Quote → Personal details → Document or registry check → Selfie or video → Payment → Policy issued
The selfie is often the only point in the journey where a live human face is required.
Insurance / NeuroVerify
Deepfake Detection API for Online Insurance Onboarding
Insurers and insurtechs that issue policies in one online session depend on a selfie or short video to bind a person to an application. NeuroVerify checks that media for deepfakes and manipulation beyond liveness and returns one unified risk score with evidence.
It runs behind the capture step already in your quote-to-policy flow: no new step for the proposer, one API call.
NeuroVerify does not underwrite risk, validate a proposer's documents or decide whether to issue a policy; it adds authenticity evidence for your own policy.
Powered byNeuroVerifyExplore the product
- #1 of 36on the public VendorBench-100
- <200 msp95 on eligible image requests
- 6 → 1six checks, one unified score
- 0new steps for your customer
The gap
A fast policy purchase can bind cover to a face that was never there.
Online insurance is built to turn a quote into an issued policy in one sitting. Life, health, motor and travel products all ask for some proof that the person buying cover is the person named on the application, and on a phone that proof is usually a selfie or short video.
Why insurance has a particular exposure
Identity at purchase matters long after the sale. A policy bound to the wrong person can surface months later at claim time, at nominee change, at surrender or at a payout to a new account. By then the original capture is the main evidence you hold.
The usual checks do not test authenticity. Registry and document checks confirm records. Face match confirms two images look alike. Liveness confirms a presentation check was satisfied. None of them ask whether the face was generated, swapped or injected.
Insurance also has a strong incentive to misrepresent: for example, a different person taking a medical or video interview for a proposer, or a synthetic persona used to hold policies that later support a fraudulent claim. See deepfake detection for insurance for the wider picture.
Threats
How a manipulated selfie reaches an issued policy.
These are techniques, not claims about any insurer. Each puts the wrong face behind a policy.
01
Stolen identity with a substituted face
A real identity is used with a swapped or synthetic face, so registry checks pass and the face match matches the document photo.
02
Proposer and insured split at the selfie
Someone else completes the video or selfie step for a proposer who would not pass underwriting, for example on health disclosures.
03
Photo-of-photo and screen replay
A printed photo or another screen is shown to the camera in the hope that basic liveness accepts it.
04
AI-generated or face-swapped selfies
A generated face is built from the document photo and submitted as the live capture.
05
Assisted onboarding with reused media
A distributor, agent or intermediary reuses captured media across several applications to speed up volume.
06
One face behind many policies
A single synthetic or borrowed face is attached to several identities to build cover that supports later claims.
Most of these look normal to document checks. Only the media itself shows the difference.
Where NeuroVerify fits
Authenticity evidence before the policy is issued.
NeuroVerify sits behind the selfie or video step in your purchase journey. You send the media you already capture. The response gives one unified risk score and evidence from six signals, and your underwriting or fraud policy decides whether to issue, step up or refer.
This supports your identity controls at the point of sale and gives reviewers evidence to reuse at claim time. It does not analyse the documents in the file.
NeuroVerify Unified Trust LayerSix parallel analysis layers
- Visual Safety Screening
- Live Presence Verification
- Face Manipulation Detection
- AI Generative Model Analysis
- AI Pixel-Level Manipulation
- AI Content Compliance
Send the proposer selfie or video
Pass the media from your existing capture step to the API using an SDK or REST.
Read the unified risk score
Receive one score with signal-level evidence in a single response.
Route by policy
Issue, request a fresh capture, refer to manual review or decline, based on your own thresholds and product risk.
Keep the evidence with the policy record
Store the output so claims and fraud teams can revisit it later.
What you get
What insurers and insurtechs gain.
Identity evidence at issuance
Add a check for deepfakes and manipulation to the moment cover is bound.
Fewer ghost and duplicate policies
Catch synthetic faces before they become a base for later claims.
Proof for later disputes
Signal-level evidence stays with the application for claims and audit teams.
No friction for genuine customers
Proposers complete the same selfie or video step.
Fits regulated deployments
Managed cloud, your VPC, on-premises or air-gapped.
Questions
What teams ask before they integrate.
How does NeuroVerify fit into your online policy purchase journey?
It sits behind the selfie or face-capture step you already run. You send the image, video or audio you already collect to one API call, and the response comes back with a unified risk score and signal-level evidence. The end user sees no new step.
Integration uses native mobile SDKs (Android/Kotlin, iOS/Swift, Flutter), backend SDKs (Python, TypeScript/Node.js, Go, Java) or the REST API. Two lines to integrate and live in under 24 hours are product claims; your own environment and review process set the real timeline.
We already run liveness. Why add deepfake detection?
Liveness asks whether a live person appears to be in front of the camera. It was designed around printed photos, screen replays and masks. It does not ask how the media was produced, so a face swap, a fully synthetic face or an injected stream can satisfy it.
NeuroVerify checks for deepfakes and manipulation beyond liveness. Keep your liveness vendor and add the authenticity evidence next to it. See deepfake detection vs liveness detection.
What does NeuroVerify return, and who makes the decision?
One unified risk score, plus evidence from six signals: Visual Safety Screening, Live Presence Verification, Face Manipulation Detection, AI Generative Model Analysis, AI Pixel-Level Manipulation and AI Content Compliance.
Your policy decides what happens next: approve, step up, route to manual review or decline. NeuroVerify does not approve or reject the proposer or the policy.
How fast is it, and how is it benchmarked?
Latency is under 200 ms p95 on eligible image requests under defined test conditions. The methodology is available on request, so you can compare it with your own traffic.
On the public VendorBench-100 benchmark, NeuroVerify ranked first of 36 systems, with a detection score (MCC) of 0.876 against 0.646 for the next-best vendor.
Where does the media go, and what deployment options exist?
Deployment options are a managed cloud API, private cloud or customer VPC, a secure container, on-premises, and air-gapped. Pick the one that matches your data residency and security requirements.
The platform holds ISO/IEC 27001, ISO/IEC 27701, ISO/IEC 42001 and SOC 2 Type II certifications. For handling details, see product data handling and the trust centre.
Can we test it on our own media before committing?
Yes. Send a representative sample of proposer selfies and video-interview frames across your product lines and compare the scores with your own labels. Include genuine captures from low-end handsets and poor lighting so you can see how thresholds behave on real users, not just on attacks.
Contact the team to set up an evaluation. See also how to evaluate a detection API.
Can it help at claim time as well as at purchase?
Yes, where your claims process collects a face image, video or audio, you can send that media for the same checks. Many teams start at onboarding and extend to claims. See claims, refunds and account recovery.
Evaluate NeuroVerify
See how your insurance selfies score.
Send NeuroVerify a sample of your proposer selfies or video-interview frames and compare the results with your current checks.