David San Filippo is Senior Vice President of Digital Content and Experience at Altudo and a CMS Critic contributor.

Walmart recently shared how it’s rethinking its enterprise architecture to embrace a future powered by Agentic AI: autonomous systems that can reason, plan, and act on behalf of humans to achieve business goals. What stands out about Walmart’s approach isn’t just the ambition, but the discipline. Walmart isn’t just sprinkling AI into workflows; it’s rearchitecting how its organization functions.

This shift is already visible in how Walmart deploys AI in the business. For its merchants, Walmart has developed a GenAI-powered assistant that surfaces insights, analyzes sales trends, and recommends actions like adjusting pricing or merchandising strategies, helping teams move from reactive reporting to proactive decision-making.

On the customer side, Walmart has introduced AI shopping assistants that guide buyers through complex decisions, from finding the perfect holiday gift to building personalized shopping lists, pulling from real-time inventory and product data. In both cases, these aren’t static chatbots, they are agents that understand goals, plan next steps, and execute tasks in a way that shortens the gap between intent and action.

Their vision for an agentic future is grounded in three core pillars: data readiness, enterprise orchestration, and trust and security. Behind the scenes, this means building an ecosystem where AI agents can securely access internal systems, reason over structured and unstructured data, and collaborate across domains, all while staying accountable to human operators. It’s an architectural shift, not just a feature drop.

And Walmart isn’t alone. Across industries, organizations are racing to understand what agentic AI means for their digital operating models. That urgency is now rippling through the Digital Experience Platform (DXP) space, where major players are repositioning themselves to ride the next wave of AI innovation. From product content and personalization to campaign orchestration, the promise of intelligent agents is reshaping how DXPs present their value, not just as publishing tools, but as platforms capable of participating in enterprise-wide decision-making and execution.

The question is no longer if DXPs will adopt agentic capabilities, but how they’ll fit into a much larger ecosystem of enterprise AI.

Digital Experience Platforms have long promised to help marketers move faster, but the past year has seen a dramatic shift in how they intend to deliver on that promise. It’s no longer just about drag-and-drop editing, reusable components, or streamlined content governance. The focus is now squarely on AI agents that can understand goals, make decisions, and act on behalf of marketers, with minimal input.

Two of the most prominent players in the DXP space, Sitecore and Optimizely, have made aggressive moves toward positioning themselves as leaders in this agentic future.

Sitecore: From Personalization to Intelligent Orchestration

Sitecore has rebranded its vision around becoming an agentic DXP, framing its platform not just as a system of delivery, but as a system of intelligence. At the center of this push is Sitecore Stream, its brand-aware foundational AI layer designed to power personalized content generation, dynamic experiences, and task automation, all governed by brand guidelines and organizational context.

The recent partnership with Gradial, a pioneer in secure agent frameworks, signals Sitecore’s intent to go beyond embedded generative tools. With Gradial’s infrastructure, Sitecore aims to support agents that can navigate complex tasks across tools, not just create content, but also understand and act within the workflows of a marketing organization. On Sitecore’s AI microsite, this vision is framed around unifying composability, personalization, and intelligence under a single architectural banner.

Optimizely: Scaling Creativity and Campaign Execution

Optimizely, meanwhile, has evolved its Opal AI framework into something more robust and extensible. Originally focused on creative assistance – suggesting headlines, generating copy, or organizing ideas – Opal is now positioned as the foundation for autonomous agents that can streamline experimentation, manage content variants, and even help optimize campaigns across channels.

In a recent blog post on AI Agents and the Future of Work, Optimizely articulates a future where marketing teams focus on strategy and oversight, while agents take over repetitive or logic-based tasks like audience targeting, personalization rule setting, and A/B test configuration. This shift isn't just about productivity; it’s about making experimentation and personalization scalable.

The Bigger Trend

Both platforms are moving beyond static AI integrations and toward dynamic, multi-step task execution powered by intelligent agents. The goal is to embed AI deeper into day-to-day marketing operations – not as a tool you open, but as a system that proactively collaborates with you.

