Key Performance Indicators

Visibility score

Definition

How frequently and how prominently your brand appears in AI search results.

What this means

The visibility score blends two things: how often your brand appears across AI answers, and how prominently it is named when it does. Because every answer counts, including the ones that never mention you, appearing more often lifts the score, and being recommended near the top lifts it more than being listed last.

How your visibility score is calculated

LunaLift runs your tracked prompts across your selected AI search engines and reads each answer to find whether your brand appears and where it ranks among the brands named. Mentions can come from your own website or from trusted third-party sources. Prompts that explicitly name your company are excluded, so the score reflects genuine discoverability.

Each answer earns a score based on where your brand ranks in it:

Named first scores the maximum for that answer.
Each position lower scores progressively less, and not being named scores zero.

Those per-answer scores are averaged across every response and shown as a percentage, so a brand named first in every answer scores 100%. The per-answer score is:

(1 + brands named - your rank) ÷ brands named

Results can also be segmented, allowing you to view visibility by product, service, location, audience, market, AI search engine, or any other tracked dimension.

Why this matters

As AI becomes a primary discovery channel, being named prominently in AI-generated answers is increasingly important. Your visibility score shows how readily potential customers encounter your brand during AI-powered research, helping you measure discoverability, spot gaps in your content strategy, and track improvements over time.

Sentiment score

Definition

The tone and framing AI uses when mentioning your brand.

What this means

The Sentiment Score measures how AI systems describe your brand when it appears in responses. Mentions are classified as:

Positive: recommended or favourably described.
Neutral: mentioned factually without judgement.
Negative: criticised, discouraged, or positioned as a poor fit.

Rather than analysing the overall response, LunaLift evaluates the language used specifically around your brand, including recommendations, comparisons, warnings, qualifiers, and contextual framing.

How your sentiment score is calculated

LunaLift analyses every brand mention across your tracked prompts and AI search engines, identifying sentiment signals associated with your business. Individual results are then aggregated to produce an overall Sentiment Score, with the ability to drill down by prompt, topic, AI search engine, market, or time period.

Why this matters

Being mentioned by AI is only valuable if your brand is presented positively. Negative or unfavourable framing can reduce trust and influence purchasing decisions, even when your visibility is high. Tracking sentiment helps you identify where your messaging, content, reviews, or supporting evidence may need improvement so AI models represent your brand accurately and confidently.

Citations

Definition

A citation is a reference link within an AI-generated answer, pointing to a source the model drew on.

What this means

When ChatGPT, Gemini, Claude, or Perplexity answers, it draws on external sources, and a citation is where it credits one. The source can be yours or a third party's:

A page on your website
A blog article or research you have published
A news article or press mention
A review or directory listing
A guest article or third-party publication

Some platforms show citations as clickable links; others just name the source in the text. LunaLift tracks both where supported.

What earns a citation

A page tends to earn citations when four things line up: relevance to the question being asked, clear structure the model can parse, domain authority that marks the source as trustworthy, and sound technical optimisation so the page can be crawled and understood. These are the same signals LunaLift scores and helps you strengthen.

Why this matters

Citations are one of the strongest indicators that AI systems recognise your brand as a trusted source of information. The more frequently your content is cited, the more likely it is to influence AI-generated recommendations and drive high-intent visitors to your business.

Tracking citations helps you:

Measure your authority across AI search platforms.
Understand which pages and topics AI trusts most.
Identify opportunities to increase your visibility.
Discover when competitors are being cited instead of you.

AI-driven traffic

Definition

Website visits that AI search sent you, both the ones LunaLift can see directly and the ones it infers.

What this means

AI platforms rarely pass a referrer the way a normal link does, so most AI-driven visits never show up as such in ordinary analytics. LunaLift closes that gap with two kinds of attribution.

LunaLift measures both types of AI-driven visits

Direct: the visit arrives with an AI platform named as its source, so the referral is known rather than inferred, for example when Google Analytics reports the session source as ChatGPT or Perplexity.
Indirect: when an AI search bot crawls one of your pages and real people land on that same page within about a five-minute window, LunaLift estimates how many of those visits AI likely drove by weighing them against the page's normal baseline traffic, then follows that journey anonymously to credit any conversions.

Because the indirect side is inferred, AI-driven traffic is a well-founded approximation rather than an exact count.

