Should We Optimise Content for AI? A Look at GEO

As AI systems consume more of the web, should creators optimise content for machines? Questioning the logic of Generative Engine Optimisation.

Generative Engine Optimisation proposes that content creators can - and perhaps should - optimise their writing to become more visible, useful, or quotable within these systems. After all, if people are getting their information from AI, then being part of those responses could seem critical to digital visibility.

But should we accept that logic? Do we actually know what works? And more importantly, do we want it to?

This piece unpacks some of the assumptions behind GEO, weighs its potential risks and rewards, and asks a more foundational question: what kind of internet do we want to write for?

1. We Don’t Really Know What Works, or If It Works at All

There are no reliable rules for how to get your content “picked up” by AI systems. Some rely on old training data. Some use live search. Some cite sources, others don’t. Some hallucinate. The ecosystem is opaque by design.

Despite this, a small industry of GEO advice has emerged, largely based on extrapolations from SEO, speculation about AI retrieval mechanisms, or intuition about what “sounds good to machines.”

In reality, we simply don’t know what makes content more likely to be surfaced or referenced by AI systems. That makes any attempt at optimisation speculative at best, and potentially misleading.

2. The Costs May Outweigh the Benefits for Creators

Even if GEO did work, it’s not clear that creators should want it to. Most AI systems do not send traffic back to source sites. Many do not credit original work. And none compensate the creators whose work they draw upon.

If your content is used in an AI-generated answer, it may never be attributed. Your voice may be flattened or abstracted. You might reach more people, but receive no signal, no feedback, and no return.

In this light, GEO may amount to unpaid training and free content for commercial systems that offer creators no meaningful stake in the value they help generate.

3. Optimising for Machines Risks Homogenising the Web

History gives us a cautionary tale. Search Engine Optimisation led to templated articles, clickbait, and shallow content designed to please algorithms rather than readers. GEO could trigger the same dynamic, but faster.

If everyone writes in machine-legible formats, using neutral tone, simplified syntax, and predictable structures, we risk creating a web that is easier for AI to consume but less interesting for humans to explore. We lose voice, creativity, and context in the name of clarity and structure.

Optimising for retrieval is not the same as writing for understanding.

4. Still, There Are Reasons to Stay Machine-Aware

There is, however, a counterargument worth considering: AI tools are now integral to how information is accessed. If your goal is education, influence, or discovery, then being referenced by these systems will become valuable, even if the benefits are currently unclear.

And it’s possible that this value will continue to grow. Future models may cite sources more reliably. They may reward high-quality content with attribution or inclusion in premium experiences. Visibility could become a proxy for authority.

This is the hopeful case for GEO: not as a manipulative tactic, but as a way to ensure your work remains relevant as information flows change.

But this argument only holds if visibility eventually comes with accountability and reward. Until then, it’s a gamble.

5. There Are Principles Worth Following, With or Without GEO

Even if you reject GEO as a discipline, some practices do align with long-term resilience. These are not intended to game the system, but rather to create content that holds up across channels, formats, and audiences.

  • Write with clarity and purpose. Whether read by a person or parsed by a system, clear thinking matters.
  • Structure your ideas. Use logical flow, helpful headings, and contextual links. It aids both understanding and referencing.
  • Share original insight or data. Repetition is noise. What’s unique to your view, your work, your evidence?
  • Explain relationships between ideas. Make connections visible. Models and people alike benefit from conceptual clarity.
  • Maintain and revise over time. The internet is dynamic. Let your content evolve with it.

These are not GEO tactics. They are durable practices that improve your work regardless of how AI develops.

6. Ask the Deeper Question: Who Are You Writing For?

This is the heart of the issue. Are you writing to be cited by machines, or to connect with people? Optimisation is always a trade-off. GEO, if taken seriously, could reorient your voice, structure, and strategy around unknown systems with unclear incentives.

If you want to teach, persuade, provoke, or build community, AI systems are not your audience. They are filters. They are not the reader. They are intermediaries.

Chasing machine visibility risks distorting the very thing you’re trying to express.

Final Thought: Visibility Without Value Is Not a Goal

GEO appears to be evolving into something useful. But today, it is a hypothesis built on opaque systems, uncertain returns, and a weak understanding of how AI uses the open web.

Rather than optimising for inclusion in black-box systems, content creators would do better to build resilience. Create work that stands on its own. Make it human-readable. Make it hard to misrepresent. Make it worth returning to.

Let AI adapt to that.

Want to hear more?

Generative Engine Optimisation is an evolving field, distinct from but complementary to SEO. While traditional SEO focuses on search engine rankings, GEO aims to make your content more likely to be referenced and recommended by AI assistants.

There are still many unknowns, but using the guidelines above to frame the type of content you produce and how you produce it provides a solid foundation for current and future AI systems.

Working out how AI changes your digital strategy - from content to product - is something we help businesses think through. Our Digital Product & AI service.

Further Reading

  • How AI Is Shaping the Future of Browsing the Internet.
  • Designing for Everyone. What the European Accessibility Act Means for Your Digital Products.
  • AI Consultancy
  • AI in Action: Google Research on Progress Towards the Sustainable Development Goals.
  • Privacy Policy

Want to discuss this further?

We're always happy to talk through ideas.