How does AI describe you?

Institutional research no longer happens only in a search engine. The question also gets asked to an AI assistant, and what comes back is often a single paragraph instead of ten blue links. If your name isn't in that paragraph, the conversation ends there as far as you're concerned.

Let's settle the guarantee question first

Nobody can guarantee that an AI system will mention your institution. If you receive a proposal that guarantees it, whoever wrote it either doesn't understand the mechanism or isn't being straight with you.

The reason is simple: there is no such thing as a “position” in these systems. Two people asking the same question on two different days from two different devices can get three different answers. With no rankable list, there is no ranking to promise.

That does not mean there is nothing to do. Instead of the part nobody can guarantee, we work on the part that can be steered.

Where AI learns about you

These systems don't know your institution. They read what the web says about you and derive a description from it. And the sources are plural:

• your own site and the structured data on it
• your Google Business Profile and map records
• education directories, sector listings, local guides
• news and press coverage
• forums, social media, third-party reviews

When those sources contradict each other, the system does not stop and ask which one is right. It either leans on whatever is repeated most often, or plays safe and doesn't mention you at all, or merges the two and produces a sentence that is wrong.

What inconsistency looks like in practice

Let's be concrete. For one institution, all three of these can be true on the web at the same time:

To a system reading all three, there is no single institution here — there are three similar-looking records. The usual outcome: you appear incompletely, or not at all. There is no bad intent involved, just noise.

What's notable is that this table was working against you before AI existed. AI only made the cost visible.

AI visibility is not a separate channel. It is your institutional facts finally being forced to agree.

The part that can be steered

What we can control is that the information about you is accurate, current, structured and reachable. All four.

• Accurate and singular: your name, campus count, programmes and conditions become identical across every source.
• Structured: the facts are published in a form machines can read, so nothing has to be guessed.
• Reachable: your pages are kept accessible to these systems. One that cannot reach you cannot describe you.
• Correct off-site too: outdated information in the third-party sources that mention you is found and fixed.

Most of this overlaps with SEO and Google visibility work — and it held its value before the AI side existed at all. What's new is that the same work now pays in a second place.

What happens over time

There is no one-off fix here. An institution whose facts agree becomes describable to these systems; one that is describable becomes, over time, recommendable. It is gradual, and no button makes it happen.

An institution whose facts contradict each other is either not mentioned or mentioned wrongly — and being mentioned wrongly can cost more than not being mentioned at all.

How we check the result

A precise ranking report is not possible, so we don't produce one. Instead we build a defined set of questions about your institution — your programmes, location, levels and competitive comparisons — test them at regular intervals and record the answers.

What you end up holding is not a score but an observation log: how you were described last quarter, how you are described now, which fact got fixed. We don't present it as a metric, because it isn't one.