But while this movement is exciting, it also introduces a deeper challenge: these agentic capabilities, as powerful as they are, were never designed to operate in isolation.

To unlock real enterprise value, organizations must look beyond what’s embedded in the platform and consider how these DXPs will fit into a much larger, AI-driven ecosystem.

AI Is Bigger Than the Platform

It’s easy to get excited about the agentic capabilities now appearing in platforms like Sitecore and Optimizely, and for good reason. These tools promise to reduce marketing effort, accelerate content creation, and automate personalization in ways that once required entire teams. But if you’re serious about operationalizing AI across your organization, it’s important to zoom out.

Think back to Walmart’s approach. Their AI strategy wasn’t built around features inside individual platforms. It was architected to give agents secure access to enterprise data, enable interaction across systems, and maintain visibility and oversight across every action taken. Their focus wasn’t just marketing efficiency – it was enterprise orchestration.

That’s the scale modern organizations need to be thinking about.

While agentic features inside DXPs are valuable, they’re only a small part of what’s required to unlock real impact. Most enterprise needs go far beyond what any single platform can handle:

Secure Data Access: Agents need to tap into CRMs, ERPs, CDPs, and DAMs, all governed by strict access controls and compliance requirements. A DXP doesn't own this data, nor does it define the access policies.
Multi-System Orchestration: Business logic often lives across tools. Launching a campaign, for example, might involve asset creation in a DAM, segmentation from a CDP, activation in a marketing automation platform, and publishing through a CMS. Enterprise AI must coordinate this, not just execute isolated tasks.
Auditability and Control: Walmart emphasized trust, and trust requires transparency. Enterprise AI must be traceable, reversible, and subject to human oversight. Agent actions must be logged, reviewed, and governed according to role-based permissions and escalation paths.
Strategic AI Alignment: Organizations aren’t just building agents inside martech tools, they’re consolidating their efforts around AI platforms like Azure AI Foundry, Amazon Bedrock, or Salesforce Einstein 1 Studio. These ecosystems offer the observability, guardrails, and deployment frameworks needed to support AI at scale.

In this context, DXPs are no longer the center of gravity, they’re a critical surface area within a broader, composable AI architecture. That changes the conversation. Instead of asking, “What can my DXP automate?”, the better question becomes: “How can my DXP participate in an enterprise agent ecosystem?”

That’s the lens Walmart is using – and increasingly, it's the lens others must adopt as well.

DXPs Must Learn to Play in the Agentic Ecosystem

This shift in perspective, from platform-level automation to enterprise-wide orchestration, has significant implications for Digital Experience Platforms. As enterprises invest in standardized agent frameworks and centralized AI platforms, DXPs can no longer operate as isolated silos of content and campaign logic. Instead, they must evolve to become interoperable, agent-aware services within a broader ecosystem.

To do that, DXPs will need to expose their internal logic, content structures, and execution capabilities to agents operating outside their walls. And they’ll need to do it in a way that supports intelligent reasoning, secure execution, and seamless collaboration across systems.

Two key concepts are beginning to define this next phase: Model Context Protocol (MCP) and Agent-to-Agent (A2A) communication.

Model Context Protocol (MCP): Making the DXP Legible to Agents

MCP provides a structured, machine-readable representation of a platform’s content model and supported actions. It allows external agents, those built on platforms like Azure AI Foundry, Amazon Bedrock, or Google Vertex, to understand what a DXP knows, how its content is structured, and what kinds of operations it can perform.

But supporting agents isn’t just about exposing APIs, it’s about rethinking what those APIs offer. If an agent needs to plan and execute work, such as preparing a campaign launch or creating multilingual variants of a product landing page, relying solely on atomic endpoints that fetch fragments of data forces the agent to assemble context manually – increasing latency, complexity, and cost.

To enable more capable, context-aware agents, DXP platforms must provide higher-order interfaces aligned to common AI planning tasks, not just “fetch article by ID,” but “retrieve all draft content scheduled for Q4 product releases, grouped by region and persona.” These types of purpose-built MCP endpoints make it easier for agents to reason about work and take meaningful action.

