A green check beside a seller’s name can be the most expensive piece of shorthand in your closing software. Somebody verified something. The question is what. An ID image passed a document check? A face matched the photograph? A person controlled a phone number? Those are useful findings, but they are different findings, and none of them, by itself, establishes that the person on the screen owns the property you are about to convey. I want the evidence behind the check before I want the check.

The Label Hides the Method

In October I wrote about the handoff problem: everybody in the transaction assumes somebody else verified the seller. Buying an identity tool can close part of that gap. It can also put a more authoritative label on the same assumption if the result arrives as a single word and nobody asks how it was reached.

NIST’s digital identity guidelines make a useful distinction between validating identity evidence and verifying that the applicant is the person that evidence describes. Checking whether a credential is authentic and checking whether it belongs to the presenter are separate jobs. The guidelines concern digital identity proofing; they are not a title examination standard. But that separation is exactly what a title company needs to understand when it buys verification software.

Ask your provider to describe a successful result in ordinary language. Which checks ran? Which sources were consulted? What could not be checked? A result that says the name and birth date agree with a database is different from one supported by document examination and a comparison to the presenter. The screen should let your closer see the difference.

A Face Match Has a Boundary

Matching a selfie to a photograph answers a narrow question about similarity. Liveness checks address whether the capture appears to involve a live person. Neither should be treated as a blanket guarantee against impersonation. Ask how the provider tests for manipulated media and for attacks that feed fabricated images or video into the capture process. “We use AI” is an especially unhelpful answer when the thing you are worried about was made with AI.

There is a practical question here for an independent with no security department: can the vendor explain the limits of the result to the person who has to act on it? If the only explanation lives in a technical appendix nobody in your office can interpret, the integration has stopped one step short of being useful.

Connect the Person to the Property

Even a well-supported identity result needs another connection. The person verified must be connected to the vested owner, and the transaction has to make sense against the record. That means resolving name differences, understanding the capacity in which someone is signing, and investigating inconsistencies rather than allowing a successful ID check to erase them.

ALTA’s September study found that 59% of surveyed firms reported a seller impersonation attempt in 2025, compared with 28% reporting prior-year attempts in its 2024 study. These are survey results, not a probability that your next file is fraudulent. They are a good reason to stop treating identity as an attachment collected at the end of intake.

Independent contact matters because evidence supplied by the same person can agree with itself beautifully. An ID, a phone number, and an email address all supplied by the purported seller are not three independent confirmations. When you corroborate contact details, record where they came from and investigate changes. An older address or phone number can also be stale; independence improves the evidence, but it does not make it infallible.

Design the Exception Before You Need It

A legitimate seller will eventually fail a check. A damaged credential, a name change, or an inaccessible capture process is enough to create a difficult afternoon. Decide in advance how that person gets reviewed. Otherwise, the closer under deadline pressure becomes the designer of your fallback procedure.

The alternative should collect appropriate evidence through an approved path and identify who can authorize proceeding. Record the reason for the exception and the basis for the decision. A failed automated check should lead to a review, not an accusation, and an override should leave more explanation in the file than a routine pass.

Retain the verification reference, method, time, and disposition in a way your team can retrieve, with access and retention appropriate to the sensitive information involved. Copying identity documents into every inbox is a poor substitute for preserving a usable verification record.

Ask to See the Result Behind the Result

On your next demo, ask to see a successful verification, an inconclusive one, and an approved exception. Follow each into the production file. Can an examiner understand what was established and what remains open without calling the salesperson?

That is the discussion I would rather have about title technology. We can talk through where identity evidence belongs in your intake workflow and where a person still has to make the decision. The green check should summarize evidence your team can examine, not end the examination.

Talk to Electra about Sentinel anti-fraud tools and identity evidence in your title workflow →