Why the guide matters
Google's guide arrives at a moment when AI-generated answers have become the dominant search experience for nearly half of all queries. The shift has produced a fast-moving ecosystem of acronyms — AEO (Answer Engine Optimization), GEO (Generative Engine Optimization), LLM SEO — alongside an equally fast-moving market for tools and tactics promising to influence what AI systems cite.
A meaningful share of that activity has been built on inference rather than evidence. Google's position, until now, was largely conveyed through scattered blog posts, LinkedIn comments and conference talks. The new guide consolidates that position in a single, authoritative document, and is direct about which tactics it considers ineffective.
For brands investing in AI visibility, the guide provides a stable reference point. For agencies and consultants, it sets a clearer benchmark for what credible advice looks like.
What the guide actually says
The guide's core argument can be summarised in one sentence: optimising for generative AI search on Google is the same discipline as SEO, applied with greater rigour to content quality and technical foundations.
Google states this explicitly: "From Google Search's perspective, optimizing for generative AI search is optimizing for the search experience, and thus still SEO." AEO and GEO, as separate disciplines, are not endorsed.
The guide explains that AI features in Google Search are grounded in the company's existing ranking infrastructure through two mechanisms:
- Retrieval-augmented generation (RAG), in which the AI response is grounded in pages retrieved by Google's core ranking systems.
- Query fan-out, in which the model generates related queries to retrieve a broader set of supporting pages before composing an answer.
The implication is that the pages eligible to appear in AI Overviews and AI Mode are, in practice, pages already performing well in Google's core index.
See where your brand currently stands in AI search. The PQ Impact AI Brand Visibility Report tracks how ChatGPT, Claude, Gemini and Perplexity describe your brand against your competitors, surfaces the prompts where you are absent, and ends with a prioritised action plan.
The four foundations Google recommends
The guide identifies four areas that drive visibility in AI search.
Create non-commodity content. Google draws a sharp line between commodity content (its example: "7 Tips for First-Time Homebuyers") and non-commodity content (its example: "Why We Waived the Inspection & Saved Money: A Look Inside the Sewer Line"). The latter — first-hand, experience-led, with a genuine point of view — is what AI systems are designed to surface. Recycled or AI-generated summaries of existing content are explicitly called out as the wrong direction.
Maintain a clear technical structure. Pages must be indexable and eligible to appear with a snippet. Crawlability, JavaScript handling, page experience, semantic HTML and the reduction of duplicate content all continue to matter. None of this is new; Google's point is that AI features amplify the cost of getting it wrong.
Optimise local business and ecommerce details. Where relevant, Merchant Center feeds and Google Business Profile information feed directly into AI responses. For hospitality, retail and local services, this is now a primary visibility channel.
Explore agentic experiences. AI agents — browser-based systems that perform tasks on behalf of users — are an emerging category. Google recommends reviewing agent-friendly website best practices, including clean DOM structure, accessibility-tree integrity and visual rendering quality, all of which agents rely on to interpret a page.
What Google says you can ignore
The guide contains a "mythbusting" section that is, in commercial terms, the most consequential part of the document. Several widely promoted tactics are explicitly listed as unnecessary:
- LLMS.txt files and other "special" AI markup files. Google does not use them.
- Chunking content into small, isolated answer units purely for AI extraction. Google's systems handle nuance across longer pages, and there is no ideal page length.
- Rewriting content specifically for AI systems. AI understands synonyms and intent; brands do not need to capture every long-tail variation.
- Seeking inauthentic brand mentions across the web. Google's core ranking systems are designed to discount this.
- Overfocusing on structured data. Schema is not essential for AI overview, although it remains useful for rich results in conventional search.
This section is significant because much of the AI optimisation market has been built on precisely these tactics. Brands that have been sold packages built around llms.txt files, aggressive content chunking or schema-heavy retrofits should re-read this section carefully.
That said, the guide does not mean structured data, clean architecture or content design are unimportant. It means they should be applied in the service of human readers and Google's existing quality systems, not as standalone "AI hacks" disconnected from those foundations.
Why the guide is harder to implement than it reads
The four foundations look straightforward on paper. In practice, each requires capabilities that most organisations have split across at least two functions:
- Non-commodity content requires editorial judgement, subject-matter expertise and an honest view of what the brand can credibly say that nobody else can. This is strategic marketing work, not content production.
- Technical structure requires engineering attention to indexability, JavaScript rendering, page experience and DOM integrity. These are development tasks, not marketing tasks.
- Local and ecommerce optimisation requires coordination between marketing, operations and the platforms that hold the source data.
