# Peter Rota on What Actually Changes in AI Search Is GEO just SEO with a new name? Peter Rota (HUB International) on falling clicks, structuring content for AI answers and the first 3 moves on a new site. Source: https://iamnoorkhan.com/podcast/peter-rota-what-actually-changes-in-ai-search Author: Noor Khan (https://iamnoorkhan.com/) Video: https://www.youtube.com/watch?v=e-mDYhwf-IU Guest: Peter Rota, Senior Technical SEO Manager at HUB International ## Summary Everyone is arguing about whether AEO and GEO are a genuinely new discipline or SEO with a new label. Peter Rota's answer, after running technical SEO for a billion-dollar enterprise and auditing sites specifically for AI visibility: **a strong SEO foundation gets you roughly 80% of the way there**. The last 20% asks questions SEOs didn't used to ask, such as how customers see a brand compared with how LLMs describe it, and whether support docs and other "unimportant" pages answer what people actually ask. We cover why impressions are rising while clicks fall (and how to report that without it looking like failure), why AI systems pull chunks instead of reading top to bottom, how Peter structures content so the answer sits at the top, the first three things he'd do on a brand-new site, and why consistency has mattered more to his career than being the smartest SEO in the room. ## Key takeaways - Good SEO gets you about 80% of the way to AI visibility. Peter got a lead from ChatGPT without optimizing for it, from links and LinkedIn mentions alone. - Falling clicks with rising impressions is the new normal. Report impressions as visibility, keep conversions as the north star. - GEO is still its own discipline at the edges: brand perception across customers and LLMs, and pages SEO undervalues, like support docs. - Structure for extraction: inverted pyramid, key takeaways at the top, semantic HTML (lists, tables, headings), strong internal links. - On a new site: fundamentals and schema first, then social profiles and third-party mentions, then crawl logs and Search Console. - Rankings over citations, mentions over citations. In 12 months, rankings will still drive more traffic. ## Chapters - 0:05 Introduction: Noor Khan and Peter Rota - 0:50 Brand, storytelling and copywriting - 1:41 A lead from ChatGPT without optimizing for it - 2:56 Impressions up, clicks down: what to report instead - 4:38 AEO and GEO: new discipline or SEO rebrand? - 6:16 AI reads chunks: structuring content for extraction - 8:05 Key takeaways first: a practical structure - 9:23 The first three things to do on a brand-new site - 10:59 The skill outside SEO that matters most - 11:59 Rapid fire: citations or rankings, and where to start - 13:36 Wrap-up and Peter's SEO audit template ## Highlights ### [1:41] You can show up in ChatGPT without "doing GEO" Peter's first ChatGPT lead came to a site he had barely touched. No AI-specific optimization, just a sound SEO foundation and links earned from his LinkedIn presence. His point: people over-obsess about exotic GEO tactics when consistent messaging and solid fundamentals do most of the work. **What to do with it:** before buying a "GEO package", check the basics are done: crawlable pages, clear positioning, schema, and a brand people actually mention and link to. ### [2:56] Impressions up, clicks down is not a crisis Peter's team has watched this pattern for about two years, since AI Overviews launched. They still report clicks, but they also present impressions to leadership as **visibility**, and keep **conversions** as the number that matters. Their insurance glossary pages lost clicks while gaining impressions, and they still earn their place. **What to do with it:** add impressions and branded search to your monthly report, so a drop in clicks is read in context, not as failure. ### [4:38] GEO is mostly SEO, but not entirely Peter started out believing "good SEO is good GEO". Running a GEO-focused audit changed his mind on the edges: brand perception (what customers think versus what LLMs say) and pages SEO rarely prioritizes, like **support docs**, turned out to matter. Some sites rank well in Google and still don't appear in LLM answers. **What to do with it:** ask ChatGPT, Perplexity and Gemini how they describe your brand and your competitors. Where the answer is wrong or thin, that's your GEO work. ### [6:16] Write for extraction: answer first AI systems lift chunks of a page rather than reading it top to bottom. Peter applies the **inverted pyramid**: the most valuable information goes at the top, with no fluffy intro, followed by semantic HTML that machines parse easily: bullet lists, tables and a clean heading structure. His team is also testing serving **markdown versions of pages to AI agents**. **What to do with it:** rewrite the first two sentences under each heading on your top pages so they directly answer the heading. See [answer engine optimization](/glossary/aeo) and [passage retrieval](/glossary/passage-retrieval). ### [8:05] Put key takeaways at the top Peter's practical habit: a short **key takeaways** list at the top of every article. LLMs get the core facts fast, and readers who want depth keep going. Add strong internal linking and good headers, and you have content that works for both audiences. Every article on this site, and this page, uses that structure. ### [9:23] Three moves for a brand-new site 1. **Fundamentals:** optimized core pages (services, about, contact), internal linking, JSON-LD schema, and crawl access for both Google and LLMs. 2. **Third-party presence:** social profiles and any credible place that talks about you. LLMs pull heavily from third-party sources. 3. **Watch the crawlers:** crawl logs and Google Search Console, to see how bots actually move through your site and what's broken. ### [10:59] Consistency beats brilliance Asked about the skill outside SEO that mattered most, Peter's answer wasn't technical. He's never claimed to be the smartest SEO, but he shows up daily, keeps learning and stays consistent. In SEO, results come from repeating the work, not from a burst of effort followed by giving up. ### [11:59] Rapid fire: rankings, mentions, and who to trust Twelve months out, Peter still bets on **rankings over citations**, because rankings drive more traffic, and he rates **brand mentions** above citations too. His advice for anyone overwhelmed: build the fundamentals, then experiment, and before acting on advice from LinkedIn or Facebook, check whether the person is trying to help you or sell you a course. ## Resources mentioned - [Peter Rota's SEO audit template (free)](https://docs.google.com/spreadsheets/d/1aUKK2QL3qgZEqsx0iU_OOA-ZwXq-0sVw2MOTOuXhrVg): The Google Sheets checklist Peter uses for SEO audits. Make a copy and run it on your own site. - [Peter Rota on LinkedIn](https://www.linkedin.com/in/peterrota): Daily posts on technical SEO, AI search and building a brand as an SEO. - [What is GEO (generative engine optimization)?](https://iamnoorkhan.com/glossary/geo): A plain-English guide to getting cited and recommended by AI assistants, and how it relates to SEO. - [What is AEO (answer engine optimization)?](https://iamnoorkhan.com/glossary/aeo): How to structure pages so the answer is easy to extract: the inverted-pyramid approach Peter describes. - [Google Search Central: AI features and your website](https://developers.google.com/search/docs/appearance/ai-features): Google's own guidance on how AI Overviews and AI Mode use web content.