LGT Solutions
Visibility practice · Private briefing

What does AI tell your buyers your company is?

A four-minute briefing, prepared for one reader. We tested it. Here is what we found — and the one move that protects you.

EVIDENCE STATE: OBSERVED · PREPARED —
01The stakes

By 2028, your first buyer won't be a person.

When someone evaluates your category, they increasingly ask an AI first. In seconds it decides what kind of company you are — and quietly builds the shortlist. If it files you wrong, you are never on the list. No lost deal. No conversation. Just absence.

What this briefing is

OBSERVED · not confirmed by the client · not measured movement · not legal advice.

02The test

So we asked the machines what you are.

Not opinions. Controlled, neutral prompts — the kind a real buyer would type — run as a repeatable battery and archived probe by probe. Every claim on the next screens traces back to a recorded answer you can audit.

Method & coverage
03The finding
0/5
times the surface placed you in the right category.
0% narrowed No compression detected Grade —
The evidence behind this number
Verbatim finding

Coverage matrix
04The gap

The danger isn't today's answer. It's the box AI reaches for when it isn't sure.

What you actually are
The narrower box

One clean surface is reassuring, not conclusive. Two of three surfaces are still dark, and the uncertainty band is wide.

Uncertainty & likely driver
Attributed driver (from output evidence)

Inference from output behavior only. Not a claim about model internals.

05The move

One page closes the gap.

Who
Effort
Then
We re-probe all three surfaces and report what moved.

One change. Specific. Testable. Not a broad content strategy, not a rebrand — a single crawlable proof point, then a controlled re-test.

How we prove it moved (the A4 gate)
Re-probe gate elements
    Preregistration
    06The ask

    Four decisions. Then we run the loop.

    1
    Tell us if being seen as a narrower tool would actually cost you deals.
    2
    Confirm or correct the ground truth — your category and product pages.
    3
    Pick the first proof change to publish.
    4
    Let us re-probe once the change is indexed.
    If movement clears the gate, we report MEASURED. If it doesn't, we say so — and learn what kind of visibility problem this actually is. Either outcome moves you forward.
    What this packet does and does not claim
    This packet
    • Shows live AI characterization behavior from archived outputs.
    • Attributes likely drivers from observable output evidence.
    • Identifies one proof change worth testing.
    • Creates a re-probe plan with preregistered gates.
    This packet does not
    • Claim you confirmed these findings.
    • Claim measured movement — that needs the A4 gate.
    • Provide buyer-outcome evidence or legal advice.
    • Claim ranking causality.