spektr agents

Every screening alert researched and ready for sign-off

False Positives AI investigates sanctions, PEP, adverse media, KYB and UBO change alerts, then hands your analyst a cited recommendation to review and sign off.

Book a demo Explore interactive demo
PEP screening hitOpened from monitoring
Position historyCompared with your customer
Likely false positive3 differing fields cited
Ready for sign-offAssigned to your analyst

True match or not

Checks position history and public disclosures to judge whether the flagged person is really your customer.

Context, not just names

Maps associates and business affiliations to show how serious a real match is.

Your analyst decides

The agent recommends. A person reviews each alert and signs off, with the reasoning logged for audit.

How it works

1

Input

A screening hit from onboarding or monitoring opens an alert for False Positives AI.

2

Investigate

The agent compares the flagged entity with your customer, checking history and relationships across your connected sources.

3

Recommend

It sorts the alert by severity and recommends a resolution with cited evidence and a confidence score.

4

Sign off

Your analyst reviews the recommendation and confirms or escalates; no alert closes without a person.

Sanctions hit PEP match Adverse media Position history Public disclosures ComplyAdvantage
Recommended resolutionSeverity rankingCited evidenceConfidence scoreDecision loggingHuman in the loop
Verification

Is this really the same person?

Name matches create noise. False Positives AI checks position history and public disclosures against your customer record, so likely mismatches arrive with the reasons spelled out.

Alert #4821PEP screening · Monitoring
PEP match · Lars EriksenScreening provider · name match 96%
Open alert
Screening hitLars Eriksen
PEP list · NO
Full name Lars Eriksen
Date of birth 12 Mar 1958
Nationality Norwegian
Position Deputy Minister
Years active 2009 – 2017
Your customerLars Eriksen
Customer · DK
Full name Lars Eriksen
Date of birth 04 Sep 1986
Nationality Danish
Position Software engineer
Years active No public role
vs
Likely false positive · 3 differing fieldsDate of birth, nationality and position differ
Confidence91%
Severity

Real matches ranked by risk

When a match holds up, the agent sorts it by severity using thresholds you set, with links to the news and regulatory disclosures behind it. Your team works the riskiest alerts first.

Alert queueScreening · Monitoring
6 open
Sorted by newest
Halden Logistics AS 12 min agoE24 · 30 Oct
Adverse media Low · 22
Mikael Strand 40 min agoStortinget register
PEP Medium · 64
Vireo Trading GmbH 1 h agoEU sanctions list
Sanctions High · 92
Anna Berg 2 h agoRiksdagen disclosure
PEP Low · 31
Brightwater Ventures Ltd 3 h agoReuters · 2 Nov
Adverse media High · 88
Jonas Vik 5 h agoOFAC notice
Sanctions Medium · 58
Your thresholds: High ≥ 80 · Medium ≥ 50 Set by your team
Audit

Every recommendation traceable

Each recommendation links back to its data source with an explanation your auditors can follow. Policies are version-controlled, and rule changes can be tested before they go live.

Suggested Likely false positive Alert #4821 · Lars Eriksen
SourceStortinget registerstortinget.no/…/eriksen
SourceCustomer recordKYC onboarding · 2 Mar 2024
Explanation

Hit born 1958, Norwegian minister. Customer born 1986, Danish engineer.

Policyv14 · version-controlledTested before going live
Signed off by Sofie Holm · 14:32

Works with

KYC OnboardingForm
Process KYC Onboarding
MonitoringProcess
Process Monitoring
AI agent KYB AI
AI agent Network Discovery AI
Integration ComplyAdvantage
"We need an infrastructure that can automate the routine work without losing the configurability each jurisdiction demands, whilst meeting increasingly complex regulatory demands."
Read case study
Mark Nelson
Head of Financial Crime Controls, at Nets Denmark
"We weren't looking for a platform that would force us to rebuild around it. We needed something that could plug into how we already operate and remove friction where it actually lives."
Read case study
Roshni Bharadwaj
Group Product Manager at Pleo
“spektr turned our most complex processes into a flow anyone can follow - bringing structure and scalability to something that used to rely heavily on internal know-how.”
Read case study
André Guardado
Head of Legal and Compliance at Santander Leasing
Welcome back, John 128,564 4,849 6,301 1,596 9,252 Try out the demo Explore the spektr platform yourself. Open the demo

FAQ

Does False Positives AI close alerts on its own?

No. It investigates each alert and recommends a resolution with evidence. Your analyst reviews it and makes the decision, and every step is logged.

Which alert types does it handle?

Screening alerts across sanctions, PEPs, adverse media, KYB and UBO changes.

How does it keep up with regulatory changes?

Policies sit in a version-controlled policy engine. Your team can test a rule set before activating it, updates go live without downtime, and every change is recorded in the audit trail.

Is our data used to train AI models?

No. Client data is never used for training and is kept only as long as needed. spektr holds ISO 27001 and ISO/IEC 42001 certification, and personal data is processed in the EU.

Fewer false positives. Same human sign-off.
Investigates screening alerts and recommends a resolution, so analysts review evidence instead of gathering it.
4.8
Book a demo