The DXP Catalyst Update - Sept 4, 2025

INTRO
Welcome to This Week’s DXP Catalyst Update

In this week’s DXP Catalyst Update, we turn our focus to agentic AI and its governance implications. The rise of autonomous and collaborative agents is reshaping how digital teams create, test, and personalize experiences. The promise of speed and scale is clear, but the risks of drift, inconsistency, and loss of control are equally real. For organizations managing portfolios of independent sites, the challenge becomes even greater. This edition explores how enterprises can put the right guardrails in place to unlock value while maintaining trust, alignment, and flexibility.

LEADERSHIP GUIDANCE

Agentic AI as a Governance Challenge

Agentic AI represents a major shift in how digital work gets done. Instead of humans directing every action, organizations can now assign agents to perform tasks independently and even collaborate with other agents. Content creation, experimentation setup, or personalization decisions can all be handled by agents that learn and adapt as they work. This level of automation introduces tremendous potential for speed and efficiency.

The challenge is not simply technical capability, but oversight. Enterprises need to ensure that agents work within shared standards, align with brand guidelines, and respect compliance rules. Without careful governance, organizations risk inconsistency, duplication of effort, and errors that damage trust with customers. The more complex the environment, the harder it becomes to keep agents aligned with enterprise objectives.

The Portfolio of Brands Challenge

Some organizations operate as portfolios of brands, each with its own independent websites and digital properties. Marketing and IT leaders may see clear value in deploying agents across these sites. One agent might optimize performance for a single property while another runs targeted experiments elsewhere. In more advanced environments, teams may configure clusters of agents to specialize in specific capabilities such as experimentation, with individual agents responsible for ideation, test setup, traffic allocation, and results analysis.

Over time, however, local teams may begin managing these agents in their own way, with little reference to shared enterprise standards. This creates a governance gap. Each brand may see short-term gains, but without coordination the portfolio gradually fragments. Design systems drift apart, data practices become inconsistent, and customer experiences diverge. Leaders might only recognize the problem after months of misalignment, at which point realignment becomes an expensive and disruptive effort. Governance must therefore operate at the enterprise level, ensuring that even locally deployed agents remain within a common framework.

Multi-Agent Orchestration Across Sites

One of the most powerful uses of agentic AI is orchestration, where multiple agents collaborate to complete larger tasks. An experimentation workflow is a good example. One agent generates content variants, another manages traffic allocation, and a third interprets results. Together they can deliver valuable insights far faster than a human-only team.

When orchestration happens within a single site, oversight is relatively straightforward. The complexity grows when multiple sites each run their own orchestration clusters. Without shared governance, the logic and methods used on one site may diverge completely from another. This creates conflicting approaches to experimentation, reporting, or even personalization.

To address this, enterprises must define rules for orchestration across sites. Some orchestration tasks can be left to local teams, while others should be guided or standardized by central leadership. For example, a central team may define how experiments are logged and how results are reported, while local teams handle which hypotheses to test. This balance preserves agility while preventing enterprise-wide fragmentation.

Balancing Centralized and Local Control

The heart of the governance challenge lies in balancing control. Centralized governance provides the consistency, compliance, and oversight required for enterprise-wide alignment. Local autonomy provides the agility needed to respond quickly to market opportunities and customer insights. Too much centralization creates bottlenecks, while too much autonomy results in fragmentation and chaos.

Rethinking Governance Frameworks for Agents

Traditional governance frameworks were designed around human workflows, approvals, and documentation. Agentic AI requires something different. Organizations must create new guardrails that account for autonomous decision-making, ongoing learning, and multi-agent collaboration.

Key elements of this new governance include escalation paths for agents that encounter uncertainty, ensuring decisions that exceed their authority are routed to humans. Ethical guidelines for data usage must also be explicit, since agents may combine data in ways humans did not anticipate. Finally, enterprises need active monitoring through dashboards that track what agents are doing across sites. Without visibility, governance cannot function effectively.

Governance in this context becomes a living framework. It is not about issuing one-time rules but about designing processes that adapt as agents evolve and as business priorities shift. This type of governance transforms oversight into an enabler of value rather than a constraint.

The Risks of Waiting

Some leaders may assume governance concerns can wait until agentic AI is more widely deployed. However, delay only increases risk. Without effective governance, enterprises may face inconsistent brand experiences, compliance failures, or missed opportunities. Each agent that operates outside a shared framework adds complexity that becomes harder to untangle later.

With governance in place, organizations gain confidence to scale. Experimentation can be conducted across multiple properties without fear of duplication or drift. Personalization can be deployed responsibly, ensuring both relevance and compliance. Content workflows can move faster, supported by agents that operate within clear rules. Governance does not slow down progress; it creates the foundation that allows agentic AI to be scaled effectively across distributed environments.

Final Thoughts

Agentic AI introduces opportunities for speed, scale, and innovation across the digital experience landscape. Yet it also brings significant governance challenges, particularly for enterprises managing portfolios of independent sites. Multi-agent orchestration multiplies the complexity, making oversight even more critical.

The organizations that succeed will be those that balance central standards with local autonomy, treating governance as an active enabler of trust and consistency. They will monitor agents across sites, define clear protocols, and ensure that every deployment aligns with enterprise objectives.

Composable stacks will continue to evolve, and agentic AI will play an increasing role in how they deliver value. By addressing governance directly, IT and marketing leaders can unlock the promise of agents without sacrificing control, trust, or long-term flexibility.