Proof

Most of what makes this platform trustworthy is not modelling. Message based authentication, knowledge based challenge questions, third party database checks, cryptographic sealing and the programmatic enforcement of each state's notarial requirements are all deterministic, and a live commissioned notary supplies the judgement in the notarial path. Models do real work in credential analysis, biometric comparison and the newer deepfake detection layer. Apply the removal test and a functioning remote notarization and signing platform remains, which is exactly what the business ran as for years before the model layer arrived.

The notarial path contains the strongest form of human oversight in this index, and it is not a design preference but a legal requirement: a commissioned notary meets the signer over live video, forms their own judgement about the person and the act, and is personally accountable for it under a state commission and bond. The platform reinforces rather than substitutes for that, programmatically enforcing each state's notarial requirements so the notary cannot omit a required element. Identity only transactions are a different matter, running automated with no described review queue or escalation route for a signer the checks reject.

The evidentiary architecture is genuinely strong: identity is bound to the signature and the result cryptographically sealed so the transaction cannot be repudiated later, which produces an audit record a court can examine and is a better answer than most vendors here can offer. The model layer is undocumented.

Completing 25 verification checks in under five seconds describes throughput and coverage, not correctness, and no accuracy, false acceptance or false rejection figures, model documentation or evaluation methodology were located.

The evidence is deep, quantified and attached to identifiable parties. Cumulative scale is stated at more than 185 billion dollars in real estate closings secured and millions of customers served, with named deployments at a title company and a credit union working with motor vehicle agencies on electronically notarized documents.

Outcome figures are specific to each role in the transaction: title agents saving up to 157 minutes per closing, lenders removing up to seven days from funding cycle time, hybrid closings cutting in person meeting time by 99 minutes and error rates by up to 31 percent, and per transaction savings of up to 110 dollars. The platform is accepted in all fifty states.

The fraud layer is described as drawing on signals from across the network as well as deepfake detection, which means intelligence derived from one customer's transactions informs protection for another, and nothing public defines what crosses that boundary or whether a customer can decline participation.

No model providers are identified for credential analysis, biometric comparison or deepfake detection, no accuracy evidence is published, and nothing describes how the detection models are maintained as generation tools improve.

The data footprint is among the heaviest in this index. A single closing session captures government identity document images, a biometric selfie, video and audio of the signer, and the complete document set for a mortgage, which together constitute a full financial and identity profile of a consumer at their most exposed moment. Knowledge based challenge questions are generated from consumer credit file data, which draws consumer reporting obligations into the flow. No published privacy framework, retention schedule for session recordings, or subprocessor list was located.

Assurance is named rather than asserted, with a service organisation control attestation cited alongside conformance to a mortgage industry data standard and to a federal identity assurance level, the last of which is a specific graded conformance rather than a general claim and is directly relevant to what the product does.

The presentation is the weak part: those standards appear in editorial comparison content on the company's own site rather than on a dedicated trust surface, with no audit period, scope statement or report request path located.

The first top grade on this axis in the index, and it is earned because licensure sits inside the product rather than beside it. Proof operates a network of notaries who hold individual state commissions, and the platform encodes and enforces the notarial law of each jurisdiction so that an act completed through it satisfies the statute where it occurred.

Acceptance is claimed across all fifty states, with the surrounding ecosystem of title underwriters and county recording offices engaged directly, and the material is careful to tell users that eligibility still depends on lender, underwriter and county rather than overstating it. Conformance is also claimed against a named federal identity assurance level and a mortgage industry data standard.

Two verification methods carry uneven failure rates and neither is analysed publicly. Biometric comparison and credential analysis vary in accuracy by skin tone, age and document type, the familiar pattern across this category. The sharper issue is knowledge based authentication, whose challenge questions are drawn from consumer credit history, so anyone with a thin file fails disproportionately: young buyers, recent arrivals, people who have used credit sparingly.

In this setting a failed identity proof does not mean a retry later, it means a home closing does not complete on the day it was scheduled. No demographic pass rates or accessibility analysis were located.

Accountability here is structural and unusual. A notarial act is performed by a commissioned individual carrying personal liability and, in most states, a bond, so a party harmed by a defective notarization has a named human and a statutory remedy rather than only a software vendor's terms. Cryptographic sealing produces non repudiable evidence that a court can test, which is the point of the product.

What is absent is the vendor's own commitment: no accuracy guarantee for the automated verification layer, and no described route for a signer wrongly rejected by identity proofing to escalate rather than simply fail.

The verification stack is openly described as combining several methods, and two of them necessarily involve outside parties, since third party database checks and credit derived challenge questions cannot originate in house. That the chain exists is therefore disclosed by implication; who is in it is not.

No data providers, credit bureau relationships, biometric or deepfake detection model suppliers or subprocessors are named, and nothing states where identity documents and session recordings are processed.

Integration is designed around the multi party reality of a closing rather than a single customer's stack, with a collaboration layer letting lenders and title agents share documents and hand off transactions inside one workflow, and interfaces available for embedding the platform into existing systems. Conformance to the mortgage industry data standard implies interoperability with loan origination systems, and the workflow reaches outward to underwriters and recording offices. What was not located is a named partner directory, public developer documentation or specific origination and title production system integrations.

Delivery is cloud hosted with a domestic focus across all fifty states, so the cross border complexity facing global vendors here does not arise in the same form. Residency still matters, because session video and audio, identity document images and executed instruments must be retained for years under state notarial journal requirements and mortgage record keeping rules. No hosting regions, retention locations, tenancy model or subprocessor chain were located in this pass.

The billing model is published even though the rate is not. Both the title and business products are described as pay per transaction rather than seat licensed or annually committed, which suits a buyer whose closing volume moves with the rate cycle and who cannot commit to a fixed subscription, and a free account is available so a title professional can create and send transactions before any commercial conversation. Actual per transaction prices, volume tiers and enterprise terms remain undisclosed.

Coverage spans both sides of a property transaction, reaching mortgage lenders and title agencies with collaboration tooling built for the handoff between them, and extends to banks, credit unions, insurers, retailers and law firms. Transaction types are enumerated rather than implied, covering first mortgages, refinances, cash and seller side closings, home equity lines and loan modifications, alongside non property uses such as vehicle sales and wills. The boundary is functional: this is transaction execution and identity, with nothing addressing lending decisions, payments, trading or wealth management.