RESPONDENT VERIFICATION • BY RIWI

Every Respondent, A Real And Unique Human — Privacy By Design

VerifyHuman screens every respondent for fraud invisibly. Bots, click farms and repeat-takers are caught before the survey starts, with nothing to see and nothing to do. When the risk earns it, it steps up to a five-second on-device check that proves a live, unique human. No images. No identity. Ever.

EVERY SURVEY IS NOW A TURING TEST

At Least 1 In 3 Survey Respondents Worldwide Isn't Authentic

01 / 03
31%
Fraudulent responses
NORC and CloudResearch, 2026
02 / 03
99.8%
AI bot evasion
Of recent bots pass traditional attention checks (Dartmouth Westwood, 2025)
03 / 03
38%
Discard rate
When verification is applied after the fact (Kantar)
THE ASSURANCE LADDER

Invisible By Default, Friction Only When It's Earned

VerifyHuman verifies real humans without interrupting them. You choose how much assurance you want, and your respondents barely notice either way. Under the hood it is a single scoring engine: the server-side floor runs for everyone, the invisible layer feeds it richer on-page signal, and the camera is invoked only when the floor is uncertain.

Server-Side Floor

Scores every respondent from network, device and behavioural-reputation signals. Zero code on the page, so it works anywhere, instantly. The respondent sees nothing.

Invisible Layer

A silent on-page script adds live behavioural signal and catches automation. Still nothing to see and nothing to do.

Camera Step-Up

Only when something genuinely looks off, and only for that one person. Proof for the few, friction for no one else.

LIVENESS VERIFICATION

About Five Seconds, On The Respondent's Own Device

The top rung is the camera step-up, the part a bot cannot fake. Passive liveness, active challenges when needed, anti-spoof and face-uniqueness together prove a live, present, unique human is there.

Face geometry is computed on the device and never leaves it. No photos are stored, and VerifyHuman never learns who anyone is. Those few seconds are the privacy guarantee: what leaves the device is a verdict, not a face.

Most respondents never see it, because the invisible layers clear them first.

OUR APPROACH

Evidence We Do & Don't Leverage

A false rejection costs a panel a good respondent, an incentive, and a reputation. So parts of the system exist purely to throw evidence away.

Shared Machines Stay Shared

People who used the same library or office computer are not a fraud ring. When a group of accounts only holds together through one shared device, VerifyHuman tests for it and declines to score the group.

A Fingerprint Is Not A Person

Two people with the same phone model and browser version produce the same device fingerprint. When an identity turns up on many unrelated networks, or in two places at once, it is withdrawn entirely rather than used to accuse anyone.

Absence Is Never A Pass

"We checked and it was clean" and "we could not check" are different answers. VerifyHuman reports which one you got, so a quietly degraded signal can never look like a good result.

Calibrated On Real Panel Traffic, Not A Vendor's Sample

01 / 04
730
Supply sources profiled
02 / 04
848K
Panel history records
03 / 04
660K
Sessions scored monthly
04 / 04
~37
Respondent languages
GO DEEPER

The Full Product, In Detail

Two scores rather than one, four verification modes, the evidence we refuse to use, and how it is calibrated on real panel traffic.

The finale
R / 01