Now in early accessLearn more

Agents do the work. Humans make the call.

cmpliance is the agentic workforce for compliance. Agents screen, classify, gather evidence, and assemble the decision pack. Humans review, approve, and sign off.

app.cmpliance.ai

Assessments

3 entities
New assessment
All active2Awaiting decision2Escalated0All3
Search entity, hash, or run id…
Kontora Risk Systems GmbHEU AI Act
#2
72
In reviewLena Brandt
#1
55
DecidedLena Brandt
Lewenis Analytics SEEU AI Act
#1
44
DecidedLena Brandt
Vendara Onboarding ABEU AI Act
#2
58
In reviewMikael Holm
#1
39
DecidedMikael Holm
Assessments·EU / Germany

Kontora Risk Systems GmbH

In review
AI-assisted customer risk scoring module · reviewed for Westmark Bank SE
AssessmentEU AI Act vendor and high-risk AI system readiness review
Summary
Evidence
Trace
Review history
System recommendation
ESCALATERequest vendor conformity evidence before approval.

A high-risk AI system influencing onboarding risk decisions lacks bias-testing and human-oversight evidence; escalate to the compliance lead before production use.

Reviewer actionRequest additional vendor evidence before approval; escalate to compliance lead because the system affects onboarding risk decisions.
Entity risk score
0
High risk
03467100
Key risk factors
High-risk AI for customer risk scoring· EU AI Act Annex III
Bias testing not segmented by customer segment· EU AI Act Art. 10
Override audit log omits reviewer rationale· EU AI Act Arts. 12/14
Human-oversight review evidence missing· EU AI Act Art. 14
Assessments·EU / Germany

Kontora Risk Systems GmbH

In review
AI-assisted customer risk scoring module · reviewed for Westmark Bank SE
AssessmentEU AI Act vendor and high-risk AI system readiness review
Summary
Evidence
Trace
Review history
Agent traceEU AI Act5 steps· 10sView run output
Run started
May 18, 2026
Classify system purpose & deployment context
0.6s
Map vendor docs to AI Act controls
3.9s
Check human-oversight & override evidence
2.7s
Review model docs & scoring evidence
2.2s
Prepare missing-evidence list
0.8s
Run completed
May 18, 2026
Trace running…RUN SHA · 7f3a91c2
Assessments·EU / Germany

Kontora Risk Systems GmbH

In review
AI-assisted customer risk scoring module · reviewed for Westmark Bank SE
AssessmentEU AI Act vendor and high-risk AI system readiness review
Summary
Evidence
Trace
Review history
Evidence pack
Hash-chained and signed
Download
Generated May 17, 2026Proofpack evp…7f3a91c2View trace
ArtifactsKontora KRS-2.4 AI system description · Model version history and change log KRS-2.4 +7
MissingNo signed conformity-readiness attestation from vendor · No current bias-testing appendix for version 2.4 +2

Extracted facts

· 1
01
Override audit-log sample
Decision rationale
RecordedAudit-log exportRetrieved May 17, 2026View source Kontora Risk Systems GmbH
RationaleThe audit-log sample shows customer risk scores overridden without a recorded reviewer rationale; insufficient logging and human-oversight evidence under EU AI Act Articles 12 and 14.
Assessments·EU / Germany

Kontora Risk Systems GmbH

In review
Summary
Evidence
Trace
Review history
Decision timeline
Reopened
09:15
Decided → Reopened·Review agent
Evidence review assigned
10:40
Reopened → In review·LBLena Brandt
Escalation prepared
15:05
In review → In review·LBLena Brandt
Every transition hash-signed and recorded.
Your call
ApproveRejectEscalate
Escalate this assessment?

This records the decision for Kontora Risk Systems GmbH and seals a signed, hash-chained evidence pack. Reopening requires a new review cycle.

CancelEscalate
Assessment escalated
01Search02Summary03Trace04Evidence05History06Decision

Pick up where you left off. Queue sorted by urgency.

Compliance teams were never meant to gather evidence.

TFP-2026-0414Dutch KVK extract
OCC-v3.2Ownership & control chart
SPS-2026-0414PEP & sanctions screening
AMS-2026-0414Adverse-media summary

They were meant to make decisions.

