Introducing Complead. One AI-native platform for financial crime compliance. Read the story
New Ready Integrations available Check the new integrations
By industry

Screen at signup. Rescreen every 15 minutes.

For banks and neobanks: The check runs inside your own account-opening call, existing customers stay in scope, and every decision leaves a trail your supervisor can read.

Why this matters here

Your programme is judged on the hours you were not looking.

A designation is published at 10 in the morning. Your batch runs at midnight. Every payment cleared in between is a line in the report a thematic review will ask you to explain.

$6 billion

in AML fines paid by banks in 2025

Monzo was fined £21.1 million and Starling £28.96 million, both for controls that did not keep pace with customer growth.

Up to 95%

of alerts are false positives

The system cannot tell two people with similar names apart, so it flags both. More analysts do not fix that. Better matching does.

No trail

when decisions live in inboxes

A name cleared in March. The supervisor asks why in September. The answer is in an email, if the analyst still works for you.

<200 ms Screening API response
15 min From a published designation to your queue
1,500+ Consolidated sanctions, PEP and watchlists
99.95% Uptime
The data layer

The model is only as good as the data under it

A bank is not fined for a slow model. It is fined for a confident decision built on a six-month-old record. We spent six years on the data before we put a model on top of it.

Lists that update as they are published 220+ countries, 3,000+ sources, 1,500+ consolidated lists. A designation published at midnight is in your morning check, tagged with one of 57 risk categories.
PEP networks, not PEP names Built from official gazettes and parliamentary records, country by country. Three risk classes from heads of state to municipal officials, with spouses, children and business partners linked to each name.
Duplicates merged before any list is touched "Mohammad Al-Hussain" and "Muhammad Alhusayn" become one question, not two alerts. Entity resolution at this level cuts false positives by up to 97%.
Four steps, one call

One applicant, from signup to sign-off

Scroll to advance
  1. Step 01 The applicant arrives

    Name, date of birth and address hit your signup endpoint. Duplicates across your core system, card system and CRM are merged before any list is checked.

  2. Step 02 Screened inside your own call

    Sanctions, PEP and adverse media in one query, with the threshold set for retail rather than corporate. The customer sees a loading spinner for less than a fifth of a second.

  3. Step 03 Approved, then watched

    The customer stays in scope. A designation published at ten is in your queue before eleven, and only that change raises an alert, not the whole book again.

  4. Step 04 Routine closed by the agent, the rest by you

    The agent collects the evidence, compares it with the customer's profile and past decisions, writes the reason and closes the routine cases. Roughly three in four never reach a human. The fourth one needs you.

complead / onboarding
  • Applications today4,182
  • Resolved to existing318
  • Median response148 ms
complead / screening / match
  • Matches2
  • Top score98
  • Threshold0.85 retail
complead / monitoring
  • Customers in scope1,284,902
  • Delta alerts today7
  • Next sweepin 9 min
complead / cases
  • Closed by agent today77%
  • Escalated to analyst23%
  • Second reviewerMLRO
Apply
Screen
Monitor
Decide
What your auditor receives

"The model said so" is not an answer a bank can give

Every decision leaves three things behind: The sources it rested on, a written reason in the format your auditor expects, and a log of every step taken to reach it.

01

ISO 27001, GDPR by design

EU data residency available in the contract, not as a workaround.

02

Five-year retention by default

Longer where your regulator requires it.

03

Agent work log

Every source the agent opened, in the order it opened them. Exportable.

04

Parameter history

Who changed which threshold, when, and what the dry run on last month's data showed before it went live.

05

A written reason and a second approver

On every closed case, by policy. Not by habit, not by whoever was on shift.

06

Evidence pack per case

The sources checked, the scores, and the version of each list on the date of the decision.

What you answer to

Who supervises you, and what they actually ask

Supervisors rarely ask whether screening works. They ask how you know it worked on a given date, and where that is written down.

Read the country guides
FATF Recommendations Risk-based approach, customer due diligence, ongoing monitoring, suspicious activity reporting Ongoing
6AMLD and EU AMLR Expanded predicate offences, stricter liability, CDD threshold lowered to €10,000 In force
AMLA (EU) Direct supervision of selected credit institutions, single rulebook reporting From 2027
OFAC and FinCEN (US) Real-time screening against SDN and Consolidated lists, SAR filing on detection 30 days from detection
FCA and national supervisors Documented risk assessment, tuning evidence, parameter history, audit trail Reviewed annually
Customer story · Digital banking

Papara screens every new customer at onboarding without slowing the signup

Papara is one of Türkiye's largest digital banking platforms, licensed by the BRSA. Its onboarding runs at consumer-app speed, so the AML and KYC check had to run inside that flow rather than after it. The customer never waits for compliance.

At signup Screening inside the onboarding call
220+ Countries, one query
Real time List updates reach the check as published
Read the case study
FAQ

Before you ask us

What does the agent decide on its own?

Only what you have configured it to. It is set up on your procedures, your thresholds and your risk appetite, and it closes routine cases with a written reason and a full log. Anything outside its remit escalates to an analyst. Nothing leaves the bank without a human signature where your policy requires one.

Can screening sit inside our account-opening call?

Yes. The API answers in under 200 milliseconds, so it runs as one step of your own onboarding flow rather than a queue after it. Most banks are live in hours. There is no integration fee.

We have several million existing customers. How are they loaded?

One file or one API sync. Duplicates across your core system, card system and CRM are merged on the way in, so a customer held under three spellings is screened once. After that, changes reach your queue the same hour they are published.

What happens to screening while we migrate from our current vendor?

Nothing stops. Complead runs alongside the existing system on the same customer file until you switch the traffic over. Most banks run both for two to four weeks and compare the two queues before cutting over.

Where is customer data stored?

You choose the region: EU, UK or Türkiye. Screening can also run against hashed identifiers if policy forbids sending names.

See it screen your own book

Bring a sample of your customer file. In thirty minutes you will see the matches, the scores, and the audit export they produce.

3,000+ Data sources checked
220+ Countries covered
15 min Always real-time data