Financial crime rarely shows up in only one dataset.
FiveEye is a Turkish-native, multimodal AML and KYC platform. It connects onboarding, sanctions screening, ongoing KYC, transaction monitoring, adverse media and investigation workflows around the same customer — so a compliance team sees one risk context instead of six disconnected alert queues.
Four harmless signals are sometimes one obvious case.
An identity record, a partial name match, a payment pattern and an adverse-media mention each look survivable on their own. Reviewed in separate systems by separate people, they stay that way. Resolved onto the same entity, they are frequently the case the team was supposed to find.
Six checks, resolved onto one entity.
Every one of these exists somewhere else as a product. Running them against the same resolved customer, account and counterparty is the part that is hard to buy.
Identity, ownership, PEP, sanctions and adverse-media checks brought into one consistent onboarding and periodic-review workflow.
Live sanctions and watchlist screening that accounts for Turkish spelling variation, suffixes and transliteration rather than failing on them.
KYC, PEP, sanctions, adverse-media, behavioural and transaction signals combined into an explainable customer risk view.
Patterns such as structuring, smurfing, velocity and layering detected in real time, beyond static rule checks.
Turkish risk-language models applied to news, social and darknet signals to surface exposure before it is public consensus.
Evidence routed into an auditable analyst case, with a MASAK STR draft prepared — while the final judgment stays with the compliance officer.
Many signals enter. One explainable case leaves.
The loop is the product. Any single check in it is available elsewhere; resolving all of them onto the same entity is not.
Onboarding and identity, KYC and ownership, sanctions and watchlists, live transactions, adverse media and darknet intelligence.
Signals are attached to the same customer, account and counterparty, across spelling and transliteration differences.
A compliance risk view is produced with the contributing signals and their weights attached to it.
Alerts are ranked by context rather than by rule count, so analysts work the cases that matter.
An analyst case carries the full evidence trail, the actions taken and who took them.
A MASAK STR draft is assembled from the case; a person decides whether it is filed.
Engineering figures, stated with their caveat.
These are product specifications and engineering targets supplied by TeamSec, not measured customer outcomes. Production results depend on configuration, infrastructure and data quality.
The control layer, on the compliance side.
FiveEye clears a counterparty before credit work begins and keeps watching afterwards. It is not the credit-scoring engine — that is Scoring Engines, and the two answer to different regulators.
Built with local domain knowledge and an explicit evidence trail.
The development brief includes academic advisory frameworks with Hacettepe University for data security and darknet research, and METU for blockchain and cryptography. It is supported under the TÜBİTAK 1501 industrial R&D programme and designed around FATF, MASAK, BDDK and KVKK requirements.
Human approval stays final
The system prepares, ranks and evidences. It does not file, and it does not close a case on its own.
Every alert carries its reason
The signals and weights behind a score are exposed for investigation, internal audit and regulatory review.
SaaS or your own infrastructure
Deployed in a local SaaS environment or on the institution's own infrastructure, according to its control and data-residency model.
The ones that come up first.
Who is it for?
Regulated institutions that cannot treat compliance as a bolt-on: development, participation and investment banks; payments and e-money institutions; MASAK-regulated crypto-asset service providers; factoring and leasing companies; and other obliged entities with suspicious-transaction obligations.
What does Turkish-native actually mean?
Name matching, transliteration, entity resolution and risk-language models are built around Turkish names, institutions and local compliance context rather than translated from an English-first product. In screening, that is the difference between a match and a miss.
Does it decide anything on its own?
No. It ranks, evidences and prepares — including the STR draft — and the compliance officer decides. That boundary is a design constraint, not a limitation.
Is it the same as the credit-scoring engine?
No. FiveEye covers AML, KYC and financial-crime risk. Credit risk is Scoring Engines and Early Warning. They share the platform and nothing else.
Bring the alert burden. See the whole context.
The demo worth running is your own false-positive queue.