Why this matters

AI is changing how people discover brands, but much of that influence is invisible in traditional analytics. Measuring AI-referred traffic helps you understand whether your AI visibility is generating real visitors, not just mentions.

Tracking AI-referred traffic helps you:

Measure the real business impact of AI search.
Connect AI visibility to website traffic and customer acquisition.
Understand how AI influences multi-step customer journeys.
Demonstrate the return on investment (ROI) of your AI search strategy.

AI-attributed conversions

A conversion on your website where AI search results played a role in the user's journey before they completed a desired action.

What this means

LunaLift tracks conversions across your website and identifies which ones originated from, or were influenced by, AI search platforms. This gives you visibility into how AI-driven traffic contributes to your business goals and highlights which pages and content are driving the most valuable outcomes.

You can define any number of custom conversion events across your site, including multiple events on the same page. These can range from lead form submissions and newsletter sign-ups to demo requests, account registrations, purchases, or any other meaningful action for your business.

Why this matters

Traditional analytics show what users converted, but not always how AI discovery influenced that decision. AI-attributed conversions help you measure the business impact of your AI visibility, identify the content that generates the highest-value traffic, and optimise pages based on revenue or lead generation rather than traffic alone. This enables you to prioritise content that delivers measurable business results and demonstrate the ROI of your AI optimisation efforts.

AEO Glossary

Accessibility = How easily people and machines can access, understand, and navigate your website. This includes factors such as alt text, semantic HTML, page structure, colour contrast, keyboard navigation, and screen reader compatibility.

Aggregate rating = Structured review data that summarises the overall rating of a product, service, or business (for example, 4.8/5 from 256 reviews). When marked up with schema, it helps search engines and AI systems understand reputation and customer sentiment.

Bot = An automated crawler or agent that visits websites. LunaLift tracks LLM bots, search engine bots, SEO bots, and social media bots to show how AI platforms and search engines interact with your site.

Buying stage = The point a user is at in their purchasing journey. LunaLift categorises prompts into stages such as Awareness, Consideration, Decision, and Post-purchase to help you understand which content supports each stage.

Content cluster = A group of related pages centred around a core topic. Content clusters help establish topical authority by connecting supporting content back to a primary pillar page. LunaLift creates content clusters as part of the recommended content strategies.

Content quality = A measure of how useful, trustworthy, and well-maintained your content is. LunaLift considers factors such as readability, promotional tone, freshness, and internal and external linking.

Content structure = A measure of how well content is organised for readers and AI systems. This includes heading hierarchy, word count, images, videos, tables, lists, question-and-answer formatting, and overall page layout.

Context documents = Documents you provide to LunaLift that give AI additional business context, such as product information, brand guidelines, policies, pricing, or internal documentation.

E-E-A-T = Experience, Expertise, Authoritativeness, and Trustworthiness. A framework used by Google and reflected by AI systems to evaluate the credibility and reliability of content.

Knowledge Base = Your collection of business information and documentation that LunaLift uses to understand your organisation and improve AI optimisation recommendations.

LLM (Large Language Model) = An AI model trained on large amounts of text to understand and generate human language. Examples include GPT, Claude, Gemini, and Llama.

Markdown = A lightweight text format that uses simple syntax for headings, lists, links, and emphasis. It's easy for both humans and machines to read, making it a common format for AI-ready content.

Prompt = The question or instruction submitted to an AI model. LunaLift tracks prompts to measure how your brand appears across relevant customer searches.

Rich snippets = Enhanced search results that display additional information, such as ratings, prices, FAQs, or product details, often generated from structured data (schema markup).

Schema = Structured data (typically JSON-LD) added to webpages to help search engines and AI systems understand what the content represents, such as a product, article, organisation, person, or FAQ. See schema.org for the complete vocabulary.

Schema attributes = Individual properties within schema markup, such as author, product name, price, or faqQuestion. These provide structured facts that help machines understand entities and relationships.

Site health = An overall measure of your website's optimisation, calculated as the average of your Technical Score and Content Score.

Source = The webpage, document, dataset, or domain an AI model uses when generating an answer, whether or not it provides an explicit citation.

Value proposition (Value Prop) = A clear statement explaining why a customer should choose your product or service, highlighting the unique benefits and outcomes you provide compared to alternatives.

Workflow = A sequence of automated or manual steps that helps you complete a task in LunaLift, such as producing new or updated content and fixing website issues at scale.