Delivering High-quality Data amid Rising AI-driven Fraud
Business Challenge
A large software company engaged TRC to conduct a multi-wave Market Map Study that targeted industry professionals. Initially, the study flowed flawlessly. But when we added an oversample group, we noted their fielding phase responses were completed at lightning speed. Something was off.
When our research and project management teams took a deeper look, they found:
- Typing speed anomalies: Some “respondents” had keystroke rates of 50–120 characters per half second, which is physically impossible for a real human being.
- Inconsistencies across all panel partners—and when compared to past data waves. Having a dedicated, consistent research team familiar with the historical data made this detection much easier.
- 76% of open-ended responses were flagged as suspect. This was an immediate warning sign our data was likely compromised. Our project managers manually review these entries every day to confirm the system’s judgment.
TRC’s Solution
We deployed our proven Identify-Investigate-Resolve framework; real-world fraud detection strategies that deliver accurate, clean data. In this case, our algorithm tool, ResponseID, was paired with a dedicated, expert research team to ensure integrity and accuracy. Our project manager manually reviewed open-ended responses to confirm the flagged responses, and another team counted the number of characters entered every half-second.
Successful Results Implementation
Once AI-driven fraud was detected in this study, we made the decision to remove 100% of the data from the new panel partner. We prefer to err on the side of caution—especially when our clients rely on us to inform major business decisions.
