Wilson confidence interval
The Wilson confidence interval is a range for a percentage measured on a sample; for AI visibility it says where the true share of answers naming a brand likely lies, given how many answers were counted.
Why AI visibility needs it
AI answers vary from run to run. Real example (Neobanks in the UK, 31 Oct – 2 Nov 2026): Revolut was named in 96 of 324 answers, a visibility of 30%. The true value could reasonably be anywhere in 25–35% (95% Wilson interval). Reporting the range is the honest way.
The formula
p = k / n, z = 1.96 (95%)
centre = (p + z²/2n) / (1 + z²/n)
half width = z · √(p(1 − p)/n + z²/4n²) / (1 + z²/n)
range = centre ± half widthUnlike the simple normal interval (p ± 1.96 · √(p(1 − p)/n)), the Wilson interval stays between 0% and 100% and behaves well for small samples and for shares near 0% or 100%.
Telling real changes from noise
Two values only differ for real when their intervals don't overlap. Example: Revolut 30% (25–35%) against Monzo 78% (74–83%). The two ranges don't overlap, so the difference is real for this week.
kennt.ai applies this rule to every trend: a change counts as real only after two full weeks and without overlapping intervals.
Example numbers: Neobanks in the UK, 31 Oct – 2 Nov 2026; real measurement, Revolut is not a kennt.ai customer.
Related
Does AI recommend you?
The free AI visibility check asks ChatGPT, Gemini and Google AI Mode 10 buyer questions each, live. 45 seconds, no sign-up.