generative engine Optimisation is no longer a single-track discipline. This week’s data confirms what our team has been testing across client campaigns: AI platforms behave differently depending on mode, bots now outpace human visitors, and video content is earning AI citations at scale. Here’s the raw intelligence and exactly how we’re acting on it.

Key Takeaways

• ChatGPT’s Thinking mode and Instant mode cite different sources — with only 25% overlap — meaning a single GEO strategy is insufficient.
• Bot traffic has officially surpassed human traffic for the first time, driven largely by AI agents crawling the web.
• Video SEO now directly influences AI citation rates across Google, YouTube, and generative platforms.
• Digital marketing strategies must be rebuilt around multi-channel AI visibility, not just traditional search.
• Enterprise-scale SEO tooling is diverging from standard platforms, demanding clearer decisions on infrastructure.

ChatGPT Cites Different Sources Depending on Reasoning Mode

This is the finding that changes how we approach GEO for every client. New research shows that only 25% of cited sources overlap between ChatGPT’s Thinking mode and Instant mode. In practical terms, a brand that earns a recommendation in one mode may be completely absent in the other — even for the exact same prompt.

Our team is now running dual-mode audits on all priority queries. We test each client’s visibility in both reasoning modes and map which content assets surface where. The content that performs in Thinking mode tends to be longer, more technically detailed, and structured with clear entity relationships. Instant mode favours concise, authoritative pages with strong topical signals. We’re adjusting content briefs accordingly.

Bot Traffic Has Overtaken Human Users

The numbers are stark. As reported in recent analysis, bot traffic now exceeds traffic from human users, with AI agents responsible for the surge. This has direct consequences for analytics accuracy, server load, and how we interpret engagement metrics.

What we’re doing right now:

• Filtering AI bot traffic in GA4 and server logs to isolate genuine user behaviour.
• Reviewing robots.txt and crawl directives to ensure AI agents can access the pages we want cited — and are blocked from thin or duplicate content.
• Monitoring crawl budgets more aggressively, particularly for larger client sites where bot activity can crowd out Googlebot.

If your analytics show a sudden traffic spike but conversions remain flat, bots are the likely explanation.

Video SEO Now Feeds AI Citation Pipelines

Video is no longer just a YouTube play. Our content strategists are treating it as a core GEO asset after reviewing updated guidance on how to optimise video for YouTube, Google, and AI platforms. Structured transcripts, descriptive titles, and schema markup on embedded videos all increase the likelihood of AI engines pulling video content into generated answers.

We’re embedding optimised video on key service pages and ensuring transcripts are crawlable. For clients in competitive verticals, this is producing measurable lifts in both traditional SERP features and AI-generated responses.

Digital Marketing Strategies Need an AI Visibility Layer

Traditional channel planning — paid, organic, social — is incomplete without an AI visibility component. A comprehensive breakdown of modern digital marketing strategy types confirms that brands must now account for how they appear in AI-generated answers, not just conventional search results. We’re building this into every strategy document we produce.

Choosing the Right SEO Infrastructure at Scale

For our larger clients, tooling decisions matter. The gap between standard SEO platforms and enterprise-grade solutions is widening, as detailed in a direct comparison of Semrush and Semrush for Enterprise. We evaluate each client’s reporting complexity, team size, and multi-domain needs before recommending a tier. Getting this wrong wastes budget and limits visibility tracking.

GEO is moving fast, and the data this week makes one thing clear: treating AI as a monolithic channel is a mistake. Our agency is splitting strategies by AI mode, filtering bot noise from real engagement data, and layering video into every visibility plan. Brands that adapt now will hold position. Those that wait will find themselves invisible in the answers that matter most.

Frequently Asked Questions

What is Generative Engine Optimisation (GEO)?

GEO is the practice of optimising content so it gets cited and recommended by AI platforms like ChatGPT, Gemini, and Perplexity. It builds on traditional SEO but focuses specifically on how generative AI models select and surface sources.

Why does ChatGPT cite different sources in Thinking mode versus Instant mode?

Each mode uses a different reasoning process, weighting source authority, depth, and structure differently. This means a page visible in one mode may not appear in the other, requiring brands to optimise for both.

How do web designers account for increased bot traffic on client sites?

We filter bot traffic in analytics to protect data accuracy and adjust crawl directives so AI agents access priority pages. Server resources and crawl budgets also need monitoring to prevent bots from degrading performance for real users.

How does video SEO improve AI visibility?

AI engines pull from structured, crawlable content — and optimised video transcripts, schema markup, and descriptive metadata give them exactly that. Embedding well-optimised video on key pages increases the chance of being cited in AI-generated answers.