Configuring Decision Logic

Learn how to configure DocV decision logic using product settings and workflow logic to optimize document verification outcomes.

Configuring Decision Logic

DocV decisioning is designed around two complementary layers. Each layer plays a different role, and together they determine both the Predictive Document Verification module’s internal decision and the final workflow outcome you return to your application.

1. DocV Product Settings

These settings control how the Document Verification module evaluates an ID. They determine:

  • How extracted data is interpreted (age, expiration, PII matching).
  • The module’s internal decision (accept, review, resubmit, reject). Note: Not applicable to DocV Omega
  • Which reason codes are generated to explain that decision.

Think of this layer as configuring how DocV will determine when to fire certain reason codes.

2. RiskOS™ Workflow Logic

Once DocV returns a decision and its reason codes, you use RiskOS™ workflow logic to apply your business rules:

  • Allowed document types
  • DocV Omega score threshold
  • Country/region restrictions
  • Reason-code or signal overrides
  • Compliance requirements


This layer determines the final outcome your app receives.

👍

Tip:

  • Product settings shape DocV’s reason code output.
  • Workflow logic produces the final decision (Accept / Reject / Resubmit / Review / Cancel) returned from RiskOS™.

Decision with the DocV Omega score

DocV Omega includes a calibrated fraud score to the DocV response. This score can be utilized in your RiskOS™ workflow logic in place of the decision.value.

The three outputs

OutputWhat it's for
ScoreThe calibrated activeModelScore from 0 to 1. This is the primary output your workflow decisions on.
Reason codesMap to your existing reject logic, review queues or resubmission UX.
SignalsThe raw intelligence detail, for teams that build custom logic on top of specific signal categories (for example, captureIntelligence or graphIntelligence).

A standard configuration decisions on the score and reason codes. An advanced configuration adds the raw signals for teams that want to build their own logic on specific fields.

Set your reject threshold

The activeModelScore ranges from 0 to 1, where a higher value indicates greater fraud risk. Configure a reject threshold, then reject any transaction that scores at or above it:

  • Stricter posture — lower the threshold to reject more transactions.
  • More lenient posture — raise the threshold to reject fewer transactions.

Your Socure account team can help you determine a starting threshold — for example, by recommending a value based on industry benchmarks and your risk tolerance, by calculating a value that holds your current accept rate steady during migration, or by optimizing against your fraud feedback and labeling data.

Decision logic

Apply the score in this order:

  1. Check for a resubmit outcome. If DocV returns a resubmit decision, the transaction is not eligible for a DocV Omega score. Handle it as a resubmit and prompt the consumer to recapture.
  2. Compare the score to your threshold. Check activeModelScore against your configured threshold. A score at or above the threshold routes to Reject.
  3. Layer in supplemental reason codes and signals. Beyond the score, you can route selected reason codes or signals to your Reject or Resubmit outcomes, depending on your risk posture.
  4. Otherwise, accept.

How reason codes work with the score

DocV reason codes fall into three functional groups for decisioning:

  • Already reflected in the score — signals the model already evaluates, such as document tampering, failed liveness, and document-to-selfie mismatch. These fire for visibility and need no separate configuration.
  • Automatic resubmit — capture or image issues that make a transaction ineligible for a score and surface as decision.value = resubmit.
  • Configured separately — business-rule codes the model does not score, such as expired documents, PII mismatches, and minimum-age requirements. You control these in your Predictive DocV Product Settings and RiskOS™ workflow logic.

Resolve individual codes at runtime against the reason code catalog in the RiskOS™ Dashboard rather than hard-coding descriptions. See the reason code list.

📘

Source-of-Truth (AAMVA) codes:

When you enable AAMVA for your use case, an additional set of field-level Source-of-Truth codes applies. These are not part of the DocV Omega score — incorporate them into your decision logic separately. See Source of Truth Checks.


Restrict accepted documents by country

DocV evaluates any supported document regardless of its issuing country, and it does not reject documents issued outside the United States on its own. If your RiskOS™ workflow logic does not include a Country/region restriction, DocV can return an accept decision or passing score for supported non-US documents, such as foreign passports.

To accept only documents issued by a specific country or region — for example, to accept only US-issued documents — add a Country/region restriction to your RiskOS™ workflow logic.


Access the RiskOS™ Dashboard

To view the full DocV Decision Logic documentation, go to Settings > Documentation > DocV Decision Logic in the RiskOS™ Dashboard or use the following links.

EnvironmentDashboard Link
SandboxGo to Sandbox Dashboard
ProductionGo to Production Dashboard

Did this page help you?