What investors are asking about your company, and why IR stopped seeing it
Investor questions used to reach IR because every one of them passed through a person. Many are now typed into a chatbot instead, where IR never sees them, and here is how to get that signal back.
In short
You know what investors ask you. You have much less idea what they ask about you. Most IR teams have a decent picture of the questions that arrive by email, on a call, in a meeting or from the AGM floor. The ones typed into a chatbot at half past eleven at night get answered by something else, and nothing about them ever reaches you. Getting that signal back means giving the question somewhere to land on a channel you own.
Key facts
- IR sees the questions that arrive through a channel IR owns: the inbox, calls, meetings, the AGM, webcast Q&A.
- A question asked in a general AI tool gets answered without you and leaves no record on your side.
- On an open model you do not control which sources the answer uses, or whether it cites yours.
- An assistant on your own site answers from the company information you approved, and the questions asked there stay with you.
How did IR know what investors wanted to know?
The channels did that work, quietly.
An analyst called before updating the model. A portfolio manager asked the same thing twice in one week of meetings, and by the third time you knew it belonged in the next presentation. Someone mailed the IR address about how the dividend policy works. The AGM floor told you which part of the strategy had not landed, usually in the least comfortable way possible.
None of that was a system. It was a by-product of every question having to pass through a person to get answered. That is what made it visible, and that is what has changed.
What changes when the question goes to a chatbot instead
The question still gets asked. It gets answered too. You are simply not in the room for either.
No mail in the inbox, no call to log, nobody to follow up with. The investor reads an answer assembled from whatever the model has about your company, forms a view, and moves on. If the answer was thin, you never get the chance to correct it, because you never learn the question was put in the first place.
This is the other half of a shift we have written about from the sourcing side: where do AI answers about a company come from. That page is about what the investor gets told. This one is about what you stop hearing.
Which questions disappear first
The ones people would rather not put to a person.
Start with the small ones. How the dividend policy actually works. Whether guidance is on a reported or an organic basis. Which segment a new plant sits in. Plenty of investors will not email IR for that, because emailing a company feels like a big move for a small question. Typing it into a chatbot takes four seconds and costs nothing socially.
Then the awkward ones. Dilution. A covenant. Why a board member left in March. Nobody is watching when you ask a model.
Then the practical ones. An institution in Milan or Frankfurt reads your reporting in English but would rather ask in its own language. A shareholder in Singapore has a question at an hour when nobody at your company is awake.
Those four groups were always the weakest part of your picture of investor sentiment. What is different now is that they have somewhere else to go, and that somewhere answers instantly.
Getting the question back onto your own channel
The question comes back to you when the best available answer is one you host.
In practice that means an investor can ask in plain language somewhere on your IR pages, and get an answer drawn from the company information you approved, with a citation for each point, and a straight "that is not covered in the published information" when it genuinely is not. MIRA is a source grounded IR assistant; what that means, and what it answers from, is set out in what is a source grounded IR assistant.
The visibility is a side effect of the setup rather than a feature bolted on top. The question was asked on your channel, so the question stays with you.
What to do with the questions once you can see them
Read the list as a backlog, not as a report.
- The IR FAQ. If a dozen people asked a version of the same question this month, that answer belongs on the page, in your own wording.
- Results preparation. Walking into a reporting day with the actual questions from the last quarter beats walking in with the ones you happen to remember.
- The wording of the equity story. When one passage keeps generating the same question, the passage is doing its job badly. Rewriting it is cheaper than answering it fifty more times.
- AGM season. Questions collected across a year are better preparation than a brainstorm two weeks out.
None of this needs a project or a new workflow. It needs somebody to read the list every other week and act on the top three items.
What the question list does not tell you
It is a record of what was asked on your own channel. That is the whole of it, and being clear about the edges matters more than the promise.
It does not tell you what investors are asking ChatGPT or Perplexity about your company, because nobody outside those companies can see that. It is not engagement tracking and it is not a view of who opened which document. It will not tell you who was behind a question unless the person chose to say so.
One channel, honestly recorded, is still a real answer to something IR has been guessing at for the last two years.
The bottom line
Investor questions used to be visible to IR because every one of them had to go through a person to get answered. That has stopped being true, and the loss is quiet, because a question you never receive does not feel like anything. IR gets the signal back by owning the place where the question is answered, which is the same place where the question gets recorded.
Want to see the questions your own investors are asking? Book a MIRA demo and we will set it up against your published company information, so the first questions you see are real ones.
About the author: Sander van Prooijen is Head of Commercial Strategy at Finvictum, where he leads the European rollout of MIRA. He works with IR teams across European markets on how investors, and increasingly AI, find and read the equity story. LinkedIn: https://www.linkedin.com/in/svprooijen/