Artificial Intelligence is becoming a core layer in commerce media as retailers and advertisers use it to tackle operational complexity, fragmented data, and inconsistent standards. The strongest current momentum is around moving from static audience segments to real-time signals, with models used to capture in-the-moment shopper intent, improve recommendation systems, and support more relevant bidding and personalisation. Committee members describe this as a practical shift toward smarter execution for advertisers, better monetisation for retailers, and more useful customer experiences.
Several use cases are already showing traction across campaign operations. These include omni-channel media planning, workflow orchestration through agentic systems and conversational interfaces, and the conversion of customer data into predictive audience building and synthetic audiences. One contributor said early adoption is delivering business value in 3 key areas: omni-channel media planning, workflow orchestration enabled by agentic workflows and conversational interfaces, and transforming customer data into tangible applications such as predictive audience building and synthetic audiences. Another highlighted tools such as algorithm cascades, gating, and business rules that let human operators steer model outputs for real-time bidding and recommendation engines.
Attention is also turning to workflow efficiency and interoperability. The goal is to reduce the number of platforms and processes involved in a typical campaign lifecycle while standardising data inputs and outputs across teams, channels, and retailers. Multi-agentic workflows powered by large language models and API connectors are presented as a route to lower cognitive workload and remove manual setup tasks from AdOps teams. MCP is described as an open standard that securely connects generative Artificial Intelligence models to local data sources, helping replace brittle integrations with faster system-to-system communication and more conversational campaign planning interfaces.
Looking ahead, the next 12 months are expected to intensify adoption of multi-agentic solutions, natural language tools, generative creative systems, and “Talk to your Data” platforms. One contributor said, “At Empathy Lab we are seeing growing demand for our Synthetic Audience solutions – creating “customer” panels for testing of new products, promotions and creative concepts at scale, with Mars reporting over 75% accuracy of results, when compared with scores from human panels.” Another expects a continued shift toward Bayesian measurement, arguing that conventional statistical methods are often insufficient for retail’s noisy and uneven data environment.
Despite the optimism, governance is positioned as the decisive requirement for sustainable growth. Predictive systems need guardrails to optimise for the right incrementality metrics, while generative systems need strict controls around brand safety, tone, and visual identity. Clean product data, usable control interfaces, and software readiness are described as prerequisites. User trust remains another major constraint, with scepticism inside companies slowing adoption until teams feel empowered and in control of how they use Artificial Intelligence tools. A measured rollout, user research, and feedback loops are seen as essential to making automation effective without overwhelming the people and brands relying on it.