Decision recordedFLAG
sha256:5d77cb56…91e · hash-signed
1 reviewer · Anna Keller · routed to second review

cmpliance · the agentic workforce for compliance

The shift

From compliance signal to accountable action.

Agents prepare the work. Humans approve the decision. Every step leaves proof. cmpliance turns alerts, document gaps, screening hits, and regulatory changes into assigned work, evidence trails, reviewer gates, and proofpacks.

  1. 01
    Unowned signals

    Signals arrive faster than anyone can turn them into owned action.

    We know. We've built and delivered systems that do exactly that. cmpliance replaces manual, fragmented processes with agents that prepare the work end-to-end.

    Agentic resolution

    Owned work with reviewer gates

  2. 02
    Fragmented evidence

    Your evidence is scattered across tools that don't talk to each other. Every one of them has a license and somebody maintaining it. The tacit knowledge isn't documented anywhere.

    cmpliance connects the numbers, the documents, and the context in a single pass, and seals it into tamper-evident packs built for regulatory review.

    Agentic resolution

    Audit trails built for review

  3. 03
    Deadline pressure

    The EU AI Act timeline may move. The obligation does not.

    Risk classification, conformity assessments, and technical documentation still need evidence a reviewer can defend. cmpliance assembles the evidence pack and preserves the proof trail.

    Agentic resolution

    Evidence packs you can export and defend

Three compliance lifecycles. One agentic workforce.

Agents prepare the repetitive work, preserve the proof trail, and hand decisions back to compliance teams.

How it works
Intake lifecycle

KYC/AML onboarding

Submit a counterparty. Agents screen, classify risk, gather evidence from multiple sources, and assemble a complete evidence pack. You review the decision and approve. From intake to audit-ready output.

screening run00:18
Sanctions + PEP checks4/4 clear
Risk classificationHigh
Source evidence linked12 records
OutcomeEvidence pack assembled and ready for reviewer approval.
Review lifecycle

Decision review and sign-off

Signals become assigned work, evidence checkpoints, reviewer gates, and decision records. Every step is logged, so the proofpack shows not just the result, but how the decision was reached.

decision record00:26
Reviewer gates checked3/3
Decision records sealed3
Proof trail completeYes
OutcomeFull run captured with end-to-end traceability.
Regulatory lifecycle

EU AI Act documentation

Agents classify your AI system's risk level, gather the required documentation, and prepare an evidence pack with regulatory references for reviewer approval.

doc assembly00:34
AI system risk tierHigh-risk
Required docs gathered9/9
Regulatory mappingsAnnex IV
OutcomeEvidence pack prepared with regulatory references for reviewer approval.

How it works

Detect the signal. Assign the owner. Prepare the evidence. Route the review. Capture the decision. Export the proof.

Get early access

INPUT

EntityNordic Components GmbH
JurisdictionEU / CH
AssessmentKYC/AML onboarding

AGENTS WORKING

Screening agent

3 sources matched

Classification agent

Risk score calibrated

Evidence agent

6 references collected

Narrative agent

Rationale assembled

YOUR CALL

Risk levelHigh
ReviewerM. Fischer

EVIDENCE OUTPUT

Evidence Pack · EP-2024-0847

Hash0x7f3a...c2d1
Timestamp2026-02-11 14:08 UTC
DecisionApproved with controls
Sources6 verified references

One case. Agents work. Traceable evidence. Your call.

What agentic actually means

Every compliance vendor says agentic now. Most mean faster alerts and better dashboards. That’s workflow software, not agency.

cmpliance agents reason, classify, gather evidence, and draft the narrative. They produce a complete, tamper-evident evidence pack. The compliance officer makes the call.

Deployed like software. Calibrated like a team member.

cmpliance runs as a subscription, but each tenant builds its own decision memory. Every decision your team makes, every approval, every flag, every override, teaches the system how you do compliance. Your risk appetite. Your jurisdictions. Your standards of evidence. Your reasoning patterns.

cmpliance calibrates to your organization’s evidence standards, escalation thresholds, jurisdictions, and review decisions. This improves future assessments for the customer without training a generic vendor model on customer data.

