Glossary

Overview#

Definitions for the terms used across Seerly. Each entry says what a term means and how to read it, rather than how it is calculated, on the grounds that the second is rarely what you need at the moment you look something up.

Generative engine optimization (GEO): the practice this product exists to serve, which is being named in the answers AI assistants compose. It also travels under AEO and AIO, among others. See Generative Engine Optimization.

Measurement#

Prompt: one realistic buyer question that Seerly asks an AI engine on your behalf, such as "what's the best project management tool for a small agency?". Prompts stay fixed and are asked again each cycle, which is what makes results comparable across time rather than a series of unrelated snapshots.

Topic: a cluster of related demand that groups prompts together. Topics are what you approve during setup and what every metric in the product is reported against, which makes them the most consequential thing in your account.

Run: one complete pass of your prompt set across the engines you track. How often a run happens depends on your plan.

AI answer surface: any place where an AI system answers a question directly instead of returning a list of links. ChatGPT, Perplexity and Google AI Overview are the surfaces tracked on standard plans.

Scores on the Overview#

GEO Score: Seerly's composite of how discoverable your brand is to AI engines, on a 0 to 100 scale. Fed by AI Discovery Reach, your site's readiness from the page audit, and brand sentiment, with reach the biggest contributor. Comparable against your own history rather than against other companies. It covers the latest week's runs and its change is in points against the previous week.

AI Discovery Reach: the share of the latest week's prompt runs where an AI engine mentioned your brand, with its change in points against the previous week. A run is a single prompt on a single engine, and every successful run counts whether you appeared or not, so this tracks how often you surface rather than where you rank.

LLM Performance Index: the per-engine breakdown on the Overview, carrying a Visibility Index and a Sentiment Score for each engine, so one weak engine cannot hide inside a healthy average.

Citation influence: how much the AI engines lean on a source when answering about your category, scored 0 to 100 and banded Dominant, High, Moderate or Minor. It appears on the Overview's source list, on Who's Citing and on Browse's domain tiles, and it is the same number in all three. It is measured across all your completed runs, not only the latest week, so it stays steady from week to week. Each of the three draws it the same way too: a bar whose length is the score and whose shade deepens with the band, with the score and band written beneath it, so a list can be read by shape before you read any of the numbers.

It combines four things: how often the source appears in answers relative to your other sources, how many engines cite it, how many of your topics it covers, and how many distinct pages of it get cited. The first of those carries the most weight.

Two things worth knowing. It is a standing within your project, so a 74 here and a 74 on another project are not the same claim, and a source's score can move when other sources change even if its own citations did not. And it spans your whole topic space rather than just you, which is why every row carries a badge counting the answers where that source cited you, amber and reading Never cited you when there are none. On the Overview and Who's Citing that badge counts the latest week: a source can shape answers in your category heavily and never mention you.

Where the sample is too small to rank against, such as very few runs or a single engine or a single topic, no score is shown at all rather than a misleading one.

Not to be confused with Backlink influence below, which is a different measure on a different screen.

Visibility#

Discovery Rank: your position among tracked competitors, normalised to a 1 to 10 scale, where lower is better. It runs in the opposite direction to most numbers in the product, so a move from 6 to 3 is an improvement. Because it is a position rather than a quantity, it can worsen in a week when nothing about your own presence changed, simply because a competitor improved.

Mention rate: the share of prompt runs whose answer names the brand. It counts presence only, so a brand named last in every answer can still have a high mention rate; position is the measure that says where it was named.

Recommendation rate: the share of prompt runs whose answer recommends the brand, meaning the answer puts it forward as an option to choose or consider. An answer that names the brand only as a source, an example or a point of comparison counts toward mention rate and not here, so over the same runs recommendation rate is never higher than mention rate.

Position: where inside an answer your brand appears. Being the first brand named carries considerably more weight than being the fifth of six alternatives listed at the end.

