This week delivered a sharp set of signals for any agency managing search visibility and paid media budgets. Amazon is facing a landmark federal lawsuit over alleged advertising overcharges, Google’s John Mueller tested whether AI crawlers actually request markdown, and fresh data shows Anthropic’s Claude products search the web in fundamentally different ways. Here’s what our team is acting on right now.

Key Takeaways

• The FTC and 22 states allege Amazon overcharged advertisers by more than $20 billion through manipulated ad auctions.
• Marketing teams need three specific role changes and a budget rebalance to earn citations in AI-generated search results.
• John Mueller tested markdown delivery for AI crawlers — and the results challenge assumptions about AI SEO best practice.
• Claude and Claude Code search the web at different rates, mention different brands, and visit different page types entirely.
• Anthropic has issued a warning that infostealer malware is hijacking Claude login sessions, creating a direct security risk for teams using AI tools.

Amazon’s Ad Auction Practices Under Federal Fire

The Federal Trade Commission, backed by 22 US states, has filed suit against Amazon alleging the company manipulated its advertising auction system to extract more than $20 billion in overcharges from advertisers. This is a significant development we’re watching closely, as detailed in Search Engine Journal’s coverage of the FTC lawsuit against Amazon.

For our clients running Amazon Ads, the immediate action is straightforward: audit your cost-per-click trends over the past 18 months and benchmark them against conversion value. If the allegations hold, advertisers may have been paying inflated rates without corresponding performance gains. We’re advising clients to diversify paid spend across channels and reduce over-reliance on any single marketplace auction system.

Restructuring Marketing Teams for AI-Search Visibility

Getting cited in AI-generated answers requires a different internal setup than ranking in traditional SERPs. According to Search Engine Journal’s guide on restructuring marketing teams for the AI-search era, three role changes and a budget rebalance can shift a team from invisible to cited — without adding headcount.

We’ve already started applying this framework across several client accounts. The core shift involves redirecting resource from pure keyword-targeting content toward entity-rich, structured information that AI models can confidently cite. Content strategists need to think in terms of authoritative statements, not just search volume. Budget should move from high-volume blog production toward fewer, better-sourced assets.

Mueller Tests Markdown Delivery for AI Crawlers

Google’s John Mueller ran a practical experiment: he checked whether AI crawlers were actually requesting markdown versions of his sites. The findings, reported in Search Engine Journal’s piece on Mueller’s markdown AI SEO test, challenge the assumption that serving markdown to bots is a priority.

Our technical SEO team has taken note. Before investing development time in markdown delivery systems, we’re now verifying actual crawler behaviour in server logs first. The takeaway is simple: test before you build. Not every AI SEO tactic floating around is grounded in real crawler demand.

Claude and Claude Code Search the Web Differently

New data confirms that Claude’s consumer-facing product and Claude Code — its developer tool — behave as distinct agents online. They search at different rates, reference different brands, and land on different page types, as outlined in Search Engine Journal’s analysis of Claude versus Claude Code search behaviour.

This matters for visibility strategy. If your site targets developers, optimisation for Claude Code’s crawl patterns is a separate task from general AI visibility. We’re segmenting our AI-referral traffic analysis by agent to understand which Claude variant drives visits — and to which content.

Claude Session Hijacking Poses Real Security Risk

Anthropic has warned that infostealer malware is actively targeting Claude login sessions, allowing attackers to drain account usage. The details are covered in Search Engine Journal’s report on Claude session hijacking. For any agency or in-house team using Claude in daily workflows, this is a direct operational risk. We’ve tightened session management protocols and enforced browser-level security policies for all team members using AI tools.

The common thread across all five developments is clear: AI is not a future consideration — it is reshaping paid media economics, organic visibility mechanics, and operational security right now. Teams that treat these signals as actionable intelligence, rather than background noise, will hold a measurable advantage through 2025 and beyond.

Frequently Asked Questions

How do web designers optimise sites for AI Search crawlers like Claude?

Start by analysing server logs to identify which AI agents actually visit your site and what content they access. Then structure pages with clear entity-based information and authoritative sourcing that AI models can confidently reference in generated answers.

What is the FTC’s case against Amazon’s advertising auctions?

The FTC and 22 states allege Amazon altered its ad auction mechanics to overcharge advertisers by more than $20 billion. Advertisers should audit recent cost-per-click trends and benchmark them against actual conversion performance.

Why does AI-search visibility require a different SEO strategy?

AI models cite sources based on authority, clarity, and structured data — not traditional keyword density or backlink volume alone. Marketing teams need to shift budget toward fewer, higher-quality content assets that establish topical authority.

How do marketing teams protect AI tool accounts from session hijacking?

Enforce strict browser security policies, use hardware-based authentication where possible, and monitor for infostealer malware across all devices. Regularly rotate sessions and restrict AI tool access to managed, secured environments only.