RecommendedOutcome

The highest-level verification output is the RecommendedOutcome. It's the culmination of all the checks, options, properties of the document, and the properties of the captured images.

Sometimes, a fraud verdict can be made, but it's clear there's uncertainty or that the capture was of insufficient quality. Other times, there could be an extremely strong signal, and we can ignore sources of uncertainty. Maybe you've elected to prioritize having high assurance, and there are confounding factors which prevent us from recommending accepting the document, even though Verify can't find anything outright wrong.

The point is - there are checks and their results, then there's the overall fraud decision, and then finally, there's the question of what to do with the document and the end user.

The purpose of the RecommendedOutcome is to answer the question of what to do with the document and the end user, and how to proceed with the transaction (whatever it may be). The core idea is that you should essentially be able to "tell" Verify how you define fraud and what you* want to do in which situation, and then have it do all the heavy lifting and just give you a single, simple answer.

This definition of fraud is done using the UseCase system, along with fine tuning specific Options if necessary.

RecommendedOutcome values

There are a few, pretty invariable things that can happen when verifying a document:

• We can be sure it's genuine and live
• We can be sure it's not genuine or not live
• We can be unsure about one or both of the above
• The document could be completely unextractable, or unsupported by our systems
• We can be fairly certain the quality of the images is poor to a degree that would affect the system's ability to make a fraud verdict

This is true of our system, and it should generally be true of any other systems or even human experts.

Thus, the RecommendedOutcome can be one of the following values:

• Accept - the document is genuine and live
• Reject - the document is not genuine or not live
• Retry - the document cannot be verified due to image quality issues with the current configuration
• ManuallyReview - the document should be reviewed manually - you still get a verdict, and detailed check results
• Undeterminable - we know the document is not supported by our systems, or the image is so poor we couldn't even find a supported document in it

A demonstration with default options

Let's look at an example of Verify's RecommendedOutcomes (using default options and a default use case) for various scenarios.

OverallFraudCheck	Image Quality	CertaintyLevel	RecommendedOutcome	Explanation
Fail	Very Poor	Low	ManualReview	Low confidence requires human verification
Fail	Very Poor	Medium	Retry	Very poor image quality may cause false positives
Fail	Very Poor	High	Reject	High confidence despite poor image quality
Fail	Good	Low	ManualReview	Low confidence requires human verification, despite good image quality
Fail	Good	Medium	Reject	Good image quality gives more credence to medium confidence
Fail	Good	High	Reject	High confidence with good image quality - clear fraud detection

API reference

For possible values, and more API details, see the RecommendedOutcome reference.