Over time, DXPs will need to ship a robust set of out-of-the-box MCP capabilities, but that won’t be enough. Enterprises will need to extend MCP definitions to reflect their own unique processes, data models, and business rules. A consumer goods brand may define an MCP endpoint that surfaces launch-critical assets by SKU and channel. A financial services firm may create one that surfaces regulatory disclosures by region and legal owner.

In all cases, the goal is the same: make the DXP intelligible and actionable for external agents, so they can plan, execute, and learn more effectively.

Agent-to-Agent Communication (A2A): Delegating Intent Across Domains

Where MCP is about making a system legible to external agents – helping them understand what the DXP knows and how to act on it – Agent-to-Agent (A2A) communication is about enabling collaboration across systems through intent-based delegation.

Rather than building brittle, cross-system integrations that require agents to know how to operate every platform, A2A lets one agent delegate a goal to another, trusting it to understand its own environment, constraints, and execution model.

This is especially important in complex enterprise environments, where no single agent can feasibly own every tool, process, and policy. Instead, domain-specific agents, like a DXP agent, become specialists that know how to reason about and act within their system. Other agents simply tell them what outcome is needed, and trust them to figure out how.

Take a merchandising scenario: a supply chain optimization agent identifies products that are overstocked and need to be promoted. Rather than manually triggering changes across commerce, CMS, and email systems, it simply instructs the DXP agent to "feature these products as promotional add-ons or related items" across digital touchpoints.

The DXP agent – understanding the platform’s component library, personalization framework, and content model via MCP – takes over, and it:

Updates relevant product detail pages.
Inserts cross-sell banners or promotional tiles.
Schedules campaign content across regions.
Adapts the display logic to match brand and UX standards.

Crucially, the supply chain agent doesn’t need to know any of that. It doesn’t care about the layout, the delivery channel, or the personalization configuration. It expresses a high-level business need, and the DXP agent takes responsibility for carrying it out in a way that is context-aware, compliant, and efficient.

This model is fundamentally different from traditional integrations or automation scripts. A2A allows for flexible, loosely coupled collaboration that mirrors how humans work, passing intent, not instructions, while preserving control and domain expertise at the system level.

To support this, DXPs will need to provide not only APIs and context (via MCP), but also agent interfaces that can accept, interpret, and act on high-level directives – all while enforcing governance, logging actions, and escalating issues when needed.

The Case for Native Agentic Features

As promising as MCP and A2A are, the reality is that enterprise-wide agent ecosystems will take time to fully materialize. Standards are still evolving, governance frameworks are still being defined, and for many organizations, the supporting infrastructure simply isn’t in place yet. Not every task requires inter-system coordination, and not every agent needs to be globally orchestrated.

There’s real value “today” in using what DXPs already offer: native, self-contained agentic capabilities that are deeply embedded in content, personalization, and marketing workflows. These aren’t just lightweight tools that help you write faster. They’re structured, goal-oriented systems that understand the platform’s internal logic and act with meaningful autonomy.

We’re already seeing this in action. Sitecore’s partnership with Gradial provides a glimpse into how native agents can transform marketing execution inside the DXP itself. For example, agents can generate a tailored landing page for a specific customer segment, complete with copy, visuals, and metadata aligned to brand standards. That same agent can then trigger downstream activity, like preparing the content for promotion via email, all within the DXP’s domain. There’s no need for orchestration across tools or departments; the agent understands what needs to be done and how to get it done using the DXP’s native capabilities.

These types of agents are also beginning to assist with personalization logic, suggesting new rules or content variants based on behavioral signals. They help automate quality assurance, flagging accessibility issues, broken links, or brand compliance violations before content goes live. In some platforms, agents can even configure A/B tests, monitor early results, and offer recommendations. all from within the campaign interface.

