Companies are scrambling to rewrite decades of content, from product pages to policy documents, as AI agents increasingly replace search engines in the customer journey.

Those still running monolithic CMS platforms risk being invisible to the LLMs that are starting to dominate product discovery.

Developers are turning to AI chatbots instead of Reddit, HackerNews or GitHub to discover new tools or query coding bugs. OpenAI and Google are developing agentic shopping where customer journeys that never include a human being are entering the mainstream. 

With OpenAI's ChatGPT and Google's Gemini overviews increasingly mediating product discovery, companies are treating Generative Engine Optimization (GEO) with the same urgency they once reserved for SEO.

How chatbots present content may already be impacting major procurement decisions. Fortune reported that chatbots and agents researching on users' behalf may push users toward larger enterprises, simply because there is more information available on the internet.

The Stack spoke to enterprise content experts about how companies are evaluating their content strategies and the infrastructure that underpins them. 

Content has an architecture problem 

Traditional monolithic CMS platforms lock content inside proprietary silos and marketing teams need IT intervention for basic updates. The same product description exists in six different systems, Sourajit Ghosh (SG), Chief Expert at SAP, told The Stack.

Whereas, a headless architecture centralises content in one API-driven repository, pushing updates to every touchpoint simultaneously. One change propagates everywhere. When information is repeated across systems, small inconsistencies accumulate, weakening confidence in its accuracy and reliability, especially for automated reuse.

SG said, “Architecturally, content is starting to play a very big role in transformation,” adding customers now want to “decouple storefronts from content platforms so marketers, platform engineers and merchandisers can work across the two experiences.” 

This kind of migration “empowers the merchandises, the marketers, it frees up the IT people not just to always go and do the technical updates,” according to SG. For digital enterprises, content visible to the public is increasingly shaping discovery and engagement and, to an extent, the business itself.

Mike Reynolds, a principal solution architect with Contentful, said customers often start with upgrading their commerce environment from a hybrid on-prem architecture to SAP’s Commerce Cloud. “As part of that, they start to think through, ‘What else should I be doing? What else can I be doing as part of this migration?’”

SG believes enterprise agentic AI systems and consumer-facing GenAI chat apps will create a “paradigm shift” when it comes to creating content as people will no longer scroll through information top to bottom: “Everybody will have to start looking at rewriting their content in such a way (...) that when you will do vectorisation and retrieval augmented generation out of it, the chunking will be accurate enough that somebody at the very end who is asking questions will be able to extract the few lines.” 

Companies who can update and structure their content to make it generative AI-friendly may have a leg-up in the race for better chatbot discoverability. To do that, it needs to be centralised, with updates that can be pushed out across platforms. 

AI isn’t always top of mind during migrations 

However, most CMS migration projects start with customers looking to improve efficiency or reduce costly duplications, not GEO. 

“The common reason that we're being brought into this most of the time is efficiency. They're looking to optimise their processes,”  Reynolds said, adding, it’s common for large customers to have multiple sites across WordPress, SmartEdit and home grown solutions, who want to centralize their content and disseminate it consistently. 

Understanding the landscape and what an organisation wants to achieve by modernising their CMS is the first step, then comes the planning.

Both Reynolds, a former consultant, and SG agreed planning was critical, however SG pushed IT and digital leaders to consider the next steps after migration. 

“You shouldn't just plan content migration from a legacy tech stack into a modern tech stack. Yes, that's the walk phase, but you need to also start defining the run and the fly phase.” 

“So what do you do at the run phase where you have a connected CX platform with the Contentful stack? And then what do you then do at the fly phase with the genetic AI and data and AI use cases?” SG asked. 

SG added, “Very few organisations have started to even figure out how to address true hyper-personalisation.” He gave the example of hardcoding a specific hero banner for one customer segment into the CMS, “That's not scalable because the customer segment itself will keep changing and evolving, and it has to be dynamic.” If an enterprise can connect its CMS, AI and customer data platforms, it can begin to scale automating content personalisation and adapting to new segments.

Integrating data can be a hurdle 

Using a modern CMS companies can collect data on how customers interact with their content and use that data to personalise and create targeted customer engagement. 

Reynolds explained one scenario where customers could integrate data collected from Contentful and SAP Commerce Cloud with data from other external platforms in a larger data lake. The combined data allows companies to better understand how customers interact with their content and refine how to engage that customer. 

However, Reynolds said this is where customers can run into hurdles with regards to how they integrate these different data sets. SG agreed, “There is a lack of integrated, connected, real-time data for that personalisation to happen.”

Reynolds said the end goal is “being able to use that data across all of those different touch points, bring it in and centralise it, and then be able to use that to identify and put people into these segments, into these groups, and then use the content management system and AI to target your content specifically for them.”

SG added, “Every organisation I've met does not have clean data.” However, for personalisation and customer experience, the data doesn’t have to be perfect. SG said data engineers need to strike a balance between agility, feasibility and outcomes.

“Data has to be absolutely accurate if you are creating a predictive model to do cancer research,” SG continued, “But not in the kind of domain where we are and where we can live with a certain level of inaccuracy in order to actually go ahead with velocity and speed, example in e-commerce personalised recommendations.” 

The Contentful Composable Content Platform is available via SAP Commerce Cloud as an SAP endorsed app.


Delivered in partnership with Contentful and SAP.

The link has been copied!