Kevel, an API-first ad serving company, has announced the launch of Kai, a comprehensive AI and machine learning feature set aimed at optimizing performance, relevancy, and revenue for retail media networks. Available as part of the Retail Media Cloud™, Kai is designed to maximize advertiser budgets through advanced ad serving capabilities.
The development of Kai has been led by Kevel's AI/ML research group, which includes industry veterans such as CTO Tim Ewald, Sr. Director of Research and W3C member Paul DeGrandis, Principal Data Scientist Richard Carter, PhD, and Retail Media Cloud™ GM and Velocidi founder Paulo Cunha. Leveraging decades of experience, the team has created a powerful suite of AI tools to enhance ad serving and audience segmentation.
Kai introduces two new features, Forecast and Custom Relevancy, alongside existing AI Audience and DecisionAPI products. Forecast utilizes machine learning simulations to predict inventory and campaign performance, offering insights into both current and future ad campaigns. Unlike traditional forecasting tools that rely solely on historical data, Forecast considers contextual and user audience targeting and pacing parameters, making it a first-of-its-kind tool for retail media.
“Forecast is a breakthrough for retail media. Traditional tools only look at past data, but Kevel Forecast uses machine learning to project future campaign performance by considering all relevant factors. This ensures advertisers have a clear view of their future performance, and retailers can maximize their inventory yield,” explained Paulo Cunha, Retail Media Cloud GM at Kevel.
Custom Relevancy allows retailers to input their own AI/ML algorithms into the Kevel Ad Server for bespoke targeting tailored to each network's specific needs. This 'bring your own model' feature enables retailers to use their advanced models to enhance ad relevancy in a secure manner.
“Retailers understand their customers better than anyone but often struggle to apply their sophisticated AI-driven optimizations to ad serving. Custom Relevancy changes that by allowing retailers to integrate their own machine learning models into our ad decision process, dynamically adjusting relevancy on a per-user basis,” commented Tim Ewald, CTO at Kevel.
Kai also includes features like ad decisioning and pacing, which rely on historical data, events, user behavior, context, relevancy scoring, and predictive trends to boost ad performance.
“The exciting part about Kai is that it demonstrates how machine learning can add value for our customers. We've developed these systems from original research using proprietary data sets and our extensive experience in ad serving,” stated Richard Carter, Principal Data Scientist. “We've collaborated closely with retail customers to identify where the most value lies—in decisioning, relevancy, and segmentation. Kai is just the beginning of many more innovative features we plan to introduce.”
