EyesClear Investigations
Evidence from the public record, quoted and sourced
Resolving an alert takes evidence from two places: the data the bank holds, and the data it does not. EyesClear Investigations is the second half — it screens a company or a person against public sources and returns a report in which every claim carries the quote it came from. It records what sources say and stops there: an allegation is labelled an allegation, and a charge is not a conviction. It is free for financial institutions and their regulators.
01 Why it exists
An alert has two sides. You hold one of them.
Take an incoming payment. The beneficiary is your customer: you have the account, the history, the KYC file, the behaviour over years. The party who sent the money is a name on a message from another bank, and you hold nothing on them at all.
No amount of internal analysis closes that gap, because the missing half was never inside the bank to begin with. It is in the public record — what the counterparty has been reported doing, where it trades, and who with. Somebody has to go and look, and today that somebody is an analyst with a search engine and no audit trail.
Investigations is that half, done properly. It searches sources that are already public and it goes out with a name — never a customer file, never a transaction, never an alert. That is what lets it sit outside the perimeter while the platform stays inside it: the two are joined in the workflow and separate in the architecture, which is the arrangement thesecurity page sets out in full.
How the platform works →Security & LLM governance →
Your data stays inside. Only a name goes out, and only to sources that are already public.
Both sides of the transaction
02 What makes it usable
A finding you can put in front of an examiner
Search is easy and evidence is hard. The difference is whether a finding can still be defended when somebody checks it.
- Every claim carries its quote
- Each relationship and each risk finding is reported with a verbatim quote, and that quote is checked against the page it was taken from before the finding is kept. Anything that cannot be located in its source is discarded rather than reported. Nothing is inferred.
- Reporting, never judging
- A finding records what a source says, with its date, its severity and its provenance. An allegation is labelled an allegation, an investigation is not a charge, and a charge is not a conviction. Where nothing is found, the report says nothing was identified in the sources searched — never that the subject is clean.
- Written to be filed
- The report is built to go into a case file and be read months later by someone who was not there when it was made. Every finding traces back to a source, a date and the words that were actually published.
A file that survives being read by someone who was not there when it was made.
03 Two ways to use it
Complementary to the platform, and useful without it
- On its own
- An analyst screens a name and gets a report back. There is nothing to integrate and nothing to deploy, so a screening can be run on the day the account is opened. It suits onboarding checks, periodic review and enhanced due diligence.
- Inside the platform
- When an alert fires, the public record is often what settles it. Investigations supplies the outside evidence that shows whether an alert is a genuine risk or a namesake, so the case is closed on evidence rather than on an analyst’s recollection.
Start with one screening. Bring it inside the platform when it earns its place.
04 Getting started
Free for financial institutions and regulators
Sign up with your work email. The service is open to people at financial institutions and at their regulators, and the eligibility check reads your email domain — so please use your work address rather than a personal one.
A new account opens with 5 credits: a simple screening spends 1, a detailed one 3.
No integration, no deployment — a screening can be run the day the account is opened.
See it against a name you already know
The fastest way to judge a screening tool is to run it on a case you have already worked and compare what it found. A demo can start there.