You ship pages yet answer engines' names someone else. The pain is missed mentions and shrinking clicks, while LLMs favor sources you do not control. This September opener sets the stage for faster audits, cleaner entities, and smarter distribution across AI surfaces.
No spoilers, just signals and fixes. Now let’s see what the first day of September has brought you.
Quick Hits: Signals That Move Mentions Right Now
The shelf moved inside answers and into places your CMS does not track by default. Here are the shifts that change how you get named and credited this week. Let’s take a closer look.
- Beauty pivots to GEO:- BoF details how beauty brands are feeding first-party product data to LLMs and testing ChatGPT Shopping to win citations. (The Business of Fashion)
- GEO basics for law firms:- Six tips stress EEAT-led writing, clear modules, schema, fresh updates, sitemaps, and mobile to win mentions in ChatGPT, Copilot, and Gemini. (JD Supra)
- Perplexity:- Nikkei and Asahi sue in Tokyo seeking an injunction and ¥2.2B each over alleged copying, robot directives ignored, and errors attributed to their articles. (japantimes.co.jp)
- ChatGPT referral slump:- Referral traffic from ChatGPT to sites fell 52% month over month after OpenAI reweighted links, reshaping SEO and GEO plans. (eMarketer)
- Fractured search rules:- New polling shows 66% think AI will replace traditional search within five years, and 82% find AI search more helpful than classic results. (Searchengineland)
- Content up, budgets flat:- 10Fold reports 89% increasing output and 46% producing 3–5x more content year over year, while 76% saw only 1–10% budget lifts, with GEO now part of the agency scope. (The Wise Marketer)
- Paid in AI answers:- How to show ads inside Google AI Overviews or Copilot via Performance Max and Microsoft placements, including targeting and reporting notes. (JumpFly)
- New voice models:- OpenAI unveils GPT-Realtime with Realtime API GA and Microsoft debuts MAI-Voice-1 for Copilot, raising the bar for spoken answers that cite sources. (SiliconANGLE)
- Gemini image prompts:- Official guide says describe full scenes, use templates and semantic negatives to improve Gemini 2.5 Flash Image outputs for brand assets. (Google Developers Blog)
Also read: The GEO Reality Check: Find Out What AI Really Says About You
Deep Dives: 5 Shifts You Need to React to This Week
These stories change the math for GEO across citations, ads, and entity quality. The angle is simple: earn mentions where models look first. Let’s take a closer look.
Beauty brands feed LLMs, not just Google
Business of Fashion reports 700 million weekly ChatGPT users, broad intent growth, and heavy citation weight on Reddit and Wikipedia. Estée Lauder pilots an OpenAI connector so ChatGPT Shopping fetches verified catalog fields, not scraped pages.
The GEO move is to build clean product entities and publish claim-level proofs that third parties can cite. Refresh attributes shoppers ask for, including shades, sizes, and ingredients.
Measure assisted revenue from answer mentions, not only the last click. This shift rewards clarity, provenance, and distribution beyond your domain and platforms. (Business of Fashion)
ChatGPT’s referral drop rewrites GEO priorities
eMarketer reports a 52% drop in publisher referrals from ChatGPT after source weighting changes. Visibility concentrates in high-trust domains like Wikipedia and Reddit that already carry authority with users and models.
The GEO response is to win those shelves and protect your brand facts where models shop for sources. Build research-backed explainers with named authors, dates, and claim citations. Seed concise summaries on pages that rank or earn community votes.
Stand up a weekly citation share log for priority queries. Track which third-party pages name you, then reinforce those placements with evidence, clearer entities, and a short answer block on the page. (eMarketer)
Ads are entering AI answers
JumpFly shows ads already appearing inside Google AI Overviews, AI Mode, and Microsoft Copilot. Eligibility flows through existing campaign types, smart bidding, and broad intent coverage, with rich assets improving placement and reach.
This matters because a reliable mention can sit next to a competitor’s paid unit on the same answer. The GEO move is to pair answer-ready pages with feed health and clean product data.
