Improving your ranking on AI engines means getting recommended inside the answer when ChatGPT, Claude, Perplexity, Gemini, or Microsoft Copilot respond to a buyer’s question. It is not a single lever. In every audit we run for a prospect, the score rolls up into six measurable dimensions, and the fastest progress comes from fixing whichever of those scores below 60 first. See the implementation guide for how each dimension is fixed in practice.
The six dimensions we score
When AirPulse runs an audit, the composite score is weighted as follows:
- AI Citability & Visibility — 25%
- Brand Authority Signals — 20%
- Content Quality & E-E-A-T — 20%
- Technical Foundations — 15%
- Structured Data — 10%
- Platform Optimization — 10%
The weights reflect what we see actually moving the needle. Citability and brand authority dominate because, in our pipeline, those two are the strongest signals that an AI engine will not just find a page but choose to quote it. We have run this audit against eight B2B brands so far, spanning categories from developer tooling and fintech to proptech, sales and HR tech, event software, and cloud communications.
The three-layer fix: Measure, Act, Maintain
The product itself is organised around three layers, and the playbook follows the same shape.
Layer 1 — Measure
Track how AI engines mention, describe, and recommend your brand across every model, every prompt, every day. The visibility metric we use is the share of a tracked query set in which the brand appears — mentioned queries divided by total queries. The five engines we cover are ChatGPT, Claude, Perplexity, Gemini, and Copilot. Which subset is included depends on the plan: Starter and Growth cover ChatGPT and Perplexity; Pro and Enterprise add Google AI and Claude.
Layer 2 — Act
Turn every visibility gap into a prioritised action. The audit produces a ranked list — critical, high, medium, low — and a content brief for the gaps that matter. The bias is toward closing gaps with content you can ship this week, not strategy decks.
Layer 3 — Maintain
Keep feeding AI engines the canonical version of your brand through the Brand Hub — positioning, ICP, differentiators, product context — so that when they synthesise an answer, they synthesise yours. The Brand Hub is the durable artefact; the audit is the snapshot.
Where mentions actually come from
AI-driven recommendation pulls from a wider source set than classic SEO. Our mention detection runs an exact pass first, then a fuzzy pass at a Levenshtein similarity threshold of 0.8, with a ±100-character context window and a per-mention confidence between 0.0 and 1.0. In practice, this means an AirPulse score reflects mentions on the AI engine’s answer, the third-party pages it pulled from, and the named-entity matches near the brand. Reddit threads, vendor comparisons, and analyst pages move the number; a single owned blog rarely does.
What to do this week
- Run a baseline audit across the five engines on ten prompts your buyers actually use. If you don’t have a tool, AirPulse Starter ($89/month) covers ChatGPT and Perplexity for 25 prompts.
- For every prompt where a competitor is recommended and you are not, capture the cited URLs. Those URLs are your working surface area.
- Pick the dimension with the lowest score under 60 and fix it first. Structured data is usually the cheapest week-one win; brand authority is usually the slowest.
- Treat AI engine output as a measurement, not a strategy. Re-run the same prompt set 30 days later and compare deltas, not absolute scores.
What this article is not
This is not a list of tools to buy. It is not a promise that AI rankings move in a week. Citation rate is the leading indicator; conversion is the lagging one. The two only correlate after the buyer journey settles, and that takes a quarter to read.
FAQ
Which engine matters most?
Depends on your buyer. For B2B SaaS, ChatGPT and Perplexity carry the most weight in our audits; for consumer queries, Google AI Overviews dominates the click-through path.
How quickly will rankings move?
Structural changes — schema, on-page — can register inside one audit cycle. Authority changes — third-party mentions, reviews, citation pages — move on a multi-month horizon. The Maintain layer is what compounds.
Is there a checklist?
Yes — the six-dimension scorecard above. Each dimension expands into specific subchecks inside the audit report.
— — —
Related AirPulse guides
AEO vs Traditional SEO: What Actually Changes Under the Hood
AI Engine Brand Visibility: The Implementation Guide
How this was made
How this was made: the draft of this article was generated from AirPulse, our own AI engine optimisation platform, then reviewed and edited by the named author. Product claims about AirPulse are sourced from internal documentation; external claims link to their primary source.
Sources
- AirPulse audit methodology and dimension weights — internal documentation
- AirPulse pricing — https://airpulse.ai/pricing
- Companion: ‘The New ROI: Top 3 KPIs for the Pre-Click Era of AI Search’ — https://airpulse.ai/insights/the-new-roi-top-3-kpis-for-the-pre-click-era-of-ai-search
- Companion: ‘Where AI Gets Its Sources’ — https://airpulse.ai/insights/where-ai-gets-its-sources-a-practical-guide-to-chatgpt-perplexity-google-ai-overviews
- Google Search Quality Rater Guidelines (Sept 2025) — policy baseline — https://guidelines.raterhub.com/searchqualityevaluatorguidelines.pdf