Boni Satani

@yuvrajv5 @OpenAI Super, thanks for the feedback

Pencil - doesn't go with the brand guidelines strictly; it sometimes changes the context Claude - consumes a lot of tokens for image generation ChatGPT - high-quality images, no token issues, this new model actually remembers the brand guidelines properly and creates images based on that

Introducing ChatGPT Images 2.0 A state-of-the-art image model that can take on complex visual tasks and produce precise, immediately usable visuals, with sharper editing, richer layouts, and thinking-level intelligence. Video made with ChatGPT Images

@yuvrajv5 @OpenAI Isn’t Pencil + Claude good at this or ChatGPT is better?

This actually solved a long-pending use case for me. I wanted to design inner blog images strictly based on brand guidelines. Earlier versions often changed the content or ignored the guidelines. But this version worked really well. Designed multiple images yesterday without those issues.

Google is indexing LESS content year by year, their threshold for indexing content has gone up significantly. Why? 1. It costs money to store and process website data 2. It doesn't need 10,000 URLS with the same copy 3. It makes no sense to index something you are unlikely to serve (quality issues, trust issues etc) So, I put together a FULL guide on using Google's URL inspection tool. If you aren't using it but you are in SEO, you should be using it. 🡲 Periodically check power pages on a domain to check index state 🡲 Check indexing states 🡲 Address and deal with anomalies 🡲 Request indexing for pages where issues have been fixed 🡲 Check rendered outputs (screenshot / HTML) 🡲 Check canonical profile and so much more. I've done a video which walks you through Google's URL inspection! Enjoy seo-stack.io/blog/the-compl…

Very informative slide via Google's Danny Sullivan explaining the difference between "Commodity" vs "Non-Commodity" content Google prefers the later IMO, lots of evidence this is spot-on, not where Google is going in the future, but where it already is now via @ChouinardJC

@lilyraynyc Wow, didn't saw that yet...who launched it?

So now multiple new GEO tools are formally, publicly launching paid linkbuilding networks? What can go wrong? 🫠🫠🫠

This AI SEO Playbook is quite literally... out of this world, freely available, and by former Head of SEO at NASA, Rodrigo Stockebrand @spanglishseo 🚀👽🛸🔬 Rodrigo shares 40 advanced plays to maximise On Running’s visibility across ChatGPT, Perplexity, Claude, Google AI and Gemini. The most comprehensive GEO/AEO resource I’ve seen to date - and applicable to any brand serious about AI search. 🤓 TL;DR: 14 content plays, 13 technical plays, 13 earned plays. Each with tactical steps, difficulty ratings, timeframes, KPIs and a “where to start in 30 days” section. This isn’t a blog post - it’s a full operating manual. What this incredible playbook covers: 🧠 Content plays - “The Citation Play” turns proprietary data into the citation AI reaches for first. “The Head-to-Head Play” publishes brand vs competitor comparisons so AI cites you instead of RunRepeat. “The Technical Lexicon” makes your brand the definition source for your own proprietary terms. “The Buyer’s Decision Tree” maps every runner-to-shoe decision path into structured, AI-extractable content. ⚙️ Technical plays - “The Schema Slam” deploys full Product + Offer + Variant + Review schema across every PDP. “The AI Crawler Allowlist” explicitly welcomes GPTBot, ClaudeBot, PerplexityBot with llms.txt and llms-full.txt. “The Server-Side Render Audit” catches content invisible to AI crawlers. “The Disambiguation Play” stops AI confusing similarly named products. 📡 Earned plays - “The Reddit Embassy” runs transparent AMAs with engineers - no astroturfing. “The YouTube Review Strategy” seeds structured review access to running creators whose transcripts feed LLMs directly. “The Podcast Citation Play” turns interview transcripts into perennial citation sources. “The Trust Shield” turns Runner’s World, Wirecutter and RunRepeat coverage into a trust network pointing back to the brand. 🔬 Things I’ve tested and can vouch for:Schema at PDP level making a measurable difference to AI extraction accuracy. FAQ blocks being the single most extractable content format (Play 11 - “The FAQ Goldmine” nails this). Comparison content earning disproportionate citation share. And SSR gaps silently killing AI visibility - most brands don’t even know their content is invisible to crawlers that don’t execute JavaScript. 🪙 My 2-pence: 💡 This is the real deal. Not a theory deck. It’s calibrated to a specific brand (On Running) with specific products, athletes and competitor dynamics - which is exactly what makes it incredibly useful rather than generic. 💡 The earned plays section alone is worth the read. Most AI SEO content stops at schema and content structure. This covers Reddit, YouTube, podcasts, press, retailer co-marketing - the full signal network AI triangulates from. 💡 Not plug-and-play. You still need to calibrate every play to your brand, your data and your competitors. But the framework is the best starting point I’ve seen. Rodrigo, take a bow👏🏼🏆 Rodrigo's AI Search Playbook play-view-sync.base44.app His announcement post: linkedin.com/posts/stockebr… #ai #seo #google #llms

@krinal Seeing Ahref making that move, offering workflows directly thought their Agentic Platform.

I can see a version of this playing out for enterprise SEO tools. No one wants to be locked in workflows they don’t use. Give them the outputs and get out of their way. DataForSEO has a headstart because their model was primarily based on API, can they/will they evolve fast?

Instead of watching Netflix tonight, watch this 2-hour Stanford lecture. You’ll learn more about how ChatGPT, Claude, and other LLMs are built than most people at top AI companies learn in years.