Tag AI intent tests, record where ads show, and watch overlap with pages that earn citations. Use UTMs and annotations, then adjust budgets toward prompts where you win both the mention and the paid placement this quarter. (JumpFly)
Content triples, Budgets stall: GEO must get sharper
Brand teams are shipping far more content without matching funds. The survey shows 89 percent plan to produce more output and 46 percent create three to five times last year, while most budgets rose only up to ten percent.
That flood means answer engines cherry-pick a few sources. Win the shelf with short claim blocks, named authors, and links to proofs. Use Reddit and high-trust media for citations that models prefer.
Bring agencies in for data pages and GEO-ready briefs. Track LLM mentions weekly, so spend follows the pages that actually get named. (The Wise Marketer)
SEO rules update for fractured search
Search Engine Land frames discovery as a split stack across Google AI Overviews, Perplexity, Copilot, and classic results that still matter for navigational checks.
Treat AI answers as a shelf where a complete response needs a short claim, evidence, and a clean entity that matches graphs. Build a 40 to 80-word first paragraph that states the answer, then link to the proofs users trust.
Add FAQs that match common prompts and publish author bios. Run cross-engine spot checks by intent class and store share links for comparison. The lesson is simple. Answer first, support fast, and measure mentions as closely as you once measured rankings. Then improve pages that never earn citations. (Search Engine Land)
Also read: Beginner’s Guide to Understanding the Difference Between SEO and GEO
Numbers That Explain Where Mentions Go
You plan better when you see the trend line. These are the numbers that steer budgets toward answer shelves and away from vanity clicks. Let’s take a closer look.
- ChatGPT weekly users are at 700 million in beauty coverage. (Business of Fashion)
- ChatGPT referral traffic to publishers is down 52%. (eMarketer)
- Generative queries 37% of ChatGPT searches, transactional 6.1%.(Business of Fashion)
- Reddit is cited in about 40% of citations in beauty tests. (Business of Fashion)
- Half of professionals expect AI to replace Google for business research by 2030. (Business Wire)
Also read: A Quick Guide to Checking If Your Brand Exists in AI Answers
Brand Spotlight: Estée Lauder Feeds Answers With Verified Facts
Big brands are moving the source of truth from web pages to structured pipes that models can fetch directly. The playbook shows how to control names, claims, and comparisons inside answers. Let’s take a closer look.
- What changed? Estée Lauder Companies expanded GEO work across ChatGPT, Gemini, and Perplexity with verified product data and third-party proofs.
- Evidence: BoF cites ELC confirming GEO exploration for brands like Clinique and Origins, with Reddit and media placements in the mix. Business of Fashion
- Takeaway: Treat catalog fields, usage claims, and safety notes as structured inputs that LLMs can quote. Build clean product entities, publish claim links, and track model mentions by query class.
Also read: Discover the 22 Best AI SEO Tools for LLM Rankings in 2025
Tactic of the Week: Answer-first Page Shape That LLMs Quote
You win mentions by making it easy for models to lift a correct snippet that includes a clear claim and a source. Build pages with a 40 to 80-word direct answer at the top, then H2 blocks for related questions, plus product, FAQ, and author schema. Keep URLs short, prune heavy scripts, and refresh often. This shape mirrors how answers chunk and cite. (JD Supra)
Also read: How to Write Your First AI-Optimized Article: A Step-by-Step Guide
Move List For The First Week Of September
These steps help you claim answer shelves you do not yet track. The focus is speed, evidence, and clear entities. Let’s take a closer look.
Claim the answer shelf:- Write a short, quotable answer atop your ten money pages, then add proofs and links models can lift. Log mentions in ChatGPT, Gemini, and Perplexity before and after. (Business of Fashion)
Win the sources models favor:- Pitch one research-backed placement to outlets that Wikipedia and Reddit editors cite. Run a Reddit expert reply series for your category and measure citation share. (eMarketer)
Cover paid inside answers:- Turn on campaign types that qualify for AI Overviews and Copilot, fix feed quality, and tag AI intent tests. Map where paid sits next to your organic mention and shift budget to those prompts. (JumpFly)
In short, AI answers drive discovery. Clicks shrink as models favor trusted sources. Expert takeaway: ship answer first pages, feed verified product data, win citations on Reddit and media, log mentions, and pair GEO with ads in AI Overviews and Copilot.
To know more about GEO, AEO, and AI SEO, read our latest articles now!