Read AEO trends over time
How do I tell a real AI visibility gain from a first measurement?
Trends plots one point per month, added automatically on the 1st. Two charts: your overall AI visibility against your competitors' average, and your score on each engine separately. Before you read either as progress, check what the first point actually was.
A jump from zero is usually not a gain
Here is the trap, and it catches almost every location in its first few months.
A Miami agency's visibility line runs from 0 in July to 81 in August. That looks like the best month any business ever had. It is not. July was this location's first report and it scored 0 — the line is measuring the difference between not being measured usefully and being measured, not eleven weeks of marketing.
The same artefact shows up as a +81 delta chip, and it will put the location at the top of Biggest gains in the all-locations roll-up.
So before reading a trend, ask which of these you are looking at:
A first real measurement
- Shape
- a single steep line from 0 or near-0
- What it means
- the earlier month was your first report, or the engines named nobody
- What to do
- treat the later value as your baseline and start counting from there
A genuine month-over-month move
- Shape
- both points are non-zero and the change is modest
- What it means
- something actually shifted — your listings, your reviews, or a competitor's
- What to do
- read the per-engine chart to see which engine moved
A small change between two non-zero months is far more informative than a large one against zero.
The per-engine chart is where the answer is
The second chart splits the same history by engine, and it is the one worth your time. An overall score that held steady can hide one engine falling and another rising, which is a completely different situation from genuine stability.
It is also where an engine-specific problem becomes visible over time. A Texas auto-repair shop sitting at ChatGPT 100, Gemini 43 and Perplexity 95 has one line to watch and two to leave alone.
Competitors on the same axes
The visibility chart draws your competitors' average alongside you, so you can see whether a drop is yours or the category's. When everyone falls together, the engines changed how they answer; when only you fall, something on your side did.
Competitor history builds up the same way yours does — a month at a time — so early on you will see your line against a single competitor point rather than a curve.
Before the second month arrives
With only one data point there is no line to draw, so the tab says so plainly: "This is your first monthly data point. A new one is added automatically on the 1st of each month, so the trend line fills in from here — no action needed."
Beside that message is a See example → link. It draws a sample upward curve so you can see what the chart will look like once it fills in. It is clearly badged Example · projected, not your data and drawn as a dashed line.
That preview is an illustration with invented numbers, and it always slopes upward. Never read it as a forecast, and never screenshot it for a client — it is not a projection of your trajectory, just a picture of the chart's shape.
Questions people ask
My score didn't change at all this month. Is the report broken?
Not necessarily. A flat line between two real reports means the engines answered much the same way twice, which is a genuine result. A flat line at 0 across several months means you are not being named at all — that is the case to act on, starting with citations and reviews.
Why does my competitor's line stop partway?
Competitor scores only exist for months where that business appeared in an answer. A gap means they were not named that month, not that data is missing.
Where to go next
- What does my AI Visibility score actually measure? — what each point on the line is counting
- How do I compare AI visibility across all my locations? — the same artefact, at portfolio scale