GEO Corporate Assessment

The crisis you thought was closed? AI tells it in the present tense.

Analysis and monitoring of corporate reputation in the answers of ChatGPT, Gemini and Perplexity — where journalists, investors and stakeholders now run their first checks.

Request the free 48-hour audit

Example sample: 480 questions to five engines, one dot per answer
ChatGPTGeminiPerplexityClaudeCopilot
  • 52%Correct
  • 17%Dated facts
  • 12%Unattended sources
  • 19%Doesn't name you
ChatGPT · the circled answer in the sample, annotated with its sources

Question“Is Velaria Group a reliable and ethical company?”

AI answerVelaria Group is a mid-sized company. In 2020 it was at the centre of a controversy covered by several outlets.1 Some sources describe its communications as not very transparent.2 [No mention of the sustainability report or the most recent initiatives.]3

  1. DatedThe 2020 controversy is told by the Wikipedia article, which stopped in that year.en.wikipedia.org · last edited 2021
  2. UnattendedThe judgement comes from a forum and a trade blog, cited without a date.forum · trade blog
  3. MissingThe 2025 sustainability report is published, but no source the AI engines read picks it up.ESG report, not cited

Illustrative example with a fictitious company. In the assessment every note links to the source URL.

The GEO Corporate Assessment is the service that analyses how generative AI engines tell the story and reliability of a company to journalists, analysts and stakeholders: dated controversies still told in the present tense, dominant unattended sources and narratives to correct, with the actions to do it.

What has changed

Whoever looks you up no longer reads ten articles. They read a summary written by an AI.

“Is [Company] a reliable and ethical company?”

“What happened with [the 20XX affair]?”

“How is [Company] positioned on sustainability and governance?”

×2ChatGPT weekly users doubled in a year: from ~400 to ~800–900 million
8%click-through when an AI answer appears (vs 15% without): your version is not read
60%of searches already end without a single click: the summary is enough

Sources: OpenAI / press 2025 · Pew Research 2025 · SparkToro 2024 — US / cross-industry data, to be read as a trend.

The problem

AI forms its view from news outlets and Wikipedia. Not from your official voice.

And every journalist, analyst or candidate starts from that summary.

The sources that matter

Where the answer comes from, for this audience

Wikipedia

The single most cited source by AI: an entry frozen for years is the most frequent cause of controversies told in the present tense.

National and trade press

Crisis articles remain authoritative sources for years; AI repeats them until a more recent source exists.

Wikidata and Knowledge Graph

They identify the company and its basic facts: headquarters, leadership, dates. An incomplete entity produces confusion and namesakes.

ESG reports and "facts" pages

They count if published as readable HTML with data and dates; a 120-page PDF is not extracted.

The solution

How the GEO Assessment works

1

We ask the AI engines

We put the real questions of your audience to several models (ChatGPT, Gemini, Perplexity, Claude, Copilot), by market and role.

2

We find the sources

We identify the actual articles and pages that form that perception, with verified URLs.

3

We give you the plan, and carry it out

Score, gaps and concrete actions per touchpoint. Your team executes them, or we do: content, sources, presence. Then we measure again.

What you receive: GEO Score A–F · Visual SWOT · Source map · Reputational risks · Action plan · Competitor comparison. One branded report, exportable to PDF, generated by the proprietary ReputAction platform. And with monitoring, the same snapshot repeated over time: you see how far your positioning has moved.

The method in detail →

A case, anonymised

Industrial group, 2020 controversy closed in 2022

Every answer to "is it a reliable company" opened with the controversy, in the present tense; the sustainability report and the closure of the case never appeared. The plan updated Wikipedia and Wikidata with sources, published a "what has changed" timeline, secured an interview with data in the trade press and made the ESG report extractable. At 90 days the controversy appears in one answer in four, always with the subsequent facts; the ESG report is cited by Perplexity and ChatGPT.

All results →

Inside the platform

The score over time, with alerts when a narrative emerges

Monthly runs and automatic alerts: you know about a review or an article entering AI answers before a journalist reads it.

Monitoring view: GEO Score trend over 12 months with an alert on a new negative narrative

ReputAction platform interface, with sample data about a fictitious company.

Frequently asked questions

GEO Corporate Assessment: the essential answers

What is the GEO Corporate Assessment?

The GEO Corporate Assessment is the service that analyses how generative AI engines tell the story and reliability of a company to journalists, analysts and stakeholders: dated controversies still told in the present tense, dominant unattended sources and narratives to correct, with the actions to do it.

How does the analysis work?

We ask several AI models (ChatGPT, Gemini, Perplexity, Claude, Copilot) the real questions your audience asks, identify the sources that shape the answers, with verified URLs, and deliver a GEO Score from A to F, a SWOT, reputational risks and an action plan per channel. The demo audit is ready in 48 hours.

Can a closed crisis really resurface in AI answers?

Yes: if the most authoritative sources on an affair remain those of the time, the AI tells it in the present tense even years later. The assessment identifies the sources keeping the narrative alive and the actions to update them or balance them with more recent content.

Can you make a news item disappear from the answers?

No, and that is not the goal. The realistic goal is for the fact to appear less often, with the right date and alongside what happened next. That is achieved by updating sources, not removing them.

Do you also monitor ongoing crises?

Yes: with monthly monitoring and alerts you see when a new narrative enters the answers, before it becomes the standard version. It is included in the GEO Program.

Who carries out the actions in the plan?

Your team, or us: every action in the report is operational (updating sources, making pages and listings readable by AI, looking after reviews and Wikipedia) and we can carry it out directly, or work alongside whoever already runs your channels.

How do I measure results (ROI)?

Through monitoring: periodic runs of the same questions on the same AI engines, GEO Score over time and a before/after comparison for each intervention. You see how far your positioning has moved against the investment, and an alert warns you if a negative narrative emerges.

How do we start, and what does it cost?

You start with a free demo audit, presented on a call within 48 hours: real AI answers about your brand and an indicative GEO Score. The full GEO Assessment starts from €5,000 for one audience and one country, with the baseline actions carried out by us and the 90-day re-analysis; the GEO Program, for serial content, PR, communities and monitoring, is quoted on the plan. Details are on the Pricing page.

Blog

Further reading

The same assessment, for your industry

In life sciences? This assessment exists in a version dedicated to pharma, within the regulatory perimeter.

Pharma version →

The next step

Start with the free 48-hour audit; then the GEO Assessment, from €5,000, with the baseline actions carried out by us and the 90-day re-analysis; the GEO Program for everything else.

Pricing and what is included →

Free audit in 48 hours

Want to see what AI says about you? We show you in 48 hours.

A free demo audit, no commitment: real AI answers about your brand and an indicative GEO Score, presented on a call.

No newsletter: we only contact you to reply.