Dan Drapeau is Managing Director at DXP Catalyst Consulting and a CMS Critic contributor.

Evaluating a DXP today often means evaluating its agentic offering alongside everything else. Vendors are racing to lead with it, with Contentstack now positioning itself as an “Agentic Experience Platform” and others following suit with varying degrees of conviction. For some forward-thinking organizations, the agentic offering has become a key evaluation criterion, sitting alongside CMS capabilities, personalization, experimentation, and the rest of the digital technology stack under review.

What gets surfaced less often during these evaluations is what "sufficient" actually means in practice. A DXP's native AI is generally not designed to power AI initiatives outside the platform's scope, so the evaluation question is more specific than enterprise-wide orchestration. It is really about whether the native capabilities are enough for what the DXP is being asked to do, and where the gaps need to be filled by something else.

Those gaps tend to fall into a few categories:

Coverage: The native capabilities to meet requirements within the DXP ecosystem on their own.
Extensibility: Assessing whether they can be supplemented with purpose-built tools where they fall short, or extended through custom agents.
Orchestration: Determining if the agentic layer can coordinate multiple agents working together within and across the products that make up the DXP, which gets more complex in multi-vendor scenarios.
Coexistence: Evaluating how the DXP's agentic layer should work alongside an external agentic framework if the organization already has one in place.

Most evaluations do not surface those questions early enough, which leads to platform decisions made on the assumption that the vendor's agentic capabilities are sufficient without a clear picture of what they are actually being asked to support.

What Native DXP Agentic Capabilities Are Actually Designed to Do

Modern DXPs now ship AI horizontally across their core capabilities, with the scope having expanded significantly over the past year. Opal, Optimizely's agentic layer, is a useful reference point because it operates across the entire content lifecycle, working as a connected layer across the products in the composable suite rather than as a feature inside any single one. Other vendors are taking similar approaches with their own naming and implementation details.

These capabilities continue to improve, and for teams working inside the platform, they reduce friction in content production, surface variant ideas, and accelerate work that previously required separate tooling or significant manual effort. The native AI layer, as it exists today, is designed to make the DXP more productive across the workflows it was built to support.

The starting point for any evaluation is understanding what the native offering is actually built to address before asking whether it is sufficient. It is broad horizontally across the platform but limited to it, which shapes the questions that follow.

Does the Native Offering Cover Your Requirements Within the DXP Ecosystem?

The most common question that surfaces during evaluations is whether the native AI layer can meet the organization's actual requirements across the marketing and content lifecycle. Some capabilities are reasonably well-covered out of the box, including experiment ideation, GEO (Generative Engine Optimization), and asset metadata generation. Others are less clear, particularly around content and creative production at enterprise scale, where enterprise content governance is still often more mature in purpose-built AI platforms than in native DXP agentic layers.

This is where purpose-built tools like Writer and Jasper enter the conversation. Writer has evolved beyond content generation into a more complete enterprise AI platform, with its own agent builder, custom agent development, and orchestration capabilities, though it operates as a closed enterprise platform rather than as an open framework. Some DXP vendors have moved to support integrations with these tools directly, with Uniform offering a connector to Writer that allows content generation results to surface within the authoring interface and hook into the platform's broader AI tooling.

In some organizations, the gaps between the native DXP offering and actual requirements are content-centric. In others, they relate to operational workflows, customer journey orchestration, knowledge retrieval, or cross-system automation, and the right supplement may not be a content tool at all. The evaluation question is whether the capabilities under consideration map to clearly defined organizational requirements, and that determination should come from a defined set of requirements rather than from what shows up in a demo.

Can the Native Agentic Framework Be Extended?

Beyond whether the native AI layer covers current requirements, there is a second question that gets asked less often, which is whether the agentic framework can actually be extended. This question has two dimensions worth considering separately.

The first dimension is within the platform itself. Most leading DXP vendors have built specialized agents across the major capabilities in their ecosystem, typically spanning multiple products, but two things vary considerably. The first is whether every specific use case relevant to a given organization has a corresponding agent.

