Navee has launched of Navee GPT, a technology that enables automatic moderation of high risk counterfeit listings, 24/7, 365 days a year.
If counterfeit production were a country, it would be the sixth largest emitter of CO2 globally, with devastating consequences for the environment, society, and economy. In light of this urgent challenge, Navee is taking the lead with the launch of Navee GPT, a large-scale language model (LLM) that takes the fight against counterfeiting even further.
This complements Navee’s earlier innovation, the CVAN (Collaborative Visual Authentication Network). Result of collaboration between brands and platforms, CVAN blocks counterfeits, at upload, across the platforms using semi-exact image matching and insights sharing, allowing them to proactively fight counterfeits. Navee GPT further reinforces this technology by using AI to analyze and tackle high risk listings that may have slipped past CVAN. This new multimodal LLM aims to help brands and platforms double their listing moderation and cut average processing time by up to 10 times.
With the integration of Navee GPT, the initial moderation capacity could multiply fivefold or even tenfold, enabling the processing of up to 15,000 counterfeit cases per month per customer to be processed. This allows Navee to empower their partners and customers to triple the efficiency of their analyst teams. The system quickly identifies and evaluates suspicious content, and when combined with CVAN, it can automatically reduce counterfeiting by up to 90 per cent without human intervention. This technological breakthrough provides prompt and precise feedback while helping brands trim operational costs, making it an indispensable tool for organisations seeking to streamline their content moderation processes and improve overall performance.
Navee GPT uses its powerful algorithm to assess the risk level of each listing, pre-moderating high-risk content not addressed by CVAN. Trained using a structured and moderated guide, reinforced by Navee’s moderation history, its accuracy rivals that of human moderators. The solution operates continuously 24/7, facilitating large-scale, high-quality moderation.
Share