If ChatGPT cites your brand but still recommends a competitor, you have a visibility problem—just not the one most teams assume. Semrush’s new AI search playbook breaks visibility into four layers so marketers can stop rewriting content and start fixing the actual gap.
Why most AI visibility fixes miss
Generating an AI answer happens in several steps: retrieving sources, evaluating claims, synthesizing a response, and deciding what to cite. A break at any step can push your brand out or describe it inaccurately.
Most teams default to content work: rewriting pages, restructuring information, or publishing new pieces. That works only when the content itself is the weak point. The first job is to identify the failing layer.
The four layers, in order of triage
- Discoverability: AI systems can’t reliably access your information.
- Clarity: AI reaches your content but describes your brand inconsistently.
- Authority: AI describes you correctly but rarely includes you in comparisons or shortlists.
- Trust: AI considers you but rarely recommends you, or adds recurring caveats.
Start with discoverability to rule out access problems. Then use the patterns you see in AI answers to decide whether clarity, authority, trust, or a mix needs work.
Discoverability is the boring first step
Priority pages need to be crawlable and indexable. AI crawlers shouldn’t be blocked by robots.txt, CDN, firewall, WAF, or bot-management rules. Key content should be readable text, not hidden behind JavaScript. For ChatGPT Search, publishers should allow OAI-SearchBot; for Google AI Overviews, a page must be eligible to appear in Google Search with a snippet.
Clarity: AI can cite you and still get you wrong
Semrush’s Geertrui Laleman explained how the company fixed its own AI description:
“AI answers used to describe us as a traditional SEO tool with AI features added on. The mentions were there, they were just describing the wrong company. We didn’t fix that by publishing more on our own site. We fixed it by changing how we were described on the third-party pages LLMs cite.”
Align owned pages with a verified source of truth first, then correct outdated or conflicting third-party descriptions that AI systems actually cite.
Authority and trust: the recommendation gap
Being cited is not the same as being recommended. In a study of 3,981 brand appearances across ChatGPT, Google AI Overviews, Gemini, and Google AI Mode, 62% were citations where the answer never named the brand. ChatGPT cited brands in 87% of its appearances but named them in just 20.7%.
To build authority, publish evidence that is hard to reproduce: proprietary data, first-hand tests, original frameworks, or primary documentation. Semrush’s content study found AI-cited pages scored higher on clear summarization (+32.83%), E-E-A-T signals (+30.64%), Q&A formatting (+25.45%), section structure (+22.91%), and visible structured data like lists and tables (+21.60%).
Trust depends on independent evidence. If reviews and third-party coverage conflict with your messaging, AI answers may add caveats or point buyers elsewhere.
Measure patterns, not single answers
Don’t treat one AI answer as a fixed ranking. Track a small set of buyer-decision prompts across AI platforms weekly. Semrush watched ChatGPT citing Reddit drop from close to 60% of responses in early August 2025 to around 10% by mid-September, before stabilizing alongside Wikipedia as one of ChatGPT’s two most-cited domains.
Build the prompt set from questions buyers already ask: sales-call questions, support conversations, on-site search, Search Console queries, and review-site language. Group by persona and journey stage—”best CRM” and “top CRM software” are the same decision, not separate prompts.
The practical order: confirm access, align brand clarity, build distinctive evidence, then monitor trust signals. Fix the layer your data points to, not the tactic you know best.
Source: Semrush Blog
