer service agents were manually conducting fraud checks, meticulously reviewing each item against a comprehensive fraud markers scorecard. This often led to subjective interpretations and time-consuming inefficiencies. As an effect, numerous cases were mistakenly escalated, burdening a specialized team with unnecessary investigations, wasting resources, and creating additional workload.

What We Found

After thorough analysis of the manual fraud-check process conducted by the Escalation Review Team (ERT), we identified several challenges. Many of the fraud criteria lacked clear definitions, leaving agents to rely on their subjective judgment for ambiguous markers like “multiple high-value orders over consecutive days.” 

We discovered the average handle time (AHT) for the ERT fraud-check process stood at 11.5 minutes, but shadowing uncovered that numerous cases took 20-30+ minutes to complete. To address these issues, we proposed implementing Laivly’s AI-driven automation solution Sidd Pro, confident that it would enhance consistency in the fraud escalation process and significantly reduce the time spent by the ERT team on each case.

Our Solution

In collaboration with the grocery delivery service and ERT leads, we implemented a comprehensive solution. Clear definitions were established for fraud markers, such as “three orders greater than $500 over three consecutive days,” and a group of new ERT agents were trained using Laivly’s Sidd Pro solution, integrating it into the fraud-check process from the outset. Now every fraud check is conducted consistently and with improved efficiency. 

The platform initiates a precheck, swiftly identifying obvious fraud indicators and streamlining the agent’s workflow by eliminating unnecessary tasks. Sidd Pro further enhances the process by assisting the agent in conducting research, partially or fully automating the fraud check, or offering guidance and navigation for manual verification as needed.

Our Results

Implementing AI-driven automation for the ERT program delivered outstanding results. The pilot group quickly achieved the baseline of 11.5 minutes average handle time (AHT) within three weeks, and further reduced it to 6.5 minutes within two months.

Key improvements include:

  • 43% reduction in AHT within two months.
  • 71% reduction in AHT by the end of the pilot.
  • Positive agent feedback, praising time savings and task simplification.


The specialized escalation team now handles cases more efficiently, and performance continues to improve, showcasing the solution’s lasting impact.

Safeguard Your Brand With AI-Powered Fraud Protection

Laivly’s AI-driven platform revolutionized our partner’s fraud-check process. What was once a subjective, time-consuming task is now streamlined and efficient, reducing unnecessary escalations and easing the burden on agents. With clear fraud markers, checks are consistent, saving time and protecting profitability. Agents now work smarter, and fraud prevention is stronger.

Learn more about our fraud detection services today.