Attempts to fully unify the media market are stalling as new protocols like MCP emerge. Instead of simplifying everything, MCP is enabling granular, automated data operations across fragmented platforms.
The Model Context Protocol (MCP) is not the silver bullet for unifying the advertising market. Instead, it’s making the complex ecosystem more operable by enabling unprecedented data granularity. As multiple protocols develop in parallel, true universality is off the table. Each player is building their own approach, environment, and logic. Paul Ripart, digital and data deputy director at Prisma Media, warns that this could even lead to new, less visible walled gardens. The idea of running a fully unified cross-media campaign remains unrealistic, given the persistent differences in buying logic, technical constraints, formats, and platform-specific rules.
The real strength of MCP lies in automating micro-tasks at scale. These include categorizing actions, applying enriched taxonomies, generating UTM parameters, structuring campaigns with consistent naming conventions, and ensuring the right data lands in each platform. While these tasks are simple, they’re repetitive and massive in volume-still largely handled by humans today. MCP brings automation and reliability, generating structured data one micro-task at a time.
Currently, the media landscape is rich but tangled. All the data exists, but it’s intertwined and hard to interpret. To unlock value, organizations must untangle this web-understanding each campaign, data point, and interaction at a granular level. MCP is valuable not for unification, but for activating these data threads individually and at the right scale. Once this groundwork is done, it becomes possible to reweave a coherent, actionable network.
Another key shift is moving beyond API-centric thinking. While MCP excels at interconnecting the media ecosystem, agentic AI goes further. Tools like Claude Cowork and Claude for Chrome combine browsing agents that operate directly in Chrome with API-manipulating agents, transforming how media operations are executed. Today’s workflows often depend on complex, limited integrations, each siloed in its own SaaS or browser tab. Agentic AI-from MCP to virtual actions-enables direct, universal interaction with any interface, shifting the focus from technical integration to practical usage. This fundamentally changes campaign operations.
However, this automation only works if the underlying data is clean. MCP doesn’t fix errors-it executes instructions. Without strong data hygiene, automation will simply replicate and amplify mistakes. Strict discipline is required, down to concrete rules like account presets, prohibitions, and buying frameworks. This rigor is essential for meaningful automation.
MCP’s value is in targeted interoperability, not in creating a single system. It allows agents to coordinate categorized actions across diverse tools and platforms. For example, it can connect media planning with project management tools like Asana or Monday.com, auto-generate media plans from briefs, assist with briefing, or distribute data across platforms. This lets users work in their preferred interface-creating a media plan from Claude within a media operations platform, or leveraging Claude’s agentic power from within that platform. The key is respecting established workflows.
The media market’s complexity isn’t going away, but it can be better managed. MCP and broader agentic AI, including agentic browsing, help streamline operations, organize actions, and optimize each purchase at the task level, making data actionable. The solution isn’t more connection, but better organization of what already exists-avoiding the automation of chaos.
