18 real German court decisions · BAG & LAG — The benchmark cases are actual judicial decisions, presented through ARAM AU research summaries. Some summaries remain subject to source review. View court sources
German employment law · court-grounded workflow

Assess AU evidentiary value. Know what to review next.

Structure the facts, evaluate relevant indicators and counter-indicators, and prepare a legally traceable next step.

A practical, court-grounded workflow for German HR, Labor Relations and employment-law counsel, referencing 18 BAG and LAG benchmark decisions.

01 Can the AU still be relied on? 02 Are there legally relevant doubts? 03 What should happen next?

Explore 18 real BAG and LAG court decisions, with ARAM AU research summaries, source references and visible verification status.

German employment lawBAG & LAG casesSource traceabilityRole-specific workflowsLocal browser processing
Structured assessmentAuditable
01
AU certificateStart with the certified period and document.
02
Relevant factsKeep facts, assertions and documents separate.
03
Legal assessmentCheck patterns, counter-indicators and Gesamtbetrachtung.
04
Determine evidentiary stateHIGH · UNDERMINED · MIXED BY PERIOD · OPEN
NextIdentify the evidence, escalation or legal action required for the selected role.
Boundary: the tool assesses evidentiary weight. It does not decide whether an employee is “really sick.”
The practical workflow

Four clear steps. One traceable legal assessment.

A simple introduction to the full ten-stage professional worksheet.

01

Describe the AU case

Record dates, documents, relevant facts and party positions.

02

Review both sides

Assess indicators and counter-indicators together.

03

Understand the result

See the reasoning, authority, evidentiary state and scope.

04

Prepare next action

Identify evidence gaps and appropriate professional escalation.

For deeper review, the Case Portal retains its full ten-stage workflow and ARAM AU Knowledge Graph.

Why this matters

An AU certificate carries significant evidentiary weight — but not in every factual constellation.

The relevant question is not whether one fact looks suspicious. The question is how the entire factual picture interacts with German employment-law principles and case law.

01

What happened?

Capture chronology, documents, work events, medical context, and the positions of both parties.

02

What matters legally?

Compare the matter against recognized patterns, counter-indicators and evidentiary principles.

03

What follows?

Determine whether the AU retains its weight and what additional proof, escalation or legal action may be required.

What the assessment checks

A structured review of the circumstances that may matter.

T

Timing & termination

Employment-end knowledge, exact alignment, AU sequencing, immediate new employment.

M

Medical reliability

Assessment method, doctor contact, medically substantive guideline issues, diagnosis context.

C

Conduct during AU

Travel, physical or leisure activity, compared with the employee’s actual contractual work.

W

Actual work during AU

Substantial contractual work, repeated work episodes and certification-capacity inconsistency.

↔

Counter-indicators

Facts that weigh against serious doubt, including treatment and non-comparable activity.

P

Independent proof

Symptoms, functional impact, treatment, physician evidence and substantiation after undermining.

Clear boundaries

What the tool does not do.

Restraint is part of the legal model.

It does not determine whether an employee is “really sick.”
It does not treat unusual behavior as automatic proof of work capacity.
It does not equate an undermined evidentiary value with fraud or misconduct.
It does not replace legal review or predict court outcomes with certainty.
It does not turn an employer allegation into an established fact.
See the reasoning in one case

A legal conclusion should never feel like a black box.

One benchmark case shows the full path from facts to legal consequence.

BAG 5 AZR 149/21

Exact timing between resignation and AU

The AU began on the same day as the resignation and covered the complete remaining employment period.

CASE FACTResignation + AU on the same day
→
PATTERNExact notice-period coincidence
→
LEGAL TESTSerious doubt in the Gesamtbetrachtung
→
RESULTUNDERMINEDAU alone no longer sufficient
What this demonstrates: the tool does not label the employee “healthy” or “dishonest.” It explains why the certificate lost evidentiary force and what proof question follows.
Benchmark decisions

Built around 18 BAG and LAG decisions.

View all cases →
For professionals

One legal engine. Five professional views.

The underlying legal result remains the same. The terminology, depth, guidance and next action change by role.

Explainability

Every conclusion should be traceable.

The product keeps source facts separate from legal reasoning and allows users to follow the path from fact to authority to outcome.

See the method
Case fact→Pattern→Rule→Authority→Assessment→Outcome
Controlled legal workflows

Local, auditable, and built for review.

In the current browser-based version, case data remains in the browser. Users can export a structured case record without a case database.

Local browser storageExportable JSON recordFacts preserved separatelySource traceability
Start with the facts

Run a benchmark case or create a new matter.

Choose a role, work through the 10-stage legal process, and see how the same result is translated for different professional audiences.

Open benchmark cases
Built for your role

One legal framework. Five useful professional views.

HR / Personalbereich

Prepare a consistent, documented AU case before legal escalation.

Labor Relations / Arbeitsbeziehungen

Keep chronology, records and handoffs structured.

In-House Counsel

Review evidentiary facts, legal authorities and burden questions.

Employer Association Counsel

Prepare source-grounded advice for a member employer without overstating the holding.

External Counsel

Evaluate a traceable record, disputed facts and unresolved legal questions.

Method

From case facts to a traceable legal outcome.

One factual model, one legal model, one assessment engine, five professional views.

1

Capture facts

Record AU periods, documents, termination events, work events, evidence and party positions.

