AI is a mirror. Make sure to put good stuff in front of it.
By 2026, most organizations have stopped asking whether to adopt AI and started asking why it isn't working harder for them. The models are trained, the tools are deployed, the experiments are behind us. Expectations are high. Yet the conversation about AI impact still circles the same territory.
By Will Robinson, CEO Hygraph, and Ola Linder, Head of Marketing SQLI Nordics
AI copilots are now standard issue across development and marketing teams. Code gets written faster. Approval requests pile up even faster. Releases ship on the same schedule. Marketing teams produce more content, but conversion rates, differentiation, and demand don't really move. Despite an estimated $30–40 billion in enterprise GenAI investment, research suggests 95% of organizations are seeing no measurable return.
Somewhere between the promise of AI-powered digital experiences and the reality of marginally faster content production, something got lost.
The question worth asking is not whether AI works. It is why, despite AI being everywhere inside the organization, the customer experience on the outside barely moves. In a recent conversation, we explored that gap and how content, structured and governed the right way, is where AI finally starts delivering value beyond the productivity layer.
Why digital experience is the key to delivering AI value
Despite these high adoption rates, businesses have frequently struggled to derive true value from AI in customer-facing and relationship-driven applications. Here are examples and insights into where AI has fallen short:
- Inability to retain client knowledge and adapt: While AI can generate text, it often fails to understand the deep context of the client. A respondent in a recent survey highlighted this limitation: "It’s excellent for brainstorming and first drafts but doesn’t retain knowledge of client preferences or learn from previous edits. It repeats the same mistakes and requires extensive context input for each session. For high stakes work, I need a system that accumulates knowledge and improves over time".
- Lack of brand voice and genuine relationships: In marketing and customer engagement, AI cannot truly capture a brand's "soul". And if you expect it to, that soul will soon look like everyone else's. AI can be creative, but only as creative as what you put in front of it.. Experts advise keeping high-touch customer engagement and influencer relationship-building strictly manual, noting that "people still prefer talking to people".
The difficulty of extracting value is a widespread enterprise issue. A recent MIT report noted that despite billions being invested into enterprise generative AI, 95% of organizations are getting zero return on those investments.
Look closely at each failure and a pattern emerges. Client context is lost because it lives in email threads and tribal knowledge, not in structured, accessible content. Personalization falls flat because the content feeding those experiences was built for pages, not for people. AI is only as good as the experience layer it operates through, and the experience layer is only as good as the content that powers it. When that foundation is missing, AI generates output. It just does not deliver value.
Your content should reflect the digital experience you want to deliver. And right now, for most organizations, it does not.
Why most digital experience stacks fall short in the AI era
Most experience stacks were designed for publishing, not for personalizing, activating, and engaging users.
Human content published in page-centric, asset-centric systems cannot explain context, relationships, or intent to a machine. This is not a tooling problem. It is an architectural one.
Consider a brand manager launching a product across three new markets. They have approved content, brand guidelines, and regional compliance requirements are ready. In a publishing-first CMS, they rebuild everything manually per market. Pages are duplicated and siloed. Brand guidelines live in a PDF nobody references. Regional variations exist as separate assets with no relationship to the original.
When AI is introduced into this stack, it encounters the same chaos. It cannot distinguish a regional variant from a separate piece of content. It cannot trace a product claim back to its approved source. It cannot understand that the German landing page and the English one are the same experience, expressed differently.
The system was never built to express those relationships. So the AI does what it can: it summarizes, it generates, it fills gaps. But it cannot reason, personalize, or govern. And the experience on the other end reflects exactly that.
According to CMSWire, only 19% of organizations say they understand their customers "well" despite increased AI adoption. When content has no structure, no relationships, and no governance, that number should not surprise anyone.
Modern experiences require information with structure, relationships, reuse, and governance. Getting there means rebuilding from the foundation up.
What it actually takes to connect content to AI value
That shift starts with how content is created, structured, and owned. And it requires getting three things right.
Quality over volume
Ola raised a point that cuts to the heart of the problem. AI is remarkable at generating concepts, plans, and first drafts. But it draws from what already exists. It is not truly creative.
In a B2B setting, where every competitor has access to the same tools and the same models, the risk is that AI produces average content at scale. If you ask ChatGPT to build a campaign and your competitor does the same, the output will likely be more similar than different. The model knows your industry, your region, your use case, but so does everyone else asking the same question. The result is a market flooded with content that meets a baseline quality bar and clears nothing above it. Who will read it all?
The answer is to use human judgment more deliberately. Quality is not a feature you add at the end. It is a decision you make at the start, about what only your brand can say, what only your customers have experienced, and what no model trained on the open web can replicate.
Creativity as a competitive asset
AI can execute. It cannot originate. The brand voice, the unexpected angle, the insight that comes from actually talking to a customer: that is still human work.
In an environment where AI lowers the cost of content production to near zero, creativity becomes the scarce resource. The brands that will stand out are the ones that produce content with a point of view that is genuinely theirs. That requires protecting the creative layer from the efficiency impulse, resisting the temptation to let AI do the work that only humans should do.
Clear ownership: content for machines and content for humans
Not all content serves the same purpose anymore. Some content is written to be read. Some content is written to be understood by machines: to feed AI systems, power personalization engines, and structure experiences dynamically.
Most organizations have not drawn that line yet. The result is content that does neither job well. Establishing clear ownership means deciding deliberately: who is responsible for the content that trains and feeds your AI systems, and who is responsible for the content that your customers actually experience? Without that distinction, quality erodes on both sides.
What’s next
AI transformation is a compelling idea. What it looks like in practice remains uncharted for most organizations. What is clear is that the foundation matters more than the model. Organizations that layer AI on top of broken content architecture will find that AI only amplifies the problem instead of becoming the solution. The brands that will close the gap between AI capability and business impact are the ones that treat content not as output, but as infrastructure.