Architecture

The model layer changes. The structural intelligence stays here.

cmpliance combines language-model reasoning over documents with graph reasoning over entity structures, in a shared embedding space. Frontier models are interchangeable.

Model layer · interchangeable
Anthropic ClaudeOpenAI GPTGoogle GeminiOpen-source models
Structural reasoning · cmpliance

Frontier models

Reason over documents

filings & reportsregistriesadverse medianarrative drafts

Shared multi-dimensional vector space

Text references and structural connections live in the same coordinate system. A textual mention can resolve to a structural link three hops away.

produces

Agent reasoning · evidence pack

Decision · Sources · Trace · Hash

Graph layer

Reason over structures

ownership chainsdirectorshipsnominee linksflow patterns

Why agents find what others miss

Same counterparty. Two systems. Different answer.

A name-match system runs a counterparty against sanctions and watch lists and returns a verdict. cmpliance reasons over the structure between entities.

Example assessment

method · graph + LLMIn review

Swipe or tap to compare systems

counterpartyidentical
Nordic Components GmbH
reasoning depth
1 hop · direct lookup
signals checked
Name string · fuzzy match
sources
OFAC · EU · UN
finding
No match on counterparty name
adverse media
No flags
audit trail
Not exposed by this class of system
evidence hash

reasoning path

1 hop · direct lookup
  1. counterpartyNordic Components
  2. name matchsanctions listOFAC · EU · UN
counterpartyidentical
Nordic Components GmbH
reasoning depth
3 hops · ownership chain
signals checked
Ownership · nominee registry · cross-jurisdiction directorship
sources
Corporate registers · sanctions lists · directorship graph
finding
Nominee for a sanctioned individual
adverse media
2 flags · ownership opacity, directorship overlap
audit trail
ep_2026_0218_0944
evidence hash
sha256:7f3a…c91d

reasoning path

3 hops · ownership chain
  1. counterpartyNordic Components
  2. ownsholding co.low-transparency
  3. directornominee directorlisted on 3 boards
  4. nominee forsanctioned individualterminal
  • Evidence-first

    Every decision includes sources, checks, and an audit trail.

  • Structure-aware

    Understands ownership links and cross-jurisdiction relationships.

  • Audit-ready

    Immutable audit trail. Exportable evidence packs.

Name matching finds a name. cmpliance finds the structure.

Virtual data room

Evidence stays inside its boundary.

Share evidence packs for review and verification while preserving access logs, source references, and hash-chain integrity.

  1. 1Document enters roomLogged on arrival with an access entry
  2. 2Agents processScreened, classified, evidence gathered
  3. 3Proof Pack assembledDecision pack with sources and trace
  4. 4Hash anchoredTamper-evident hash written to the ledger
  5. 5Auditor verifiesChecked against the audit ledger
Verification00:00.4s
Documents in room47
Access events126
Hash anchors47/47verified
Recent access3 of 126 · last 7 days
External auditordownloaded packEP-2026-02182h ago
M. Fischeropened packEP-2026-02175h ago
Legal counselopened packEP-2026-02141d ago
OutcomeAll evidence packs verified. Zero tampering detected.
Evidence sharingShare evidence packs across reviewers and auditors.
VerificationEach pack verifiable against the audit ledger.
Access logsEvery view and download recorded automatically.
Source referencesEvery claim points to a stored source record.
Hash-chain integrityTamper-evident continuity across the run timeline.

Compute scales anywhere. Secrets stay put. Only hashes and commitments cross the boundary.

M&A due diligenceRegulatory submissionsCross-border compliance

Early access

Evidence-first
agent labor.

We are not launching a product. We are launching the agentic workforce for compliance.

For teams that have to defend decisions. Agents prepare the work and preserve the proof trail. Humans make the call.

Request early access
Calibration depth
Live signal
DeepMidThinCold0102030405Joining order
Pioneer cohort
Compounding signal · live
Tap to pin

Why early access matters

The first companies in compound first.

