VerIQ
Due-diligence terminal
Automated forensic triage for global capital markets: KYC packets, SWIFT messages and commodity logistics documents read, cross-referenced and scored in seconds.
Due diligence in commodities and correspondent banking means reading MT103, MT760 and MT799 messages, tank storage receipts, bills of lading and inspection reports by eye, then checking names against Companies House, OpenCorporates and KVK one browser tab at a time. VerIQ runs that as staged due diligence with GPI payment tracking and IMO vessel lookups built in. Most onboarding teams re-key data from PDFs, search registries by hand in a browser, and store the proof in a folder nobody can audit. VerIQ collapses that into one case file — extraction, deep search across official sources, ownership resolution, scoring and monitoring — and keeps the reasoning for every decision.
Visit the live site — myveriq.com< 90s
Median case
3
Staged DD levels
MT103/760/799
SWIFT parsed
The live product
VerIQ as it ships today
A screenshot of the running terminal — staged due diligence, SWIFT parsing and commodity document checks on the front door.

Walkthrough
From upload to cleared case
Step through the five stages of a case. Each view is a representative mockup of the production interface, including the applicant's phone app.
veriq.cases / VQ-40218
document extraction — 6 fields, 1 flag
Step 2 of 4
Photograph your passport
Applicant capture app
The applicant uploads once; the model does the typing
Passports, licences, incorporation certificates, utility bills and structure charts are read by a vision model. Fields are extracted, cross-checked against each other, and any inconsistency is raised before the case moves on.
- Multi-language document extraction with per-field confidence
- Liveness and face-match against the submitted photo ID
- Tamper and template checks on the document image itself
- Applicant chased automatically when something is missing or unreadable
AI & automation
Where the models actually earn their keep
AI is used where it is measurable and reviewable, and kept away from the final call on anything material.
Document understanding
Vision models extract structured fields from any layout, with confidence per field rather than a single pass/fail.
Entity resolution
Fuzzy name, transliteration and alias handling so the right registry record is found without guesswork.
Adverse media triage
Articles summarised, relevance-ranked and de-duplicated so reviewers read four items, not four hundred.
Policy engine
Deterministic rules make the decision; the model only supplies inputs. That keeps decisions explainable.
Human in the loop
Anything above your risk threshold goes to a reviewer with the evidence pre-assembled.
Full evidence chain
Raw API responses, model outputs, scores and reviewer notes stored together per case.
Build one like it
We can build you a system comparable to VerIQ
Any business with an onboarding or compliance queue can run this pattern: banks, brokers, payment firms, lenders, insurers, marketplaces and trading desks. We adapt the document set, the registries and the policy matrix to your regime, then plug it into whatever you already use for cases.
Free quote & initial consultation