2

Identify legal patterns

Compare the matter against benchmark-derived temporal, medical, conduct and capacity patterns.

3

Check counter-indicators

Record facts that weaken or neutralize the challenge rather than searching only for suspicious facts.

4

Gesamtbetrachtung

Consider the overall factual constellation rather than applying a one-fact automatic rule.

5

Determine evidentiary state

HIGH, UNDERMINED, MIXED BY PERIOD, or OPEN — with plain-language explanation and source traceability.

6

Identify next action

Translate the legal result into a role-specific employment-law workflow step.

CASE GRAPHWhat happened?
+
STANDARD GRAPHWhat rules apply?
→
ASSESSMENT GRAPHWhat does it mean?
→
KNOWLEDGE GRAPHWhat becomes reusable and traceable?
ARAM AU Knowledge Graph

See why the assessment reaches its result.

The Knowledge Graph turns an assessed case into a clear legal reasoning map. It connects facts, legal standards, authorities, evidentiary state and the limits of what may be concluded.

18benchmark decisions BAG + LAGauthority levels preserved EN / DEcontrolled legal terminology
ARAM AU Knowledge Graph
Central legal questionAU Evidentiary ValueCurrent assessment + reusable legal meaning
Designed for professional use:HR sees the operational meaning. Legal users can open the same graph with deeper authority, burden and scope detail. The underlying legal result stays identical.
One case, explained end-to-end

BAG 5 AZR 149/21

UNDERMINED
CASE FACTAU covers the exact remaining employment period
→
CHALLENGE PATTERNExact temporal alignment
→
LEGAL RULEConcrete circumstances may create serious doubt
→
AUTHORITYBAG 5 AZR 149/21
→
ASSESSMENTSerious doubt threshold met
→
STATEUNDERMINED
→
NEXT CONSEQUENCEIndependent proof must be assessed
Inference boundary: UNDERMINED does not establish that the AU was false, that fraud occurred, that the employee was healthy, or that dismissal was justified.
18 benchmark decisions

The legal foundation of the assessment.

Compare the recurring legal questions behind the cases — timing, medical reliability, conduct and demonstrated work capacity.

TIMING & TERMINATIONTemporal patternsTermination knowledge, exact alignment, AU sequences, leave adjacency
MEDICAL CERTIFICATIONReliability & processAU-Richtlinie, doctor contact, retrospective and foreign certification
CONDUCT & CAPACITYFunctional comparisonTravel, activity and actual contractual work during AU periods
PROOF, BURDEN & SCOPELegal consequenceIndependent proof, dismissal burden, open issues and scope boundaries
Five user journeys

One legal engine. Five realistic professional workflows.

Each role is asked only for decisions that belong to that role. The legal engine remains consistent underneath.

Legal sources · 18 benchmark decisions

Authority and provenance remain visible.

All 18 benchmark records are shown with court level, topic, evidentiary status and source links. Some records require further judicial-source review.

BENCHMARK CORPUS18 decisions · 6 BAG / 12 LAG

This is a research and demonstration corpus, not a certification of all 18 judgments. The source layer preserves authority hierarchy and research status. A LAG application is never presented as a BAG-level general rule.

About ARAM

Structured legal workflows for explainable employment-law analysis.

ARAM AU Evidentiary Analysis connects case facts, curated legal rules, German court authorities and role-specific workflows without hiding the reasoning path.

LEGAL FOUNDATION18 benchmark decisions

The product is grounded in the complete 18-case benchmark corpus used across the website, Legal Sources and Case Portal.

Built for professional review

The product separates facts from allegations, distinguishes BAG authority from LAG application, preserves open questions, and keeps evidentiary analysis separate from misconduct or fraud conclusions.

One engine, five professional views

HR, Labor Relations, In-House Employment Counsel, Employer Association Counsel and External Employment Counsel work from the same facts and legal model. Wording and workflow depth adapt to the role; the underlying legal result does not.

Local processing model

Case data entered in the browser-based version is stored locally unless the user exports it. No case database or account is required.

Professional pilot

Explore ARAM AU with a controlled professional pilot.

We are preparing structured practitioner evaluation for German HR and employment-law teams. This demonstration does not collect pilot applications online.

Professional pilot / practitioner feedback

Use fictional or effectively anonymous input for your own demonstration matter. The 18 benchmark decisions are real court judgments. Review the decisions, evaluate the workflow with your professional team, and document usability and legal-quality feedback.

Contact pathway: Enquire about professional pilot participation by email at mano@aramalgorithm.ai. Do not include identifiable employee or medical records in your message.

Legal notice · Impressum

ARAM AU — Website provider information

Operator information for the independently presented ARAM AU Evidentiary Analysis website. Separate website · Shared operator.

Privacy · Datenschutz

Privacy information for ARAM AU

Data protection information for this Netlify-hosted AU demonstration.

Data processing · Datenverarbeitung

Local-first AU demonstration

Understand case-data entry, local retention, exports and the boundaries of this public demonstration.

Governance & Trust

Traceable legal reasoning. Human responsibility.

AU-specific decision-support boundaries, authority provenance and review discipline.

Assessment workspace

Case Overview

Saved locally← Zur Website
Hinweis zur Demoversion: Bitte nur fiktive oder wirksam anonymisierte Falldaten eingeben. Eingaben werden im Browser gespeichert; JSON-Exporte können Ihre Angaben enthalten. Keine personenbezogenen Gesundheitsdaten eingeben.