Marketers Embrace Modular Agentic AI Architectures

In an opinion piece for AdExchanger, Dan Hagen, Global Chief Data & Technology Officer at Havas, argues that marketing should adopt modular, interoperable architectures rather than rely on end-to-end, agency-owned platforms. The article draws a direct analogy to the shipping container revolution credited to Malcom McLean in the 1950s, arguing that standardization, not ship redesign, delivered faster, cheaper, and more secure logistics. The piece frames the contrast as a "black box" approach versus a "black book" modular approach that integrates with clients' existing technology stacks.

Industry-pattern observations: Building agentic AI as composable services typically emphasizes well-defined APIs, standardized data contracts, and lightweight orchestration layers rather than single monolithic agents. This modular pattern reduces coupling between data sources, model components, and activation logic, making it easier for organizations to incrementally adopt capabilities and to replace or upgrade individual pieces without full-platform migrations.

Editorial analysis: Public discourse has oscillated between promises of end-to-end platforms and the pragmatic reality of heterogeneous enterprise stacks. The shipping-container analogy highlights a broader pattern where interoperability and standards often unlock scale and efficiency more reliably than centralized, closed systems.

For practitioners and vendors, useful indicators include adoption of common data schemas for marketing events, increased use of adapter layers for legacy systems, emergence of vendor-neutral orchestration tools, and case studies showing incremental modular rollouts that preserve existing workflows.