Kevel Launches Kai to Boost Performance Optimisation, Relevance & Revenue for Retail Media Networks

Kevel, the API-first ad-serving company, is announcing its new branded AI feature set: Kai (Kevel Artificial Intelligence), a suite of AI and machine learning technologies that enable performance optimisation and drive relevancy, profitability, and revenue. Kai is available as part of the Retail Media Cloud^TM, the ultimate SaaS platform for building retail media networks with ad-serving that maximises share of advertiser budgets.

The new tools were developed and spearheaded by Kevel’s AI/ML research group, chaired by CTO Tim Ewald, Sr. director of research and W3C member Paul DeGrandis, principal data scientist Richard Carter, PhD and Retail Media Cloud^TM GM and Velocidi founder Paulo Cunha. The group has decades of combined experience in AI, which has led them to develop this powerhouse suite of AI features to power ad serving and audience segmentation for a premium retail media experience.

With Kai, Kevel introduces two new features, Forecast and Custom Relevancy, alongside its existing AI Audience and DecisionAPI products. Kevel Forecast predicts inventory and campaign performance for existing and future campaigns using machine learning simulations to generate insights on both current and future ad flights.

“Forecast is a first of its kind for retail media. Traditional forecasting tools look simply at historical data to predict future campaign performance, whereas Kevel Forecast uses machine learning algorithms to project future campaign performance when considering all contextual and user audience targeting and pacing parameters in conjunction with other running or future ads. This way, advertisers always know exactly what their future performance looks like and retailers can maximise their inventory yield,” Paulo Cunha, Retail Media Cloud GM at Kevel explains.

Kevel’s Custom Relevancy allows for retailers to input their own AI/ML algorithms into Kevel Ad Server for custom targeting geared towards the individual performance of each network. Functioning as a unique ‘BYOM’ (bring your own model), Custom Relevancy helps retailers utilise their own advanced models to determine relevance as part of their ad stack in a safe and secure way. 

“Retailers know their customers better than anyone else, but struggle to influence their ad serving with the exceptional AI-driven optimisation they use for promoting a customised user experience,” commented Tim Ewald, CTO at Kevel. “That all changes with Custom Relevancy, which allows customers to plug their own ML models into our ad decision process to dynamically adjust relevancy and improve ad serving a per-user basis."

Kai encompasses not just new features like Forecast and Custom Relevancy, but existing features like ad decisioning and pacing. Kevel’s approach to pacing, delivery, and decisioning leans into historical data, events, previous behaviour, context of the experience, ads viewed, and relevance scoring, plus trends and predictions to drive ad performance. 

“What excites me about Kai is that it's a set of features that showcases how machine learning can be used to deliver more value to our customers. We’ve developed these systems from original research using proprietary data sets, harnessing our many years of experience in ad serving,” stated Richard Carter, principal data scientist. “We’ve been working closely with retail customers to prove where the most value sits and it’s in decisioning, relevancy, and segmentation. KAI is just the start of many more innovative, unique features in our pipeline.”