Share of Voice (SoV): your presence relative to your tracked competitors across the same prompts. Visibility asks whether you appear at all; share of voice asks whether you appear instead of your competitors, and the two can move in opposite directions in the same week.

Citations#

Citation: a source an AI engine referenced while answering. Citations matter because they shape what the answer says, which means influencing the sources is how you influence the answer.

Citation gap: a topic and engine pair where you are absent and something else is being cited in your place. A gap names both the absence and whatever is filling it, which is what makes it actionable rather than merely discouraging.

Earned citations: citations from sites the brand does not own. A citation from your own site shows the engine found your page; an earned one shows somebody else vouched for you, which is harder to come by and usually worth more. Seerly can only split the two once it knows which site is the brand's own.

Favoured by: the AI engine that leans on a source most, adjusted for how many prompts each engine answered, so no engine wins just by answering more of them. It is measured across all your runs and does not change when you filter. Only an engine you still track can be named, and only once at least 3 of its answers cite the source. A project tracking a single engine never shows it, since there is nothing to compare. Either way, a source without one is not a finding.

Recurrence: how many consecutive runs a source has continued to be cited across. A page cited once is noise; a page cited every run is the page you are genuinely competing with.

Sentiment#

Sentiment: whether AI answers describe your brand positively, neutrally or negatively. A large neutral share is the normal shape rather than a problem, because most factual mentions of a brand are neutral.

Aspect: the specific theme driving a sentiment judgement, such as pricing, support, quality or ease of use. Aspects are the layer you can actually act on, since nobody can fix "negative sentiment" while "negative on support response times" is a piece of work with an owner.

Where each term appears#

Most of these map to a particular screen. The GEO Score, AI Discovery Reach and the LLM Performance Index live on the dashboard Overview. Citation influence reads the same wherever it appears: drawn as a bar on the dashboard Overview's source list, on Monitoring → Citations → Who's Citing You and on the domain tiles of Monitoring → Citations → Browse Citations, and as a score and band in that page's List view. Favoured by reads the same everywhere it appears: the dashboard Overview's source list, the Sources citing your topics and Sources you're missing lists on Monitoring → Citations → Overview, a column on Monitoring → Citations → Who's Citing You, and both views of Monitoring → Citations → Browse Citations. Backlink influence is a separate measure on Foundations → Web Presence. Discovery Rank, Share of Voice, Discovery Reach Trend and Weekly drivers all live on Monitoring → Visibility. Mention rate and Earned citations appear side by side for you and one competitor on Monitoring → Competitors. Aspects live on Monitoring → Sentiment → Aspects. Citation gaps live on Monitoring → Gap Analysis, and clusters live on Monitoring → Keywords → Universe.

Web presence#

Backlink influence: how much a linking site counts for, rated high, medium or low. On Web Presence the column is labelled simply Influence, and it is a different measure on a different screen from Citation influence above: that one is about the sources AI engines read when answering, this one is about the sites linking to you. If you are comparing the two, they are not the same number and not on the same scale.

Backlink target influence: the same judgement applied to a site you are not yet linked from, so an outreach list can be read strongest-first. Labelled Target influence on the screen.

Site health#

Smart Audit: an analysis of your public pages covering how easily an AI engine can read, understand and cite them. It overlaps with a traditional SEO audit without asking the same question, and its actions include generated fixes and an llms.txt draft for your site.

Readiness check: a focused check on whether AI crawlers can reach your content at all, which is the prerequisite everything else on the site health side depends on.

Content#

Content credit: the unit consumed when you generate content. See seerly.app/pricing.

Brand Hub: where your brand voice, positioning and the terms you would rather avoid are stored, so that anything generated stays consistent with the rest of your material.

What to keep in mind#

These are product definitions rather than industry ones, and the industry has not settled on much yet. Where a metric is expressed as a percentage, treat direction and relative comparison as the meaningful signal rather than the absolute figure, which rarely means anything on its own.