Manual due diligence costs €5,000–15,000 per artwork, takes 4–8 weeks, and still leaves gaps. MIRAS gives you patented authentication and quantified risk scores — in a single platform, at a fraction of the cost.
The Problem
41% of insurers report rising fraud claims on fine art policies. Without biometric proof of identity, you are underwriting a photograph and a paper certificate — not the physical object.
Authentication alone costs €1,500–15,000 per work. Add KYC (€1,500–3,500 per review), AML checks, and provenance research. A single policy can require 4–8 weeks of manual work before you can even quote.
Post-claim disputes rely on photographs, paper provenance, and expert opinion. There is no computational, independently verifiable proof that the claimed object and the insured object are the same physical thing.
The Cost of the Status Quo
How MIRAS Works for Insurers
Six photographs from a standard smartphone. The patented computer vision pipeline generates a unique biometric fingerprint from the artwork’s physical surface — texture, cracks, pigment patterns. Unforgeable.
Patent #1 · 18 claimsThree scoring engines analyse 26 risk variables: Client (9), Artwork (9), Transaction (8). Dynamic weighting via 10 contextual modifiers produces a composite score (0–100) with four risk bands.
Patent #2 · 17 claimsAutomated checks against OFAC, EU/UN sanctions lists, Art Loss Register, and Interpol stolen works database. Provenance gaps and wartime-period exposures (1933–1945) flagged automatically.
A structured compliance report documenting authenticity verification, provenance depth, risk scoring, and screening results. Machine-readable (JSON-LD) for integration with your actuarial systems via API.
Portfolio Scenario
A mid-size fine art insurer processing 20 new policies annually. Manual costs based on verified industry data.
Savings depend on artwork complexity and existing vendor contracts. MIRAS subscription pricing replaces all four cost lines above. Contact us for a customised cost comparison.
Protected by Two Patents
Patented computer vision combining pHash, DINOv2 transformers, and Gabor texture analysis. Generates an irreplicable 1.5KB fingerprint from the artwork’s physical surface. For insurers: proves the insured object and the claimed object are the same physical thing.
The art market’s first computational AML risk model. Three engines (Client, Artwork, Transaction) analyse 26 variables with 10 contextual modifiers. Produces a composite score (0–100) with four risk bands. For insurers: quantifies compliance risk before you underwrite.
Join our founding network of institutional partners. Request a pilot to see how MIRAS transforms your underwriting workflow.