Explore checkers, demos and evaluation routes →
Summarize this page with
Signup-risk assessment examines the registration’s session, device, network and related activity. Identity verification checks specified claims about an identity. Synthetic-identity detection evaluates manipulated or fabricated identity relationships. These layers can overlap inside one provider; a product category should not be reduced to its oldest feature.
- 1Name the risk
- 2Identify required evidence
- 3Choose the relevant layers
- 4Evaluate the combined flow
Start with the question each control answers
A device result can establish that the environment resembles one seen before. It does not establish the legal identity of the person using it. A document check can assess submitted evidence without establishing that the account’s later activity is permitted. An email confirmation can show access to an inbox without proving either conclusion.
The NIST Digital Identity Guidelines distinguish identity proofing, authentication and federation in their framework. That is useful vocabulary for product evaluation, not a reason to claim every application must use the same verification flow. Define your product’s requirements with the appropriate specialists.
| Layer | Main evidence | Example question |
|---|---|---|
| Session/device risk | Device, network, behavior and available history | Is this registration connected to suspicious activity? |
| Identity verification | Specified identity attributes, documents or other evidence | Do the submitted claims meet this verification requirement? |
| Synthetic-identity assessment | Identity consistency, relationships and network history | Do the identity elements suggest manipulation? |
| Challenge or additional verification | A completed interaction or possession check | Has this specific extra check been satisfied? |
Avoid treating vendor groups as fixed boundaries
Socure now offers Digital Intelligence, combining device, behavioral and entity evidence, alongside identity and synthetic-fraud models. Socure Launch provides a self-service route to selected workflows. Describing the entire provider as document-only or universally sales-gated would miss that scope.
Persona’s passive-signals product describes background device, network and behavioral context, linked accounts and adaptive routing. It is broader than a mandatory document-and-selfie screen. Confirm the selected product and plan rather than assuming every verification vendor produces only pass/fail.
Conversely, ShieldLabs provides live device/network and account-abuse context; that does not replace authoritative identity or document evidence when your flow needs that separate function. Evaluate the output you need, rather than forcing one product to answer every question.
Separate unfamiliarity from synthetic identity
A genuine new customer can have little public activity, a new phone number, a recent relocation or a newly created inbox. A synthetic identity can also develop a plausible history. Missing or inconsistent data is a reason to inspect the evidence and its coverage, not a universal rule that the applicant is fraudulent.
Deduce’s product material describes activity-backed identity intelligence outside a merchant’s own first session. Socure’s synthetic-fraud offering addresses identity relationships. Compare coverage, required inputs and unmatched handling, keeping device/account conclusions separate from identity conclusions.
Design the combined journey before buying modules
Decide which evidence is collected during registration, what result is needed before activation and which condition sends a user to an additional step. Record both the risk assessment and the actual action. A review state is not the same as a denied account; an unavailable result is not a verified identity.
Choose the minimum interaction that satisfies the actual requirement and risk case. Some checks can run passively; others require active participation. Test the selected sequence on supported web and app channels, including accessibility and genuine shared-device use. Do not promise that all users avoid verification simply because a device layer is present.
Evaluate two kinds of mistakes separately
An incorrect identity result and a mistaken abuse conclusion can have different causes. A genuine customer might pass identity verification but be incorrectly linked to another household account. A correctly identified person might still operate several ineligible accounts. Record which conclusion was wrong before changing a threshold or replacing a provider.
Use independently known identity and abuse facts. Do not label every failed document submission as fraud, every abandoned step as a false positive or every completed payment as proof of a legitimate signup. Keep unresolved cases separate and inspect the shared-device protocol.
FAQ
Can a device fingerprint replace identity verification?
It provides different evidence. Use it alongside the identity checks required by your product’s actual verification need.
Do identity providers only perform document checks?
No. Current products can combine passive device/network signals, identity data, synthetic-fraud models and workflow decisioning. Confirm the exact module.
Does a thin online history prove a synthetic identity?
No. Review coverage, missing data and independent identity facts, and include genuine thin-history applicants in the evaluation.
Sources
Checked October 21, 2026: NIST, Socure Digital Intelligence, Socure Launch, Persona, Deduce, ShieldLabs. See alternatives by layer.
Continue reading
- New account fraud red flags: what to check before activation
- How to prevent automated account creation and coordinated signup abuse
Compare all ten tools or read the complete deployment guide.