Why does AI give an incomplete picture of smaller listed companies?

In short

General AI tools are only as good as the public material available about a company. For large, heavily covered names there is plenty, so the answers are usually solid. For small and less-covered listed companies there is far less. The tools then tend to produce a tidy but incomplete story, one that can leave out material facts from the company’s own filings. A source-grounded investor relations assistant works the other way round. It draws on the company’s approved information alone, so it can surface those facts and cite each one.

Key facts

• How good a general AI answer is depends on how much public material exists about a company.

• For thinly covered small and mid-caps, that material is sparse, so answers can be incomplete.

• Tools that rely on secondary coverage can miss material facts that sit in the filings.

• A source-grounded assistant answers from the company’s own disclosures and cites each point.

For a large-cap, asking a general AI about the company usually works well. There is a lot of public material, and the tool has plenty to draw on. For a smaller listed company, the same question can produce something that reads well and leaves out what matters most. The technology does not dislike small companies. There is just less about them to read.

Why do general AI tools struggle with less-covered companies?

Because they depend on how much public coverage there is, and how good it is. A general AI answers from the material it can find, such as analyst notes, news, reference sites and commentary. For a widely followed company that pool is deep. For a small or mid-cap that few analysts cover and few journalists write about, it is shallow. With little good material to work from, the tool tends to produce a smooth, plausible story that fills the gaps. What it does not reflect is the specific reality in the company’s own record.

That is where a less-covered company is most exposed. The investor gets a confident answer that sounds complete, with no sign that the important detail is missing.

What kind of material facts get missed?

The ones with consequences. In practice, the facts most likely to drop out of a thin, secondary-source story are the ones that most change an investor’s view.

• a difficult balance-sheet position
• a formal restructuring or rescue process
• a corporate action such as a reverse split
• dilutive recapitalisations

All of these sit clearly in the company’s own disclosures. But if the AI leans on limited secondary coverage instead of the filings, they can drop out of the picture entirely.

The result is not a small inaccuracy. It is a materially rosier or vaguer story than the company’s own record tells, and the investor receives it as fact.

Why does this hit small and mid-caps hardest?

Because they combine the two conditions that cause the problem. They matter to the investors looking at them, and they are thinly covered. A large-cap has enough public material for a general tool to lean on. A small or mid-cap in a specialist niche often does not, and such companies are common across Belgian and Nordic markets. So the companies with the least third-party coverage are the ones whose story a general AI is most likely to tell incompletely. That happens at exactly the moment an investor is trying to understand them.

How does a source-grounded assistant handle a thinly covered company differently?

It does not depend on outside coverage at all. A source-grounded assistant works only from the company’s own approved, published information. Whether or not analysts cover the company, the material facts in its filings are there to be surfaced and cited, point by point. When something is not in the disclosures, the assistant says so. It does not invent a tidy story to fill the space.

MIRA is built around this. For a less-covered company, the difference is direct. Without it, an investor gets a plausible summary assembled from scarce secondary material. With it, they get an answer drawn straight from the company’s record, with a citation behind each point and an honest note where the filings are silent. The value is in the sourcing, and in the completeness that comes from working off the primary record instead of whatever coverage happens to exist.

What should a smaller listed company do about it?

See what a general AI currently says about you, and compare it with your own disclosures. If you are lightly covered, look specifically for the material facts that a thin story would smooth over, and check whether they appear at all. Then decide where you want investors to form their first impression. Is it from that answer, or from an assistant that answers only from your own record and cites it?

If your company is thinly covered, this is not a marginal concern. It is the difference between an investor seeing your actual position and seeing a plausible sketch of it.

The bottom line

General AI is only as complete as the public coverage of a company. For a thinly covered small or mid-cap, it can produce a tidy story with the material facts left out. A source-grounded assistant works from what the company itself has approved and published, and from nothing else. So those facts can be surfaced and cited, however little analyst coverage exists. For a less-covered company, that is the difference between an accurate picture and a plausible sketch.

Want to see how your own filings would come across? Book a MIRA demo and we will show you MIRA answering investor questions about your company from your published documents, so you can compare it with what a general AI says about you today.

MIRA is an investor relations assistant that answers investor questions from a company’s own approved, published information. It is one of the Finvictum group’s tools for direct, compliant communication between companies and their investors.