AI Search in 2026: BOFU Content and Agent Transactions

AI search in 2026: why BOFU content wins and what comes after citations

Two shifts are happening at the same time in 2026: Google AI Overviews are stripping clicks from informational content, while AI agents are preparing to complete transactions directly on your website without any human input. If you are still optimising mainly for informational queries and classic SERP positions, you are building on a foundation that is actively moving.

Key takeaways

• AI Overviews reduce CTR for top-ranking pages by an average of 58%, and 95% of affected keywords have low or zero commercial value.
• BOFU comparison content is cited significantly more often by LLMs than general informational articles.
• From June 2026, Chrome auto-browse is live on 200 million Android devices and AI agents are starting to complete full transactions on behalf of users.
• LLM referral traffic appears as direct traffic in GA4, so standard metrics undercount the real contribution.
• Robots.txt and technical site readiness now affect visibility in both AI citations and agentic transactions.

Informational content is losing clicks, not relevance

The data is clear. Research cited by HubSpot shows that pages ranking first with an AI Overview lose an average of 58% of their clicks, with some studies pointing to a 61% drop between June 2024 and September 2025. Nearly half of HubSpot's own keyword portfolio already triggers an AI Overview at the top of the page.

Here is the important nuance: 95% of the keywords that trigger AI Overviews carry low or zero commercial value. In other words, Google is replacing answers to questions you were never making much money from anyway. Transactional and comparison pages, including pricing, demo, and vs. pages, are largely surviving in the classic SERP format for now.

That does not mean you abandon informational content. According to Search Engine Land, it should play a different role: building topical authority and channelling internal link equity toward BOFU pages. Retrofit existing top-of-funnel articles with product-connected sections, contextual CTAs, and expert quotes so that informational content starts to pull its weight commercially.

Why BOFU content dominates LLM citations

Search Engine Land describes a concrete example: a comprehensive comparison guide for time-tracking software in the construction industry became the most-cited article in LLM responses and generated more pipeline than a dozen informational posts combined. That is not a coincidence.

LLMs favour content that is specific, trustworthy, and answers purchase-stage questions directly. Crucially, trustworthiness here includes honesty: listing the downsides of your own product, presenting real comparisons, and not writing like a press release.

The recommended shift is to redirect 60 to 80% of content output toward BOFU and mid-funnel formats. Comparison guides, high-intent listicles, and review pages are the formats that AI models cite and that also convert.

For finding the right topics, Semrush recommends combining traditional keyword mapping with AI-style prompts grouped by intent (learn, compare, buy). The Prompt Research report lets you see which queries your audience is typing directly into AI tools and where competitors are already dominating.

AI visibility in 2026: from citation to transaction

This is where the biggest shift sits. Search Engine Journal reports that from June 2026, Chrome auto-browse arrives on 200 million Android devices as an OS-level feature, with no installation required. The AI agent is simply there.

And this agent does not just want to find your website. It wants to complete a booking, navigate a checkout flow, and confirm an order. Google has published a ten-step roadmap (UCP protocol, A2A, AppFunctions) that forms the foundation of what is being called the agentic web stack, and Project Mariner has reached 83.5% on the WebVoyager benchmark.

The problem: sites built on visual builders like Wix, Bubble, or Figma Sites render by default through JavaScript hydration. The agent sees an empty shell, the same way AI crawlers see empty content. Booking forms, calendar widgets, and multi-step checkouts need to work without human clicks, otherwise agentic traffic simply passes you by.

AI visibility now rests on four pillars: Identity, Structure, Content, and the new one, Interaction. That fourth pillar is load-bearing for transactional sites, not an optional extra.

How to measure when the metrics are lying

Here is the uncomfortable truth: traffic from ChatGPT, Perplexity, or Claude shows up as direct traffic in GA4. Standard SEO reports do not see this contribution at all.

HubSpot recommends segmenting keywords into four buckets based on the combination of ranking position and AI Overview presence, then comparing CTR across them. This lets you separate what was caused by a ranking drop from what was caused by the AI Overview itself.

For LLM tracking, build a regex segment in GA4 covering ChatGPT, Perplexity, and Claude domains. Track brand search volume growth as your primary metric, because it reflects actual AI visibility better than total traffic figures.

The same principle applies from cross-platform PPC measurement: always validate conversion tracking before evaluating attribution. Short conversion windows hide upper-funnel contributions, and direct customer feedback through CRM data will surface what platform reports miss.

Technical preparation that pays off now

A few concrete steps worth doing immediately:

• Robots.txt: never block AI crawlers if you want visibility in AI-generated responses. Semrush flags this explicitly.
• Bing Webmaster Tools: set up access to the AI Performance report, which tracks performance in Bing Copilot and AI summaries. Google Search Console does not yet offer this separately.
• Server-side rendering: test your entire booking and checkout flow from an agent's perspective. Client-side rendering hides transactional elements from AI agents the same way it hides content from crawlers.
• Schema markup and Q&A sections: structured data improves machine readability, and concise Q&A blocks help answer engines extract and cite your content.

If you want to track these shifts systematically and respond to them without building everything from scratch, Charlie's AI specialists cover SEO analysis, content strategy, and cross-platform campaign measurement in one place.

The changes are significant, but the logic is straightforward: content that helps people decide and buy will work in classic SEO and in AI search. Technical readiness will determine whether agents complete your transactions or skip straight past you.