Somewhere between the roadshow and the results call, a new kind of investor has quietly joined every fundraising process, annual report review, and activist campaign. This shadow investor group has read your filings in full, cross-referenced them against your competitors’, and formed a view on your capital allocation before your IR team has finished breakfast. But they have never met your CFO, and they never will — because they are generative AI, and the growing number of investors who are using it to conduct investment research.
In fact, Brunswick’s 2026 US Investor Survey of 100 institutional investors finds that 54% of those surveyed call GenAI moderately important to very important in their research. Forty-two percent use it as a leading tool for diligence on new positions. And more strikingly, 68% say GenAI has already changed how they approach earnings calls. One survey respondent even remarked, “Now I rely on AI-generated transcripts, which tell me the tone and manner of how management Q&A was interpreted.”
Statements and statistics like these demonstrate the importance of generative engine optimisation (GEO) for IR. Companies need to optimise their IR content for AI, recognising that this increasingly important second reader has joined the room. While an AI engine may never ask a follow-up question or signal that an answer didn’t land, it remembers what you said and repeats it to the next investor who asks.
Leading with “why”
The most important factor in your GenAI-ready content, the Brunswick study found, is a clear and simple story that tells investors “why you will win.” Sixty-one percent of Brunswick’s respondents chose this as the top driver of investor confidence. But the survey also revealed that only 27% of respondents believe companies do that well. While most companies know where their own competitive advantage lies, very few have written it down anywhere that a machine — or a human — could find it, cite it and correctly repeat it.
What’s needed: a model that summarises your equity story by pulling from what’s structured, current and discoverable online. Because whatever lives only in a banker’s slide deck or your CFO’s head will never be seen. Nearly half of investors (47%) already rank company websites and IR pages as important sources. Another quarter give GenAI results the same weight. If your own site isn’t the clearest, most current account of why you’ll win, something else will fill that gap — a stale sell-side note, or (worse yet) a competitor’s version of your story.
From audience intelligence to AI visibility
This is where GEO for IR earns its keep, and where it connects directly to the audience intelligence work we described in an earlier piece: you can’t structure a narrative for an audience whose actual questions you haven’t measured. Before touching any content, we run a perception and coverage diagnostic — by investor type, by geography — alongside a baseline of what AI models currently say about you when nobody’s briefed them.
The diagnostic tells you where the gaps are, and how to close them with content strategy. We map the actual questions investors and their AI tools put to a prompt window, then test what today’s models return for each one, and look for cases where the honest, complete answer isn’t currently surfacing. Closing that gap means publishing specific, accurate, well-sourced answers on the properties a model already treats as authoritative — your own IR page and disclosures.
The cost of getting it wrong
Brunswick’s data sharpens the cost of getting this wrong. “Overpromising and underdelivering” was, by a wide margin, the top trust-killer investors named — 62%. This polled well ahead of “dodging tough questions” or “blaming external factors.” Ninety-three percent said they wouldn’t invest in a company they don’t trust, regardless of the numbers. GEO built on structured, consistent information makes every future statement measurable against what’s public. It provides discipline that makes sure there is no divergence between your promise and your delivery — in front of a model that never forgets what you said last quarter.
Managing two tracks
For listed companies, our Investor Relations offer folds this into the existing IR calendar. The AI visibility baseline sits inside the same diagnostic as your analyst coverage audit, and GEO-structured content rides alongside the annual report and quarterly notes you’re already producing.
Investment funds are on a different track. Rather than working on a quarterly basis, LPs reconvene when you’re raising — perhaps three or five years from now. The track record deck that a GenAI model finds today needs to still hold up, unedited, whenever an LP asks what your fund did with its last vintage. We fold GEO structuring into the same materials, so what an AI surfaces during due diligence matches, word for word, what your placement agents or fundraising teams present in the room.
Keeping the right conversation going
None of this means you have to write for a machine instead of a real, live investor. Everything should start from the same principle: know what the people who allocate your capital are actually asking, then make sure the true, well-told answer is the easiest one to find — for anything, human or otherwise.
The AI reading your filings this quarter doesn’t need convincing — it needs the facts in one place, correctly. That’s a smaller job than most IR teams assume, and a bigger miss than most have budgeted for.
Start that job with a simple audit: see what today’s models already say about you on the handful of queries that matter most to your investors, and identify where that falls short of the truth.
The Editorialist helps you navigate your GEO challenges, ensuring your content is visible, accurate and relevant across generative AI platforms.
Talk to our experts