Calibrates future assessments
Does not train a vendor model on your data

Calibration is cumulative. The earlier a company joins, the more the system understands their entity landscape, their escalation thresholds, their jurisdictions, and their reviewers. That depth does not transfer to later adopters. Early access is a position you claim, not a waitlist you join.

Pricing

Per-seat tools scale with people. Regulation scales with volume.

Per-seat pricing solves for the slow line. Agent labor pricing solves for the steep one. cmpliance meters the compliance work delivered, not the people who log in.

capacity / costtimethe gap per-seatpricing cannot close
headcount · slow, linearregulatory surface & volume

Pricing

We don't sell seats. We sell agent labor. The system meters the compliance work actually delivered, not the people who log in. Traditional seat-based tools can cost €50K–200K/year per organization. We scale with throughput.

First assessment

For compliance officers evaluating cmpliance

CHF0
  • One full assessment (KYC/AML or EU AI Act)
  • Evidence pack with audit trail ID
Get early access

Switch plans or cancel anytime

Recommended

Compliance team

For teams running ongoing KYC and EU AI Act assessments

CHF2'500/mo
  • Metered agent labor for ongoing KYC and EU AI Act assessments
  • Assessments, evidence packs, and documentation delivered, not licensed
  • Agents apply your compliance policies consistently
  • Continuous monitoring, not one-off exports
Get early access

Switch plans or cancel anytime

Full department

For organizations needing department-scale agent capacity and dedicated support

CHF5'000/mo
  • Department-scale metered capacity across your compliance function
  • Dedicated compliance graph
  • Custom review paths and policy configuration
  • Priority support and onboarding
Get early access

Switch plans or cancel anytime

Enterprise

For regulated industries with complex compliance structures and multi-jurisdiction requirements

Custom

  • Usage-metered across jurisdictions
  • Private infrastructure and data residency
  • SSO, SCIM, and audit log integrations
  • Dedicated compliance engineering support
  • Custom SLA with guaranteed response times
Get in touch

Switch plans or cancel anytime

2 of 4 · Compliance team

Transaction screening

Agentic screening across the compliance graph.

Private banking

Regulatory documentation prepared end-to-end for compliance team review.

Evidence infrastructure

Deterministic evidence packs with hash-anchored audit trails.

Get in touch →

Selected regulated-industry deployments. Details available under NDA.

The EU AI Act timeline may move. The obligation does not.

Swiss hosted · AES-256 · Hash-chained audit trail · Your data is never used for training

Insights

Featured articles

Evidence-led notes on audit readiness, control design, and keeping accountable humans in control of agentic compliance.

Common questions

You choose. cmpliance offers data residency in Switzerland, the EU, and the United States. All data is encrypted at rest and in transit. Your compliance data never leaves the region you select.

KYC/AML onboarding under 4AMLD, 5AMLD, and 6AMLD. EU AI Act risk classification and documentation under Articles 6–9 and Annex III. We’re actively expanding coverage; if you need a specific framework, get in touch.

No. Your data is never used to train or fine-tune any model. Agent outputs are generated per-run using your data in isolation. You own everything that goes in and everything that comes out.

All infrastructure runs on encrypted, access-controlled environments with full audit logging. Every agent action is traced and every evidence pack is hash-chained for tamper detection. Our controls are designed to support SOC 2 Type II requirements.

cmpliance is built by paterhn.ai, a Swiss AI development studio with production AI systems in operation since 2001, multi-agent systems running in enterprise environments since 2017, and transformer-era agent work in production since 2019.

One assessment is a complete compliance run: KYC/AML onboarding or EU AI Act documentation, including all screening, scoring, and evidence generation for a single entity or system.

Agents prepare compliance assessments in parallel. Higher tiers give you more concurrent agent capacity, so your team can process more assessments without waiting.

Yes. No lock-in contracts. Switch plans or cancel from your dashboard. Annual plans are refunded pro-rata for unused months.

Upgrade instantly from your dashboard. Your compliance graph, decision history, and evidence packs carry over; nothing is lost.

Per-seat pricing charges for humans in chairs. cmpliance charges for agent labor: the actual compliance work being done. Your whole team can use cmpliance; you pay for throughput, not headcount.