A Jev that can ask Opus for help gives Opus accuracy at 4x lower cost (hightouch.com)

A recent study has unveiled a hybrid model combining Jev and Opus, which boosts identity resolution accuracy while cutting costs significantly—over four times. Both Jev and Luna, when used independently, delivered satisfactory results with accuracies ranging from 88.8% to 97.3%. However, when Jev’s outputs were coupled with Opus for further validation, the hybrid model achieved frontier accuracy while remaining cost-efficient, yielding less than 10 cents per 1,000 identities processed. The significance of this development lies in its ability to automate and enhance identity resolution (IDR), a crucial process for companies aiming to accurately map customer data to real identities. The challenge of IDR is underscored by potential data noise, such as typos and misreads, which often lead to over-merging—where multiple identities are linked incorrectly to a single individual. By leveraging Jev's well-calibrated confidence levels, the hybrid model can more effectively evaluate identity claims, allowing for rapid processing of large datasets, which significantly streamlines workflows in customer analytics and outreach initiatives.