How Cipher works

How Surbee scores response quality, from bots and AI-written answers to careless respondents.

Cipher is Surbee's response quality engine. Every response to a Surbee survey is scored automatically. Cipher looks for bots and AI-written answers, and also for careless, ill-intent and wrong-fit respondents, so you can trust your data without manually reviewing each submission.

Cipher combines four kinds of evidence:

  1. Behavior. How the respondent moved, typed, scrolled, and paused.
  2. Device and network. Browser fingerprint, automation flags, VPN, proxy, and datacenter signals.
  3. Content. Whether open answers look human, original, and consistent.
  4. Coordination. Whether many responses look like the same actor or a fraud ring.

The risk score

Each response receives a Cipher risk score from 0 to 1, where higher means more suspicious. Responses at or above 0.5 are flagged and surfaced in the dashboard (and in your Slack channel, if connected). Every flagged response carries a list of reasons so you know why.

You stay in control

Cipher recommends, it does not delete. Flagged responses are still collected. You decide whether to keep, review, or discard them.

What Cipher can tell you about a response

A score on its own is not very convincing. For every response, Cipher also turns the raw signals into a few plain-language observations, so you see not just that a response is trusted or flagged but why. For example:

  • "Answered on an iPhone, held in hand, with natural typing and one fast answer on Q2."
  • "Pasted a long answer into the feedback question and switched tabs twice."
  • "Desktop, steady mouse movement, consistent answers, no quality concerns."

Each response shows the two or three signals that moved its score the most, and you can open any response to see the full breakdown, question by question. The same plain-language treatment works on desktop and on mobile.

On phones and tablets

A phone has no mouse and no keyboard, so Cipher does not pretend to grade one as if it did. On a touch device it streams a different set of signals, including touch pressure, swipe shape, and the device's own motion, and scores the response on those. See Cipher on phones and tablets.

Cipher levels

Cipher runs at one of three levels. Each level includes everything in the one below it. Higher levels add more checks and cost credits per response, so you can match depth to how much risk a study carries.

LevelPlanFocusExample checks
BasicEvery plan, including Free. No charge.Speed, care, attention and automationRapid completion, straight lining, careless rating patterns, attention checks, hidden trap fields, pasting, bots, device spoofing
ProHiveHow each written answer was producedTyped, revised, pasted or copied out, hidden characters, typing rhythm, tab switching, pointer, hover and scroll behavior, sentence rhythm, contradictions, copied text, light AI review
MaxSwarm and EnterpriseAcross responses, network and the webFraud rings, shared answers, coordinated timing, shared devices, reputation across your surveys, VPN, datacenter, Tor, proxy and timezone, thorough AI review, web search for copied answers

Every flag comes with plain-language reasoning at every level.

How Cipher spots AI-written answers

Cipher catches AI-written answers from human behavior: how each answer originated, where it came from, whether it was typed or pasted, how quickly it arrived, and when. The wording is read on top of that evidence.

The Max web search checks whether long written answers match text already published online. It is context aware, so questions that ask people to quote, cite sources, or paste text are skipped.

What gets checked

Some of the signals Cipher looks at:

  • Timing. Completion faster than the survey can be read, or timing that doesn't look like a person working through it.
  • Interaction. Too little interaction for a real attempt.
  • Typing. Typing a person doesn't produce, and heavy pasting.
  • Answers. Straight lining (the same option repeatedly), low answer entropy, duplicated or near duplicate responses.
  • Automation. Browsers driven by software, and devices whose details don't fit together.
  • Network. VPN, proxy, Tor, datacenter IPs, and timezone mismatches.
  • Content. AI-written open answers, copied text, and answers that contradict each other.
  • Coordination. Shared devices, shared answers, and synchronized timing across responses (fraud rings).

Cipher keeps learning

Cipher is more than a fixed rulebook. Scored responses, and the verdicts you give when you mark a flagged response as genuine or not, are kept as labeled examples for training future versions of Cipher's model.

Desktop and mobile responses are labeled on signals that fit the device, so a phone is never taught as a bot for having no mouse. See Cipher on phones and tablets.

Using Cipher

  • Inside Surbee. Cipher runs automatically on every published survey. Turn it on or off, pick a level, and override individual checks in the survey's Cipher settings.
  • From your own app. Use the Cipher API and SDK to score responses you collect anywhere. See Cipher SDK and API.