And in larger operational tasks like content migration, agents are proving especially useful. They can analyze legacy content, propose new structures based on target models, and assist with tagging, metadata generation, and transformation. The agent doesn’t need to ask where to start, it understands the publishing model, governance rules, and approval workflow.

What makes these capabilities powerful is that they’re truly agentic, even if they’re scoped. They plan, reason, and act with autonomy, but within a controlled environment where security, governance, and outcomes are clearly defined. And unlike the cross-system vision of A2A, these tools are available now, or will be very soon, in most modern DXPs.

For organizations just beginning their AI journey, or those looking for tangible productivity gains without major architectural shifts, these native agents represent a smart, immediate next step. They reduce friction, improve quality, and lay the foundation for a more intelligent and responsive marketing operation.

Composable DXPs Are Ready for an Agentic Future — Are You?

The rise of Agentic AI marks a fundamental shift in how digital experiences are created, delivered, and optimized. What once required teams of specialists can increasingly be orchestrated by intelligent agents – not just automating tasks but making decisions, adapting in real time, and collaborating across systems to drive outcomes.

Digital Experience Platforms are beginning to evolve to meet this moment. Sitecore, Optimizely, and others are embedding agentic capabilities that make marketing operations faster, smarter, and more scalable. And emerging models like MCP and A2A hint at a future where DXPs don’t just automate tasks but participate in enterprise-wide agent networks – exposing their capabilities and collaborating through intent-driven, cross-system orchestration.

But that future won’t arrive all at once. For most organizations, the path forward will be incremental and composable. Native agentic features offer an immediate opportunity to unlock value without waiting for standards to mature or infrastructure to be overhauled. At the same time, it’s critical to choose platforms and architectures that are future-aware, designed to expose context, support secure agent interfaces, and play well within an ecosystem that’s only going to grow more intelligent.

Composable DXPs are uniquely suited to this dual mandate. Their modular nature makes it easier to integrate with external agents, extend native capabilities, and evolve over time. They allow you to start with embedded agents that accelerate your teams today and grow into more ambitious, enterprise-scale agent strategies tomorrow.

The question isn’t whether DXPs will support agents. They already do. The question is whether your organization is thinking big enough to align those capabilities with a broader AI strategy. Because Agentic AI isn’t just another martech feature. It’s the beginning of a new operating model and the platforms that embrace that shift early will be the ones that lead it.

Upcoming Events

Storyblok JoyConf 2025

October 14-15, 2025 – Amsterdam, Netherlands

Join us at JoyConf 2025, where the future of content meets the people shaping it. Built for developers, marketers, and digital innovators, this free conference goes way beyond product updates or yet another strategy playbook. It’s where you’ll get inspired, get real and get connected – to ideas, to the community, and to what’s possible when you build joyfully. Whether you're building lightning-fast frontends or launching cross-channel campaigns, this is where content becomes connection and collaboration becomes second nature. Come for the talks. Stay for the community. Leave with joy. Save your spot.

Sitecore Symposium 2025

November 3-5, 2025 – Orlando, Florida

At Sitecore Symposium, you’ll learn how to meet your customers where they are – across channels, moments, and touchpoints that didn’t exist a year ago. You’ll hear how the boldest brands in the world are transforming content into outcomes and creating experiences that are fast, relevant, and human. Join 1,500 marketers, digital leaders, and technologists in Orlando to see what’s possible when personalization scales, AI accelerates creativity, and trust becomes your competitive edge. This is the Third Wave of the Internet. And this is your moment to lead in it. Get your tickets today.

CMS Kickoff 2026

January 13-14, 2026 – St. Petersburg, Florida

Meet industry leaders at our fourth annual CMS Kickoff – the industry's premier global CMS event. Similar to a traditional kickoff, we reflect on recent trends and share stories from the frontlines. Additionally, we will delve into the current happenings and shed light on the future. Prepare for an unparalleled in-person CMS conference experience that will equip you to move things forward. This is an exclusive event – space is limited, so secure your tickets today.

News

Agentic AI Is Coming. Is Your DXP Ready?