Introducing Claude Opus 4.7, our most capable Opus model yet. It handles long-running tasks with more rigor, follows instructions more precisely, and verifies its own outputs before reporting back. You can hand off your hardest work with less supervision.

@PeterMindenhall @timsoulo Yeah, and I've been seeing this a lot! A lot of GP/Link Farm sites seems to have figured this out at scale.

@BoniSatani @timsoulo SV data is generally crap at best anyway (always has been) but that does appear to be overinflated more than usual.

Hey @timsoulo is it possible to inflate the keyword search volume? The below site ranks for a lot of kws having high search volume. Feels something is off either search volumes are inflated or this website genuinely is ranking for keywords like 'g15tools com contact' (5300/mo)😮

We've redesigned Claude Code on desktop. You can now run multiple Claude sessions side by side from one window, with a new sidebar to manage them all.

Do you ever notice your rankings go up after a surge of unrelated traffic? Many reasons why this could happen, but in the Google antitrust trial, they found Popularity is based on both links + "Chrome visit data" Importantly, this data is separate from Google search results 1/3

Our 2026 SEO report template helps you track and communicate your website’s organic traffic, AI visibility, backlinks, and keyword rankings. Get it for free here! social.semrush.com/4tqzUrH.

Optimizing your site for SEO & for LLMS? It's CRUCIAL that user-agents / bots are actually able to see the rendered output. Not all tech stacks are suitable out of the box - JS tech stacks are notorious for: 1. Not being easy to render even for Googlebot which can render JS but isn't overly efficient at it 2. Not being suitable for LLMs to crawl / ingest content Want to know if your site is? 1. Visit a page on your website, hit F12 and open DevTools 2. Press CTRL + SHIFT + P 3. Type Javascript and click Disable Javascript 4. Click NETWORK in the header menu of devtools 5. Ensure Preserve Log and Disable Cache are set 6. Now on your website URL append ?test and hit enter 7. On your network tab, click DOC (shown with red box in image) 8. Where it says NAME you should see ?test, click it 9. Click on RESPONSE tab (shown with red box in image) This will show you the OUTPUT of what an LLM would see. Google CAN render Javascript but, not all Javascript tech stacks are the same - if you want to see the RAW TEXT OUTPUT, do the following: In the bottom half of DevTools you will see menu items "Console, Issues, Coverage" etc. click on CONSOLE (selected by default). Click the CLEAR icon (circle with line through it) Copy and paste this command into console and hit ENTER: const text = document.body.innerText; for (let i = 0; i < text.length; i += 1000) { console.log(text.slice(i, i + 1000)); } This will show you the RAW, NON JS output of your URL. THIS is the content that LLMS will see, and potentially Google/other search engines if there are any issues with SSR. IF YOU ARE USING JS based tech stacks you SHOULD consider SSG (static site generation) or pre-render. Effectively you want to MAXIMISE the output of RAW TEXT/HTML in the dom post JS processing before that document output is served. The amount of SITE OWNERS who have these fancy tech stacks are often oblivious to the risks/issues with using JS for a web app / site framework. Here are some solutions based on common platforms: React Next.js – [SSG + SSR + ISR] Most popular for React, supports hybrid static rendering and server-side rendering Gatsby Fully static, React-based. Uses GraphQL for data and pre-renders everything at build time Astro Partial hydration. You can use React components but ship less JS by default Vue Nuxt Offers static site generation (nuxt generate) as well as server-side rendering VitePress Lightweight Vue-based SSG, ideal for docs and content-driven sites VuePress Vue-powered static site generator, originally designed for documentation Svelte SvelteKit Can be configured for full SSG, SSR, or hybrid deployment. Excellent for fast, SEO-friendly output Elder.js Performance-focused SSG for Svelte apps (less common now but still useful) Angular Scully The go-to SSG for Angular. Pre-renders Angular apps into static HTML Analog.js Newer project aiming to bring Next.js-like DX to Angular with SSR and SSG #seo #seo2025 #aiseo

While Yoast SEO and Rank Math claim llms.txt helps AI bots find your content, we checked. 6 months. 57 AI bots. 180k AI hits. 0 llms.txt requests. Full data and methodology: #issuecomment-4226935622" target="_blank" rel="nofollow noopener">github.com/sybrew/the-seo…

Update: the fake Google Search Console impressions have started to come crashing down. I'm mostly seeing a clear decline within Image Search for the moment (the original surface I was alerted by), with other surfaces still to follow. Here is a timeline of the key events: 1. March 30th, 2026 I shared that there was a clear issue with impressions (and a smaller portion of clicks) that appeared to be unrelated to the disabling of &num=100. 2. April 3rd, 2026 Google recognises that there has been a logging error that has prevented them from accurately reporting on impressions. It has apparently been an ongoing issue since May 13th, 2025. Will take a few weeks to resolve. 3. April 9th, 2026 The first big decline in impressions happens when using the Images search appearance filter in GSC across many properties I have access to (not just US-based data). If you're also seeing a decline in impressions, I would suggest adding an annotation to your GSC dashboard, or within your 3rd party GSC data warehousing tools such as @seogets and @seo_stack. In the example shown, Image Search impressions dropped by around 200K p/day. The difficult aspect here is that Google just finished rolling out the March 2026 Core Update on April 8th – which doesn't appear to be a related impact here. Never a dull day in SEO.

The irony is that this stuff only makes real SEOs more employable. As I often say, “call me in 6 months.”