The second is the maturity of true multi-agent orchestration, where specialized agents can work together autonomously, whether by handing off to each other, running in parallel, or coordinating under a manager agent, to execute multi-step workflows across the content lifecycle. Vendors are at different stages on both fronts, and the relevant evaluation question is whether the framework allows the organization to fill gaps by writing custom agents that work within the platform, or whether it is limited to the agents the vendor has predefined.

The third dimension is beyond the platform. Custom agents that leverage DXP data for use cases outside the platform's core scope are an emerging possibility, but vendor support varies. In a more composable, multi-vendor scenario, this gets more nuanced. MCP servers (Model Context Protocol, an emerging interoperability standard for agentic systems) may exist at the product level rather than at the suite level.

If a particular product within the DXP comes from a different vendor than the rest of the platform, the question becomes whether the broader agent capabilities work in a platform-agnostic way or whether agents need to connect to that product via its own MCP server. That has implications for how orchestration is handled, particularly when the use case spans capabilities across multiple vendors. Understanding where MCP support sits and what it actually covers is part of evaluating the offering during the selection process.

A Recent Evaluation Worth Referencing

In a recent DXP evaluation I was involved in, this question surfaced directly. The evaluation itself was well-structured, with multiple parts of the business represented beyond just IT and marketing or experience teams, requirements defined, vendors shortlisted, and their agentic offerings being assessed as part of the broader scope.

The prior digital ecosystem at the organization had powered pre-acquisition primarily, and the DXP evaluation was still focused mostly on pre-acquisition, but the goal had expanded toward capturing a more complete view of the customer and delivering personalized content across multiple channels, which meant the platform was beginning to shift toward supporting post-acquisition as well. Running in parallel was a separate initiative, led by a different part of the technology organization, focused on a post-acquisition journey use case.

The work was being done with a specialty AI firm building a custom solution with its own LLM and a set of custom agents. Because the prior digital ecosystem had never been able to support this use case, even partially, it had been scoped as an independent build from the start, and the two efforts were proceeding on separate tracks with little coordination between them.

Part of the work on our side became surfacing where the modern DXP could complement what was being built independently. The CMP could power content publishing into a separate application. The CDP could close the loop on customer data, enabling more personalized messaging within the app while writing engagement signals and intent back to a unified profile.

The agentic layer's role was different. It was less about replacing what the parallel team was building and more about leveraging the DXP's data and platform capabilities, potentially through custom agents connecting via MCP rather than expecting native DXP agents to power use cases outside the ecosystem directly.

None of that conversation was happening, though, and duplication of investment, integration debt, and the possibility of building something the DXP could have partially supported were all live risks when those two tracks did not intersect during the evaluation.

In many enterprises, these conversations are further complicated by the fact that AI initiatives, digital platform ownership, data strategy, and customer experience operations often sit under different leadership structures, which makes overlap harder to surface even when the evaluation itself is well-staffed. The broader observation is that as DXP capabilities expand their footprint, particularly around agentic capabilities, the likelihood of overlap with parallel AI initiatives inside the same organization grows. The evaluation is the right moment to surface that overlap rather than discovering it after a platform has been selected and the architecture has started to harden.

The Co-Existence Question

For organizations with their own agentic framework already in place or being built, the question shifts from whether the DXP's native offering is sufficient to how the two should work together. This is where MCP becomes part of the conversation. The DXP's native agentic layer often includes its own multi-agent orchestration, which means the practical question becomes which orchestration layer takes the lead and where the two connect.

One path is to continue leveraging the DXP's multi-agent orchestration within the platform, using the specialized agents the vendor has built for ecosystem-scoped use cases, and treating the external framework as the layer that handles broader enterprise orchestration outside the DXP's scope. The two coordinate where they need to, but neither tries to absorb the other.

Another path is to lead with the organization's own agentic framework and connect to the DXP via MCP servers provided by the vendor. In this model, the DXP becomes one node within a broader agent architecture, and the organization's framework owns the orchestration logic across systems. The trade-off is that the DXP's native agents are highly specialized and scoped to the vendor's products, so if you are building your own agents outside the DXP's framework, you are potentially limited by what the vendor's MCP server actually exposes. MCP support across DXP vendors is also at various stages of maturity, which factors into how viable this approach is in practice.

