GEO for Investor Relations: what AI tells investors about your company

An investor weighing your stock once opened a search engine, scanned a page of ranked links, and clicked through to your investor relations site; increasingly, that same investor now poses the question directly to ChatGPT, Gemini, Copilot or Perplexity, requesting the investment case, the principal risks, and the circumstances under which the share price might fall.

The model responds in its own words, synthesising an answer from sources you largely neither own nor monitor, and the investor may never arrive at your website at all. That synthesised answer, assembled beyond your control, has quietly become a component of your equity story, and generative engine optimisation for investor relations is the discipline of understanding it and, where the levers permit, shaping it.

Few IR functions have yet audited what the models actually say about their company, even though the answer is already circulating in conversations that influence allocation decisions. This article sets out what generative engine optimisation (GEO) is, how it diverges from the search optimisation your team already understands, and how to conduct a repeatable audit of your own AI presence before your investors conduct one for you.

What is generative engine optimisation (GEO)?

Generative engine optimisation is the practice of influencing how AI systems characterise your company when a user queries them, and it differs from conventional search in a consequential respect: where a search engine returns a ranked list of links and leaves interpretation to the reader, a generative engine composes a single answer and presents it with the authority of settled fact. The optimisation objective shifts accordingly, from securing a position on a results page to becoming the source the model trusts when it assembles its reply.

For a listed company the exposure is concrete, because the models are already answering questions that inform investment decisions, and they compose those answers from whatever they treat as authoritative: reference encyclopaedias, financial media, your own disclosures, analyst and consensus material, and the less disciplined precincts of the internet. Where that composite is stale, partial or simply inaccurate, it operates against you in exchanges you will never witness and cannot correct in the moment.

GEO vs SEO: what changes for investor relations

Search optimisation and generative engine optimisation share a family resemblance, but the operative levers have moved. Traditional SEO rewarded keywords, backlinks and site architecture, all instrumented toward a click; GEO instead rewards narrative coherence, machine-readable factual structure, and demonstrable authority, all instrumented toward being cited inside an answer that may generate no click whatsoever.

Three consequences follow for investor relations. First, because the destination is no longer your own property, the accuracy of the material that exists about you elsewhere weighs as heavily as the polish of the pages you publish yourself. Second, consistency becomes a ranking determinant in its own right, because when your equity narrative, headline figures and corporate descriptors diverge across your website, your regulatory disclosures and third-party profiles, the model's confidence in all of them erodes. Third, the discipline is continuous rather than episodic, since the answer drifts as the sources beneath it are updated, superseded or contradicted.

4 steps to running a GEO audit for investor relations

A worthwhile audit is repeatable and costs nothing to initiate, and it proceeds in four stages.

1. Run the audit

Pose the questions your investors ask to ChatGPT, Gemini, Copilot and Perplexity, and repeat them in each of your principal market languages, since the source composition changes with language. A representative set:

What is the investment case for [Company], and why would an investor buy the stock?

What are the principal risks of investing in [Company], and why might its share price decline?

 

What is [Company]'s strategy, and how did it perform in its most recent results?

 

How does [Company] compare with its peers as an investment?

 

Who are [Company]'s largest shareholders, and is it a credible ESG investment?

 

For each response, record two things: the substance of what the model asserted, and the sources it cited. Perplexity is particularly instructive here, because it exposes its provenance explicitly.

2. Build a risk map

Plot where the narrative is inaccurate, outdated or unfavourable, and identify which sources are driving each instance, so that anything demonstrably wrong and traceable to a source you can influence becomes an immediate priority.

3. Feed the machine with sources you control

A poor answer is not suppressed but out-published, displaced by accurate, machine-readable signal. Begin with your reference pages, since a Wikipedia entry is frequently the model's largest single source and the one IR audits least; then establish your IR site as the authoritative record, with a structured FAQ that answers the questions above directly, publish consistently, and hold the facts identical across every surface.

4. Monitor every quarter 

Re-run the audit, track how the answers and their underlying sources migrate, and you acquire a defensible, quarter-on-quarter demonstration that the narrative is shifting in your favour.

The compliance line

Two cautions are material for a regulated function. Because models hallucinate, every assertion a model makes about your company should be independently verified before it is acted upon, since an erroneous answer is, in most cases, a content problem to be corrected at source rather than an incident to be managed.

And the narrative should be shaped exclusively with accurate, published information disseminated through legitimate channels, because planting misleading or selectively disclosed information to influence perception would constitute a fair-disclosure problem under the Market Abuse Regulation. GEO is a discipline of becoming the trusted source, not of manipulating the model.

Put it into practice

The GEO check is one component of AI Tools for IR Communications, a practical toolkit from Euronext Corporate Solutions. Alongside the complete GEO audit methodology, it equips investor relations and communications teams with ready-to-run prompts for the earnings call, Investor Day, AGM and roadshow cycle, a checklist for evaluating any AI tool against the Market Abuse Regulation (MAR) and GDPR, and guidance on which category of AI is appropriate for which class of IR data.