Short definition. Algorithmic visibility is the capacity of a company to be recognised, described correctly and recommended by the algorithmic systems that sit between the client and the market: generative engines (ChatGPT, Claude, Perplexity, Gemini, Microsoft Copilot), the AI summaries of search engines and recommendation systems. An algorithmic visibility audit measures that capacity with a reproducible method and returns five answers:
- Whether the brand appears in AI answers to real client queries.
- Whether AI describes it accurately (what it does, where, for whom).
- Which sources the engine relies on — and whether the brand is in them.
- Whether the website is technically legible to AI crawlers (Schema.org, NAP, robots).
- Who appears in its place when the brand does not.
Why this discipline exists
For twenty years the question was 'where do I rank on Google?'. That question still matters, but it is no longer the only one and, in many B2B sectors, no longer the main one. Decision-makers research suppliers by asking generative engines, and those engines do not return ten links: they return one answer with two or three names. How those mentions are shared out is the new field of play, and most companies do not even know whether they are on it.
We measured it with data in our Madrid Bar Test, run on the Spanish market: of some fifty established firms, half scored an absolute zero for visibility in AI, and the ones being recommended were not the largest or the most prestigious — they were the most legible.
The five dimensions the audit measures
1. Presence: does the brand appear?
A standardised set of real queries is run — the ones a client would actually ask, in their own words, not in jargon — across the five generative engines, and we record which ones the brand appears in, in what position and how often. This is the base benchmark. In our methodology we call it the Bar Test: twenty queries across five engines, with a dated screenshot of every answer.
2. Accuracy: does it describe the firm well?
Appearing badly can be worse than not appearing. We audit whether the engine attributes the speciality, the location and the services correctly — and we record the hallucinations: invented facts the model states with confidence.
3. Authority: which sources does the engine rely on?
Every generative answer is built on sources: directories, press, profiles, corporate websites. The audit identifies which ones the engine uses in the client's sector and checks whether the brand is present and consistent in them. A good part of the result is decided here: models only recommend entities they can corroborate in independent sources.
4. Technical legibility: can a machine read you?
We audit the data infrastructure of the website: Schema.org markup in JSON-LD, NAP consistency (name, address and telephone identical across every asset), access for AI crawlers in robots.txt, sitemap and effective indexing. It is the fastest part to fix and the one that surprises clients most in diagnostics: websites that are excellent for humans and mute to a machine.
5. Competition: who occupies your place?
The most uncomfortable dimension, and the most useful. When the brand does not appear, somebody else does. The audit names them: which competitors receive the recommendation, why the engine recognises them, and what they have that the brand does not. It turns an abstract problem ('AI cannot see me') into a concrete and contestable one ('this specific name is taking my query').
Who carries it out: the auditor's profile
It is an emerging profile that does not fit the traditional boxes. It is not an SEO consultant (the object of study is not the Google ranking), nor a marketing consultant (no campaigns are produced), nor a classic systems auditor. It is a data infrastructure auditor: someone who works with structured markup, entity graphs, data consistency and reproducible verification — and who signs off a finding with a method behind it, not an opinion with slides.
AuditScale carries out this type of audit as a specialism for law firms, tax and accountancy practices and professional boutiques. Sector specialisation matters: the queries a law firm's client asks, the directories engines cross-reference in the legal sector and the professional-conduct limits on what may be claimed are not those of an e-commerce business.
Which deliverables to insist on
- A reproducible benchmark: dated screenshots, by engine and by query. Without a date and an engine, there is no measurement.
- A map of algorithmic competitors: who appears in place of the brand, and why.
- A technical audit: state of the Schema, NAP, robots, sitemap, indexing.
- A prioritised 30/60/90-day plan: what to correct first and what result to expect from each phase.
- A re-measurement method: how and when the measurement is repeated to verify progress.
Its relationship with GEO
The algorithmic visibility audit is the diagnostic phase of Generative Engine Optimization: first you measure, then you correct. If you want to understand the full discipline — what it is, its phases, prices and how to choose a provider — we set it out in the guide GEO consulting for law firms.
Frequently asked questions
What is algorithmic visibility?
It is the capacity of a company or firm to be recognised, described correctly and recommended by the algorithmic systems that sit between the client and the market: generative engines (ChatGPT, Claude, Perplexity, Gemini, Copilot), the AI summaries of search engines and recommendation systems. A firm can have an impeccable website and a real reputation and still be invisible to these systems if its information is not legible and verifiable by a machine.
What exactly does an algorithmic visibility audit measure?
Five dimensions: presence (whether the brand appears in AI answers to real client queries), accuracy (whether AI describes correctly what it does and where), authority (which sources the engine relies on and whether the brand is in them), technical legibility (Schema.org, NAP, accessibility for AI crawlers) and position against competitors (who appears in its place). The result is condensed into a reproducible benchmark with a date and an engine.
Who carries out an algorithmic visibility audit?
It is an emerging profile, halfway between technical auditing and semantic engineering: the data infrastructure auditor. It is not a marketing consultant in the usual sense — the work is structured markup, entity consistency and reproducible verification. AuditScale carries out this type of audit, specialising in law firms, accountancy practices and professional boutiques, with a Bar Test of 20 queries across the five generative engines.
Which deliverables should a serious audit include?
At a minimum: the visibility benchmark with dated screenshots by engine and query, the identification of who appears in place of the brand, the technical audit of the website (Schema, NAP, robots, sitemap), a prioritised 30/60/90-day implementation plan, and a re-measurement method to verify progress. If the provider cannot show you the measurement method, it is not an audit: it is an opinion.
Update (July 2026): the metric, measured on ourselves
So that this does not stay theoretical: we applied our own audit to AuditScale and measured the result over two months on the Spanish market. We went from zero citations in AI answers to more than 300, and the progression was not a straight line but an irregular one — days at zero included. It is exactly the kind of progress a visibility audit should be able to measure and explain, not promise. The full experiment, here.
Want to measure your firm's algorithmic visibility?
The GEO audit includes a bespoke Bar Test of twenty queries across the five generative engines, the map of who appears in your place, the full technical audit and a 30/60/90-day plan. From €1,500.
Request a diagnostic