In a multi-vendor composable scenario, this gets sharper. As noted earlier, MCP support may sit at the product level rather than at the suite level, which means the external framework will likely need to own orchestration across the end-to-end content lifecycle since no single vendor is positioned to provide it.

There is no single right answer, and the right approach depends on where the organization is in its AI maturity, what existing investments look like, how the use cases are scoped, and whether the underlying architecture is composable or suite-oriented. What matters is that the question is asked explicitly during the evaluation rather than left to be answered later by default.

Evaluation Questions to Surface Early

Does the DXP's native agentic offering adequately cover requirements across the marketing and content lifecycle, or are there specific gaps where a purpose-built tool like Writer or Jasper would supplement it well?
Can the native framework be extended to support custom agents, both to fill gaps where the vendor has not yet built native agents for specific use cases, and to support use cases that reach beyond the platform's core scope entirely?
How mature is the DXP's multi-agent orchestration, particularly the ability for specialized agents to coordinate autonomously across the content lifecycle?
If a separate agentic framework or AI initiative is already underway, where does MCP support sit, at the suite level or at the product level, and how does that shape which orchestration layer takes the lead?

Closing Thoughts

The agentic offering is becoming an increasingly significant part of any DXP evaluation. Most platforms have offered AI assistants for some time, but the shift toward more comprehensive agentic frameworks has accelerated over the past year, and the native capabilities are improving quickly enough that they will continue to cover more ground.

The organizations that get the most out of these investments are the ones that treated the DXP evaluation and the broader AI conversation as connected yet distinct, and that deliberately decided where the two should intersect rather than assuming the contract settled the question. That kind of deliberate decision is harder to revisit once the architecture has started to take shape, which is why the evaluation is the moment to ask the question while the right stakeholders are still in the room.

Upcoming Events

Umbraco Codegarden 2026

June 10–11, 2026 – Copenhagen, DK

Join us in Copenhagen (or online) for the biggest Umbraco conference in the world – two full days of learning, genuine conversations, and the kind of inspiration that brings business leaders, developers, and digital creators together. Codegarden 2026 is packed with both business and tech content, from deep-dive workshops and advanced sessions to real-world case studies and strategy talks. You’ll leave with ideas, strategies, and knowledge you can put into practice immediately. Book your tickets today.

CMS Connect 26

August 5-6, 2026 – Montreal, Canada

The best conferences create space for honest, experience-based conversations. Not sales pitches. Not hype. Just thoughtful exchanges between people who spend their days designing, building, running, and evolving digital experiences. CMS Connect brings together people who share real stories from their work and platforms and who are interested in learning from each other on how to make things better. Over two days in Montreal, you can expect practitioner-led talks grounded in experience, conversations about trade-offs, constraints, and decisions, and time to compare notes with peers facing similar challenges. Space is limited for this exclusive event, so book your seats today.

Open Source CMS 26

October 20–21, 2026 – Utrecht, Netherlands

Join us for the first annual edition of our prestigious international conference dedicated to making open source CMS better. This event is already being called the “missing gathering place” for the open source CMS community – an international conference with confirmed participants from Europe and North America. Be part of a friendly mix of digital leaders from notable open source CMS projects, agencies, even a few industry analysts who get together to learn, network, and talk about what really matters when it comes to creating better open source CMS projects right now and for the foreseeable future. Book your tickets today.

Contentstack ContentCon 2026

September 30 - October 1, Amsterdam / October 27-28, 2026 – Chicago

Contentstack’s annual customer conference is the premier event for executives, marketing leaders, and developers to redefine their digital experience strategy. This is your opportunity to step out of the "status quo" and into "elite" status, learning exactly how the world’s most successful brands are using the technology you already own to do the impossible. Enjoy a full day of interactive workshops, certifications, and inspirational on-stage sessions designed to help you become an expert on cutting-edge digital strategies and how to turn Contentstack's CMS and adaptive personalization tools into your greatest competitive advantage. Book your seats today.

News

Where Native DXP AI Fits Within an Agentic Enterprise Strategy