Risk Scoring

The risk score (0–100) is computed by a deterministic, config-weighted rule engine — not a machine-learning model. Each matched risk flag adds a configurable weight to the score; the higher the total, the greater the estimated risk.

Score bands

The actual allow/review/block decision depends on where the score falls relative to your own configured review_threshold and block_threshold (see Quickstart) — the bands above are descriptive, not the literal decision boundaries.

Signal categories

Shield evaluates several independent categories per transaction and combines their weighted flags into the final score:

Reacting to decisions

const res = await fetch('https://api.wajub.com/payments', {
method: 'POST',
headers: { Authorization: 'pk_test.xxxx', 'Content-Type': 'application/json' },
body: JSON.stringify(params),
}).then(r => r.json());

switch (res.shield?.decision) {
case 'block':
  return showError('Payment declined for security reasons.');
case 'review':
  return showPending('Verification in progress, you will be notified.');
default:
  return redirect(res.authorization_url);
}