"Certain formats consistently trigger AI Overviews: clear explainers and definitions, step-by-step how-tos, detailed comparisons and buying guides, and local discovery content with consistent business details," writes Laughlin Rigby, Digital Transformation & AI Director of Core Optimisation

The search experience has fundamentally changed. Users no longer just scroll through ten blue links, they are reading answers. Google's AI Overviews, now reaching 1.5 billion users across 200+ countries, synthesise trusted sources into concise, cited summaries above the fold. Meanwhile, agentic browsers like ChatGPT Atlas, Perplexity's Comet, and Edge Copilot sit beside the page, comparing options and completing tasks on behalf of users.

This shift demands a new approach: be citable, machine-readable, and action-ready. The classic pillars of SEO remain foundational. What's changed is how we express them, ensuring machines can reliably extract and act upon our information.

The behavioural shift

When AI-generated answers appear above the fold, click-through behaviour transforms. Traffic concentrates on pages offering canonical facts, original insights, and proprietary data, which is precisely what AI assistants quote.

Expect fewer low-intent sessions but higher post-click quality. KPIs must evolve to emphasise assistant-layer visibility through citations and mentions, alongside completion metrics like enquiries and bookings. The content strategy follows: write answer-first, then layer in nuance. Address likely follow-ups on the same page.

Becoming citation-worthy

Earning citations from AI systems requires deliberate choices.

Remain discoverable. Allow reputable AI crawlers like OAI-SearchBot in your robots.txt. Keep high-value pages indexable and track assistant-origin traffic.

Structure for dual audiences. Employ semantic HTML and WAI-ARIA. Write one idea per paragraph. Use proper tables for data. Implement schema.org markup for Article, Product, FAQPage, Organisation, and Person, always including datePublished and dateModified timestamps. Clear headings and concise sections improve both human comprehension and machine extraction.

Deploy focused FAQs. On high-intent pages, add four to six genuine questions marked up as FAQPage schema. Place direct answers first, then link to details.

Demonstrate provenance. Feature named authors with verifiable credentials. Cite reputable sources and maintain claim consistency across channels.

Maintain currency. Regularly update priority pages, eliminate duplicates, and establish a single canonical source for each claim, price, or specification.

Provide reliable APIs. AI assistants act, not just read. Publish OpenAPI endpoints for dynamic information like prices and availability with stable IDs and clear errors. For complex interactions, develop an official integration.

Measure rigorously. Track AI Overview presence, citation share, and SERP features. Monitor assistant referrals using server-side tagging. This is continuous, data-driven optimisation for the assistant era.

Campaign strategy in the AI Era

Use consented, verifiable sources and label AI-assisted copy where appropriate. Develop a concise brand style guide documenting preferred phrases and model summaries to help your tone survive AI paraphrasing.

Establish a visible "source of truth" for prices, policies, and key figures. Position this prominently and mirror it in structured data, ensuring AI assistants quote your facts directly.

Agent-friendly design

Agentic browsers navigate via the accessibility tree, making inclusive design both ethical and strategic. Bind inputs with proper labels, use descriptive button text, apply ARIA landmark roles, keep states predictable, and design forms for autofill and programmatic submission. Your goal: the agent completes the intended action correctly, first time.

Preparing for agentic commerce

September 2025 brought two competing standards: OpenAI and Stripe's Agentic Commerce Protocol (ACP), now powering live...