By Rob Crisp, Head of Solutions EMEA

‍

Key Takeaways

  • MCP creates a shared AI connection layer across platforms like Adobe Experience Manager, Marketo, and Figma, dramatically reducing integration friction and speeding up the content supply chain.  
  • AI agents are absorbing repetitive workflow coordination, shifting human roles toward creative judgment, governance, and strategic oversight.  
  • Governance becomes embedded directly into workflows through automated brand rules, rights management, approvals, and audit trails.  
  • Clean metadata and structured DAM systems become critical because AI agents depend on organized, queryable content data to operate effectively.  
  • The largest operational opportunity may be content reuse. AI agents can rediscover, adapt, and redeploy underused assets at scale, improving efficiency, personalization, and brand consistency.

What the MCP Layer Actually Changes

MCP (Model Context Protocol) is an open standard that gives AI agents a consistent way to connect to external tools and platforms. MCP provides a common interface that any compliant agent can use to read/write data and take actions within connected platforms. In the context of the content supply chain, this means an agent can interact with systems like Adobe Experience Manager, Figma, Marketo, and Pardot through the same protocol layer. It creates a shared connective tissue across platforms, eliminating the need for custom point-to-point code for every individual integration.

Traditionally, the content supply chain is a sequence of human handoffs, from brief to creative to review to approval to publish. MCP is most clearly valuable at the boundaries of that chain, as spoke connections radiating outward from the work management hub. Connecting activation platforms like Marketo and Pardot, pulling assets from the DAM, and retrieving performance data from AEP are natural MCP use cases. Live integrations are now emerging across each layer of the content supply chain.

The work management platform itself sits at the centre as the orchestration hub. Here, the more likely pattern is native AI capability within the platform rather than MCP. Task prioritisation, routing, and status management are functions that platforms like Workfront are building their own intelligence around, operating within their existing data model. MCP connects these platforms to the systems around them; it does not replace internal orchestration logic. The supply chain compresses because agents handle those boundary crossings; humans intervene at the points that require judgement.

How Roles Change

The shift to agentic orchestration is significant across every function that touches content operations. The changes are not uniform: some roles are disrupted and new hybrid capabilities emerge at the boundary of marketing and engineering.

Content Producer

The producer role shifts from execution toward quality control. Rather than building assets from scratch they are reviewing and approving what agents assemble. The skill set moves toward prompt craft, brand judgement, and output evaluation.

Project and Traffic Manager

This is the most directly disrupted role. The coordination work inside Workfront, assigning tasks, chasing approvals, tracking asset status, is exactly what CX Enterprise Coworker is designed to absorb. The role pivots toward exception handling and maintaining the business rules the agents operate within.

Creative Director

The creative director becomes more important, not less. Adobe Brand Intelligence needs to be taught, calibrated, and corrected. Someone has to own the evolving definition of what on-brand looks like, and that is a creative leadership function that cannot be delegated to an agent.

Marketing Operations

The marketing ops role expands significantly. Defining agent skills, maintaining the governance layer, owning the business rules that determine how agents behave: this is marketing ops work, but at a much higher level of technical sophistication than today. This role becomes a critical bridge between business intent and agentic execution.

Developer and Solutions Architect

The developer or solutions architect becomes embedded in the marketing function rather than sitting separately in IT. MCP endpoint configuration, agent orchestration design, and integration maintenance require engineering capability living close to the business problem.

How Governance Improves

This is one of the strongest parts of the CX Enterprise story and an area where BPX has a clear point of view.

Today governance in the content supply chain is largely manual and inconsistent. Brand compliance depends on reviewers catching things. Legal sign-off is a bottleneck because it happens late. Rights management on assets is poorly tracked. Channel adaptation often bypasses the original approval process entirely.

With an MCP-connected agentic layer, governance can be codified rather than checked:

  • Brand rules become parameters the agent operates within rather than guidelines a human reads and interprets.
  • Rights and usage data attached to assets in the DAM can be queried by the agent before it selects or repurposes content.
  • Approval workflows can trigger automatically based on content type, channel, or audience segment rather than relying on a human to remember the rules.
  • Audit trails improve substantially. Because agent actions are logged and attributable, you get a traceable record of what was used, adapted, approved, and published.

The audit trail value is particularly strong for regulated industries and for post-campaign analysis, and positions governance as a commercial advantage rather than a compliance burden.

How Data Improves

Several things happen simultaneously when the agentic layer is introduced over a content supply chain.

Asset Metadata

Agents require structured, queryable data to function. If your DAM has inconsistent tagging today, the agent will underperform. This creates a business case for metadata remediation that was previously difficult to justify. The agentic layer surfaces the cost of poor data hygiene in a way that procurement and finance can understand.

Performance Feedback Loop

Because the agent layer sits across both content production and engagement measurement in AEP, it can connect content attributes to performance outcomes at a level of granularity that humans rarely manage consistently. Over time this creates a genuine learning loop: content that performs well informs the parameters the agent uses when assembling future content.

Customer Data Enrichment

Rather than a marketer manually segmenting an audience and briefing a creative team, the agent can pull real-time segment data from AEP and either adapt existing assets or flag where new content is needed for a segment that is currently underserved. Personalisation becomes systematic rather than periodic.

Content Reuse

This is arguably the biggest operational win and the one that is easiest to quantify for a business case.

Most organisations have enormous libraries of content that is chronically underused because discovery is poor and adaptation is manual. An agent with MCP access to the DAM can find, retrieve, and adapt existing assets far more efficiently than a human searching through folders. Combined with rights data and performance history, it can prioritise reuse of assets that have proven effectiveness rather than defaulting to new production.

The reuse case is also strong from a brand consistency perspective. If agents are drawing from a governed, well-tagged asset library rather than humans creating net new content under time pressure, the output is more consistent by design.

BPX can help companies frame a focused engagement around content reuse uplift. The work can begin with auditing the existing asset library and improving metadata and tagging standards. From there, teams can configure the DAM for agentic consumption and establish a baseline reuse rate to measure future improvement against. The result is a commercially attractive initiative tied to a measurable operational outcome.

‍