What happens to SEO in an AI-first world?
The old playbook of SEO is breaking down. Keywords, backlinks, structured rankings. These still exist, but they’re no longer the levers they once were. But the new rules haven’t been written yet.
In 2024, Dotdash Meredith – America’s largest digital publisher – reported that Google search traffic to its sites had halved compared to four years ago. That’s a canary in the coal mine. Something deeper is shifting in how people find information online. The old routes to visibility are being challenged. And AI summaries are starting to replace them.
The browser is becoming the agent
Traditionally, search engines pointed users to pages. Now, AI tools are beginning to bypass the web altogether – delivering answers directly, not just results.
ChatGPT, Claude, Copilot and others are replacing the search box with natural language prompts, and skipping past the list of links we’ve optimised for years. Instead of search engine optimisation, we’re seeing the rise of answer engine interaction.
Meanwhile, a new generation of AI-powered browsers – Arc, Comet, Opera’s Aria – are shifting from passive tools to active agents. They browse for you. Summarise for you. Soon, they’ll make choices for you.
This isn’t just a UX update. It’s a redefinition of how discovery happens.
The comfort of metrics and the danger of false clarity
With every disruption comes a new promise of order. “Share of Model” is the latest buzzword – a tempting shorthand for ranking in generative AI outputs. But we don’t fully control what AI models see, remember or return.
Ask ChatGPT about your brand today, and you might get one answer. Ask again tomorrow with a different phrasing, region or model, and you’ll get another.
The model is opaque. Its logic unknowable. Its memory imperfect. So any system claiming to offer reliable “AI visibility” metrics should be met with healthy scepticism.
From visibility to legibility
So how should brands think about visibility now?
Shift the focus. From trying to rank in a list of results, to being understandable to the systems generating the answers. That means being legible, both to people and to machines.
The future of discovery won’t reward gaming the algorithm. It will reward clarity, credibility, and content that’s structured enough to be recognised, retrieved and reassembled.
AI tools rely on signals: consistency, formatting, trustworthy sources, contextual strength. If your brand shows up the same way in multiple places, with clear messages and accessible structure, you’re more likely to be surfaced.
In a world of infinite synthesis, the brands that are most easily understood will be the ones that endure.
What to do next:
There’s no fixed roadmap. But there are some durable principles:
Be clear – focus on usefulness, not fluff.
Be structured – use semantic HTML, clear metadata, accessible formats.
Be consistent – align your tone, messages and facts across all channels.
Be adaptive – revisit your strategy regularly as tools and interfaces evolve.
We're not just seeing a new version of SEO. We're seeing the early stages of AI-first discovery—where interfaces act, interpret and influence what people see.
Visibility still matters. But it’s being rewritten in real time. In this environment, legibility is the new competitive edge.
