EXAMPLE · ILLUSTRATIVE
Why this transaction was flagged
SCORE FACTORS
Contributing rulesweighted
Graph signalsweighted
Decline patternsweighted
Baseline deviationsweighted
Unclear synthesizes every payment signal into one explainable decision, turning transactional noise into operational clarity. A risk score is only useful if you can defend it, so Unclear hands you the number and the reasoning every time.
Most fraud tools hand you a score and ask you to trust it. Unclear gives you the score and the reasoning, on your own data, in language a reviewer and an auditor can both follow.
Each scoring engine examines a different dimension of risk. Their signals combine into one explained score, and each engine's contribution is shown so you know which lens drove the call.
Rules, graph signals, decline patterns and baseline deviations each show their point weight. Any run is reproducible against the model and ruleset version that scored it.
Suspicious transactions are connected into rings, so you see the pattern and not just the single event.
Every scored run is queryable. Ask what you would ask an analyst. Unclear answers from your own data and shows the runs and factors behind the answer.
Unclear reads the transaction exports you already produce and maps your columns to its canonical fields. No re-plumbing your stack and no connectors to certify.
Upload CSV or XLSX in the portal.
Push files over the API and receive webhooks on the Understand plan and above.
Have a processor drop files into a watched folder. Same pipeline, same results.
Same engine, three vantage points: ISOs, ISVs and processors. Whatever seat you are in, Unclear turns a pile of transactions into a decision you can defend.
Unclear scores and explains. It never touches money, which keeps it out of money-transmission regulation by design.
Card numbers are truncated and emails hashed on ingestion. Full PANs are never stored.
Every report is isolated to its owner. Analysts sign in through your identity provider, such as Google SSO.
Priced by payment events per month. Prices in US dollars. Current terms are always at unclear.ai/pricing.
Unclear reads your transaction files, scores each transaction for fraud, and shows the reasons behind every number, plus how suspicious transactions link together into rings. It runs after transactions have occurred, as an analysis layer, not in your authorization path.
CSV and XLSX in, scored verdicts out. Upload in the portal, push files over the API, or have a processor drop them into a watched folder. Unclear maps your columns to its canonical fields, so you do not have to reshape your exports.
Unclear is built to stay out of PCI scope. Card numbers are truncated and emails hashed on ingestion, full PANs are never stored, and it does not move money, so it stays out of money-transmission regulation by design. Every report is tenant-isolated to its owner.
Every score breaks down into the factors that produced it: the contributing rules, graph signals, decline patterns and baseline deviations, each with its point weight. Any run is reproducible against the model and ruleset version that scored it.
Start free on the Observe plan, or request a scoped trial and Unclear will set you up on your own files. Once you are live, analysts sign in through your identity provider such as Google SSO, and your administrator controls who is provisioned.
Start free on your own files, or ask for a scoped trial.
Get started for free →