Every month there is a new acronym to panic about. AEO, GEO, LLMO, take your pick. Strip the jargon and the same unglamorous SEO that always worked is what gets you quoted in AI answers, just pointed at a new set of boxes. Here is what actually earned citations this week, and the exact moves you can copy before your competitors do.
🧭Do you really need a separate AI-search playbook?
Google just said the quiet part out loud: AI search is still SEO. Its own optimization guide explains that AI features run on RAG grounded in ordinary search ranking, so the page that ranks is the page that gets cited. No llms.txt, no content chunking, no AI-only rewrites, no schema obsession required. Google spelled it out plainly. The move: stop rebuilding your site for robots and go write the single clearest answer to a question your buyers actually ask.
So what is "AI search optimization," really? It is the boring checklist a nine-year practitioner keeps coming back to, eleven things, none of them new:
Reviews, earned media, and brand mentions
Backlinks and owned content
YouTube presence
Rankings and topical authority
Entity consistency across the web
Comparison pages
Community threads (Reddit, forums, Slack groups)
All of it is SEO with a wider surface area. Run down the list and be honest about how many you are actually doing. That gap is your roadmap.
📡How do you tell if AI search is actually working?
Your buyers now fact-check you inside ChatGPT. One SEO with nine years in the game says clients stopped trusting ranking graphs and started sending messages like "I asked ChatGPT and it didn't mention us." He now reports on AI-mention visibility because that is how buyers check. Your ranking report is no longer the scoreboard your customer is looking at, so start checking your own name the way they do.
And you can finally measure the traffic. GA4 quietly added an "AI Assistant" channel that auto-tags visits from ChatGPT, Gemini, and Claude instead of dumping them into direct. Two-minute job this week: open GA4, filter for the ai-assistant medium, and see what is already trickling in. You cannot grow a number you were never looking at.
But keep your head on: citations are not clicks. The field data is messier than the hype. YouTube is climbing in AI Overviews, Reddit's citation share swings from 60% to 10% to 70% of ChatGPT answers depending on the week, Wikipedia still owns plain facts, and publishers who do get cited see under 1% referral traffic. Being mentioned builds recall far more than it builds sessions. Chase the mention for the brand lift, not for a click that mostly is not coming.
🔗Where do your first 100 backlinks actually come from?
Your first 100 backlinks, as a checklist you can start today. None of it is clever, all of it compounds:
📁 Get on every relevant product directory
📝 Add yourself to listicles where competitors already sit
🤝 Offer case studies to the tools you already pay for
✉️ Guest-post in niche newsletters
📊 Publish one piece of original data worth citing
🔧 Reclaim broken links on outdated resource pages
The full playbook is here when you want the detail.
Then work the 300+ list. After Product Hunt, Hacker News, and Reddit, a founder compiled 300+ startup, AI-tool, and SaaS directories into one public sheet, and every submission doubles as an early backlink to lift a new domain's rating. Block two hours and go top-down. This is exactly the grind BacklinkBot automates, but nothing is stopping you from starting by hand tonight.
🛠️What's winning right now without a single new blog post?
A £45K month from fixing plumbing, not publishing. An agency's best month ever came from technical SEO alone: 31,000+ stray HTTP URLs cleared, 273 broken or missing canonicals fixed, security headers added, robots.txt unblocked from 223 PDF datasheets, and zombie categories killed off. Zero new content, zero new links. The full teardown is a useful gut-check: before you spend on content, confirm Google can even crawl what you already have.
The programmatic SEO math, told honestly. Pick a "best X for Y" or "service in city" pattern, scrape structured data that gives each page genuine value, build one template, generate copy that is actually unique per page, then ship a 100-page MVP and watch indexation before scaling to thousands. Skip that "unique value per page" step and all you have built is index bloat. Read the end-to-end breakdown before you generate a single page.
✍️Should you put your real name on your content?
Does putting your real name on it actually move E-E-A-T? A healthcare site with 150+ articles under a pen name, backed by a genuine credential, pulls steady Bing traffic but stays stuck in Google's sandbox. The open question: in a sensitive niche, when does attaching your real license beat a credentialed-but-anonymous byline? Nobody has a clean answer yet, but if you write in health, finance, or law, it is a test worth running on your own site this quarter.
⭐ THE ONE TO STEAL
The HubSpot numbers that end the argument. HubSpot reports 15x growth in AI-search traffic in a year, AI-search visitors converting 5x and in some cases up to 13x higher than Google search, and 60% of AI citations coming from outside the top 20 Google results. Read that last one twice. The incumbents winning Google are not automatically winning AI search, which is precisely the opening a smaller, sharper SaaS needs. The play to steal: find the exact questions your buyers ask an AI, and become the clearest answer on the web to each one. See the data.
BacklinkBot clears that 300+ directory list for you, so the only work left is the part that actually moves the needle: earning one genuinely citable data point. That is the trade worth making this week.
What's BacklinkBot? BacklinkBot gets your SaaS listed across 300+ directories, niche listicles, and review sites, building the backlinks and citations that Google and AI search both reward, without the manual grind.
Thanks for reading 👋
— Piyush, BacklinkBot AI
BacklinkBot · SEO and AI-search growth for SaaS founders