Claims platforms move submissions through workflow. They were not built to ask whether the evidence inside those submissions is authentic. Veritura is the forensic layer that fills that gap, returning a single scored evidence package on every claim, every policy application, and every renewal.
Image forensic ensemble. Document intelligence. Asset and identity verification. Cross-channel intelligence. Returned through one REST integration, in seconds, into the platforms and processes already in place.
Veritura runs every engine on every submission. The four capabilities below operate together on each claim, application, and renewal. They return a single risk score, the weighted composition that produced it, and the forensic artefacts attached to each finding.
The same forensic stack runs at three distinct moments in the policy lifecycle. At claim, the engine analyses the evidence accompanying a loss notification, including photographs, repair documents, and statements, before the claim is approved. At renewal, the engine re-analyses prior submissions against current models, surfacing claims whose evidence is now known to be manipulated even if it passed review at the time. At policy inception, the engine analyses the photographs and documents in the application itself, refusing the policy when the evidence the cover would be priced against is synthetic. Fraud blocked at inception or renewal eliminates the claim entirely. That is the structural shift this stack is built around.
Every submitted image runs through a dual-model forensic ensemble: eight methods spanning classical computational forensics and modern neural detection. Manipulated photographs, AI-generated submissions, and reused images surface as scored signals before a claim is approved or a policy is bound.
Repair estimates, police reports, claimant statements, and registration documents are analysed for internal consistency and cross-checked against each other and the rest of the submission. Contradictions on falsifiable facts such as dates, amounts, named parties, and declared workshops surface as structured signals.
VIN and plate data extracted from photographs, validated against national vehicle registries and the EUCARIS network. For property and commercial lines, addresses cross-referenced against cadastral registries; equipment serial numbers validated against declared inventories; contractor details verified against trade registries.
Every submission is fingerprinted on entry and indexed in a carrier-scoped intelligence pool. Reused images, documents, and identifiers surface across channels, whether direct app, broker portal, or bancassurance, without personal data crossing any boundary and without creating new sub-processor relationships.
Veritura returns a 0–100 risk score and the weighted composition that produced it. The score lands in one of three bands, and each band routes to a specific action. The handler always makes the final call.
Claim proceeds through the standard workflow. No required action from the handler. No second screen.
The handler reviews the composition and decides: approve, request further information, or escalate. Veritura provides the evidence; the handler provides the judgement.
The handler refers to the SIU. The case file is assembled and attached to the referral automatically: artefact images, signal log, API response payload, chain-of-custody audit trail. The investigator opens an already-prepared file.
Veritura is designed to the standards that high-risk AI systems are held to in the EU: explainability, auditability, and a defensible record. That those standards apply to fraud detection is our judgement, not the regulation's. The composition behind every score is itemised. The artefacts are retained. The audit trail is immutable. Veritura returns evidence to the handler, never a verdict in place of one.
Most carriers already run fraud detection systems: behavioural fraud databases, network analytics platforms, and case management workflow tools. These operate on transactional, identity, and relational data, and they are good at what they do. None of them analyses photographic or documentary evidence for manipulation or fabrication. That gap is what Veritura fills.
Cross-carrier identity and behavioural pattern matching against shared submission corpora.
Answers: is this submitter on record?
Relationship mapping between people, vehicles, addresses, and businesses across claims.
Answers: who is connected to whom?
Investigation lifecycle management, evidence orchestration, referral routing.
Answers: where in the process is this case?
Per-submission analysis of every image and document for manipulation and fabrication.
Answers: is the evidence real?
The systems are additive. Veritura's high-risk referrals can travel directly into existing investigation tools through a structured case-export API, so the SIU investigator opens the case with the forensic file already attached.
Veritura integrates through a single REST API, compatible with the open API models of the leading P&C claims platforms and with direct insurer-system connection. There is no platform migration. There is no second screen for the handler. The forensic score and composition render inside the existing claims interface.
The same API call handles claim submissions, policy inception submissions, and renewal submissions. The forensic stack runs identically across all three; only the workflow integration point changes. Veritura is designed for the open API models of the leading P&C claims platforms, and for direct insurer-system integration where no major platform is in place.
The API response is structured for handler-rendered display: a single risk score, the weighted composition, the per-engine outputs, the forensic artefact references, and the elapsed time per engine. Everything required to render the finding inside the existing interface, no UI components shipped.
Veritura hosted. Veritura operates the infrastructure and assumes full regulatory responsibility: data processing agreement, transfer controls, retention obligations. This is the deployment model for the pilot and for carriers where managed-service operation is preferred.
On-premise at insurer. The Veritura software and models are deployed inside the insurer's own infrastructure perimeter. No data is transmitted to Veritura systems. The insurer operates the deployment; Veritura provides the software licence, model updates, and support. No new sub-processor relationship. No new cross-border transfer. Data residency and GDPR posture remain entirely the carrier's.
Veritura is the forensic layer between submission and approval. Every claim, every application, every renewal is analysed by every engine in the pipeline. The handler keeps the decision. The SIU receives a case-ready file. The audit trail is immutable.
A thirty-minute scoping call. A sample claim cohort. API access.
We bring the rest.