Rokt mParticle Launches Performance Engine to Automate Audience Building

The Performance Engine functions as an automated layer atop a brand’s existing data architecture. Its central component, the Audience Agent, interprets marketing goals in plain language to scan enterprise data, identifying the specific signals that correlate with conversion. Unlike standard AI assistants that focus on creative copy, this agent performs the analytical heavy lifting of identifying high-value users, leaving the final approval to the marketer.

To bridge the gap between data identification and campaign execution, the platform deploys four distinct Performance Accelerators. Alongside the existing Match Boost and Predictive Audiences tools, the company introduced Audience Expansion and Household Reach. The former identifies lookalike behaviors within a brand’s own dataset, while the latter targets collective household purchasing patterns, such as family plans or shared services. For CKE Restaurants, these tools proved immediate, with the Hardee’s and Carl’s Jr. brands reporting significant lifts in ad platform reach across Google and Meta.

According to Sam Dozor, CTO of Rokt, the system’s primary advantage lies in its real-time feedback loop. Because the platform continuously ingests conversion data, the agent’s recommendations evolve based on campaign performance. This creates a compounding benefit where the technology becomes more precise the longer it interacts with a brand’s unique customer signal.