RPA agentic AI partnership with Newton Research accelerates media buying models

The transition from correlation-based measurement to causal modeling allows agencies to prove the direct impact of media spend on sales outcomes. By automating the data-heavy setup of marketing-mix models, RPA can pivot from reactive reporting to predictive simulations in near real-time. This shift reflects a broader industry move toward agentic systems that handle complex optimization tasks, potentially reducing the reliance on third-party measurement firms. As agencies adopt these tools, the focus will move toward evidence-based budget reallocation across fragmented digital channels. Watch for whether this speed in modeling leads to more frequent mid-campaign budget shifts between social and programmatic platforms.

Newton Research has been building momentum in the agency analytics space by positioning causal inference as a replacement for correlation-based marketing-mix models. The firm raised $12 million in Series A funding in early 2025 to expand its causal AI platform, with investors citing the company's ability to automate the data preparation and model specification steps that traditionally make MMM projects slow and expensive. That funding round signals growing investor confidence that agencies will internalize measurement capabilities previously outsourced to specialized consultancies like Dentsu's Merkle or WPP's GroupM analytics division.

RPA's parent company, IPG, has been pushing its agencies toward proprietary AI tooling across the network. In March 2026, IPG announced its proprietary AI platform, IPG.AI, would be deployed across all its agencies including RPA, UM, and Mediabrands, covering creative production, media planning, and audience analytics. The RPA-Newton partnership fits within that broader corporate mandate to embed AI directly into workflow layers rather than treating it as a bolt-on capability. Meanwhile, competing holding companies are pursuing similar strategies: Publicis Groupe acquired Lotame in January 2026 to strengthen its first-party data and measurement infrastructure, and WPP has integrated its own causal modeling tools through its partnership with DataRobot.

The technical approach Newton Research brings to RPA centers on causal discovery algorithms that can identify which media channels drive incremental sales rather than merely correlating with them. Newton Research published benchmark data in Q2 2026 showing its platform reduced MMM setup time by 78% compared to traditional econometric workflows, with model accuracy on holdout tests within 3 percentage points of manually specified models. For RPA specifically, the integration targets its Unlimited Analytics product, which the agency uses for cross-channel budget optimization. The speed gains matter because mid-campaign reallocation decisions in programmatic and social channels often require updated model outputs within 48 to 72 hours, a timeline that traditional MMM firms cannot meet.