How MentionX turns AI answers into evidence

How a scan works

AI engines do not publish what they say about a brand, and they do not say the same thing twice. A MentionX scan treats each answer as a piece of evidence: it asks the same questions a buyer would ask, records exactly what came back, and turns those records into numbers that can be checked.

The dataset

The research dataset holds 6,813 completed AI engine answers to 2,311 prompts, with 36,944 citations captured from those answers. Every answer is stored with its engine, its prompt and its date, so any figure on this page can be traced back to the answers that produced it.

The answers span ten engines and answer surfaces: ChatGPT, Gemini, Google AI Mode, Google AI Overviews, Perplexity, Microsoft Copilot, Claude, Grok, DeepSeek and Meta AI. The brands scanned cover B2B software categories such as CRM, workflow software, real estate technology and AI visibility tools.

The dataset is repetitive by design. A large share of scans are repeat runs of identical prompts, which is what lets us measure how much AI answers change from one run to the next. Failed runs are excluded from every count, and the prompt count includes scans that did not complete.

One visibility score, three signals

Being named is not the same as being chosen, and being chosen is not the same as being trusted. The visibility score blends all three: how often a brand is named, where it sits when engines list options, and how much of the cited evidence points to it.

A worked example

Take an illustrative brand that is named in 30% of answers, sits at an average position of 3 when ranked in a scan whose lowest position is 10, and holds 20% of brand-linked citations.

A brand counts once per answer however many times it is named. Because citation share only counts citations linked to a tracked brand, it can swing widely in scans where few answers cite any tracked brand.

Every KPI, defined

Each KPI is computed per scan from the structured record of every completed answer, using the same formulas as the MentionX product.

A blended score can hide the real picture

We tested a young marketing software brand with the same 20 prompts across 3,384 AI answers. Its overall mention rate held steady between 33% and 36%. That looked healthy, until we split the prompts by whether they named the brand.

When a buyer asks about the brand by name, engines mention it almost every time and put it in a ranked shortlist in 56.8% of answers. When a buyer asks a category question that does not name it, the brand appears in 0.05% of answers. The blended number moved so little because it mostly measured the share of prompts that already contained the brand name.

The lesson for anyone tracking AI visibility: report category questions on their own. They are the questions where buyers discover options, and they are the honest baseline to improve against.

The same question, different answers

We ran a large business software brand through the same category prompts on ChatGPT and Gemini, again and again, every one to three days. Nothing about the prompts changed. The answers did.

Broad, high-intent questions such as “which platforms are best for customer relationship management?” mentioned the brand every time. Narrow questions about specific jobs, such as following up stalled deals or data protection for European companies, rarely or never did. The prompts that flipped mentioned the brand in about half of their runs.

The lesson: one scan is one sample. A move of up to 25 points between two runs is normal noise, so trends should be read across repeated scans on the same prompt set, never from two adjacent points.

What AI engines cite

On buyer questions that do not name the brand, engines cited 27,007 sources. The brand’s own website made up 0.9% of them, about 1 in every 108 citations. Engines decide who to recommend mostly from what other people publish.

8.8% of citations were Google grounding links that hide the real source. Among visible sources, tech media, review sites, communities and video each carried more weight than the brand itself. TechRadar, Gartner, Wikipedia, G2, Reddit and YouTube were among the most-cited domains across the software categories we scanned.

When a buyer asks about a brand by name, the pattern reverses: engines cite the brand’s own pages heavily. Owned content explains who you are. Third-party coverage decides whether you are recommended.

What AI says when it is negative

MentionX reads every answer for individual statements about the scanned brand. Across 2,455 such statements, 33.1% were negative. Most of them were not complaints about quality: they were about fit and price.

Use-case fit (31.2%) and pricing and value (13.9%) account for 45.1% of all negatives. A further 15.3% were doubt rather than criticism: knowledge gaps, unverified claims and hedged answers, where an engine simply did not have enough to go on.

The negative share also depends on where buyers ask: it ran from 25.0% on Google AI Mode to 42.6% on DeepSeek. A brand that only reads one engine sees only part of its reputation.

When two scans can be compared

Prompts are tailored to each brand and stored as a fixed set. Because prompts that name the brand are answered with a mention almost every time, category questions are always reported on their own.

What the method does not claim

The general method, including how ranges are calculated, is on the methodology page. Standards for every statistic on Learn are in the editorial policy.

Questions