- Agentic readiness requires the technical layer to support clean rendering, accessibility-tree quality and structured interactions — work that sits squarely with engineering.
The pattern is familiar. Marketing owns the content layer; engineering owns the technical layer; and the work that determines AI visibility falls into the gap between them, which neither side owns.
This is the gap most brands underestimate. Following Google's guide well is less a matter of executing a list of tasks and more a matter of having both disciplines operating against the same brief.
Where our expertise sits: marketing strategy and engineering depth, in the same team
PQ Impact is structured to close that gap. Our strategic marketing and custom software development practices operate as a single integrated capability. Our team has built and scaled our own products from the ground up, which means the technical decisions that influence AI visibility — indexability, performance, JavaScript handling, schema where it adds value, page experience, agentic readiness — are decisions we have implemented on our own platforms before recommending them to clients. That depth sits alongside our marketing strategy, demand intelligence and content work, not in a separate function.
For brands working to act on Google's guide, this means the editorial work and the technical work are scoped, prioritised and delivered against the same brief. There is no handoff between an agency that writes the content and a development supplier that implements the infrastructure to support it.
The starting point: a clear picture of where your brand stands today
Google's guide tells you what to do. It does not tell you where you currently stand, or where competitors are winning the recommendations you should be winning. Without that baseline, optimisation work is unfocused by definition.
The PQ Impact AI Brand Visibility Report is built to provide that baseline. It tracks how ChatGPT, Claude, Gemini and Perplexity — together accounting for over 90% of consumer AI traffic — respond to real, neutral prompts in your category. It measures your share of voice across those four assistants, benchmarks you against competitors, surfaces the prompt categories where you are absent or underrepresented, and ends with a prioritised action plan that your team or ours can implement.
The methodology is consistent regardless of sector. For a hotel in Toronto, that means tracking 10 neutral prompts (the kind a real traveller types — "Are there any accessible luxury hotels in Toronto that comply with Ontario's accessibility standards?" rather than "is Hotel X good?") across 22 guest-intent categories, with 30 snapshots per month to capture how answers drift over time. For a B2B SaaS brand, a professional services firm or a consumer business, the prompt set changes; the structure does not.
The report itself is priced from €490 and delivered without ongoing commitment. The deliverables include:
- Share of voice across the four major AI assistants
- Performance by category, so you understand exactly where you are visible and where you are not
- Competitor benchmarking and sentiment analysis, including the signals behind their visibility and how AI systems describe your brand
- A prioritised 12-point action plan, written so your in-house team or agency can implement it directly
From Google's guide to delivered results: how we implement the work
A report that lists 12 actions but leaves the client to find someone to deliver them is a half-finished engagement. Google's guide acknowledges, implicitly, that visibility work spans content, technical foundations and ongoing iteration — work that requires multiple disciplines acting in concert.
Every PQ Impact engagement can move directly from the report into implementation, with the same team across four areas:
- Content and messaging. Refining existing pages and building new ones — including dedicated landing pages and structured Q&A — so the site answers the questions your audience actually asks AI assistants. This is the editorial work Google's guide describes as creating "non-commodity content."
- Technical foundations. Schema markup where it adds value, page structure, internal linking, performance and information hierarchy, so AI and search systems can accurately understand and represent your brand. This is the technical work Google describes as "clear technical structure."
- Automation and workflows. Processes and integrations that keep content accurate and current over time, built on top of your existing CMS, CRM and other platforms rather than replacing them. This is the freshness layer Google's guide quietly depends on.
- Monitoring and iteration. Continued measurement of visibility across AI assistants, so priorities can be adjusted as competitors, prompts and models evolve.
Every brand is different, so the scope adapts to what makes sense for your business. The work runs alongside your existing marketing, content and technical teams rather than displacing them.
What to take from the guide
Google's guide is a useful reset. It strips out a layer of speculative tactics that have been sold to brands over the past two years, and it returns the conversation to the foundations: unique, expert-led content, a sound technical layer, and consistent execution against both.
It is also a more demanding standard than it first appears. "Create non-commodity content" is easy to say and difficult to deliver consistently. "Maintain a clear technical structure" assumes engineering capacity that many marketing-led teams do not have. The guide rewards organisations that have both capabilities operating against the same brief, and quietly disadvantages those that do not.
For brands looking to act on the guide, the most efficient starting point is a clear view of where you currently stand. The PQ Impact AI Brand Visibility Report provides that view, with a prioritised plan you can implement yourself or with our team. Where the work spans both strategic marketing and custom software development, our integrated capability allows you to deliver against the guide without coordinating multiple suppliers.
Get in touch to discuss the report and how it would apply to your business.
