AI Support
AI support where expert decisions are made
ANYVA uses AI as professional support – integrated into processes, requirements, risks, and TOMs. Responsibility for decisions remains with the user.
Function 1
Plausibility justification
An LLM checks technical assessments for comprehensibility and consistency. It doesn't replace a decision – it provides a second level of review.
Risk assessment
- Is the risk assessment comprehensible?
- Does the justification match the chosen risk level?
- Are there contradictions in the evaluation?
TOMs and Measures
- Are the measures appropriate?
- Are obvious TOMs missing?
- Are all requirements fully covered?
DSFA-Plausibility check
- Are risks fully captured?
- Is the need for protection justified?
- Are countermeasures appropriate?
Audit preparation
- Are the proofs complete?
- Are there gaps in the audit trail?
- Are reasons documented?
Function 2
RAG-powered expert system
ANYVA uses Retrieval-Augmented Generation (RAG) to specifically integrate existing specialist knowledge into the work context – not as a generic chatbot.
Embedded context
Own processesTechnical ServicesAssetsTOMsRequirementsRisks
RAG knowledge base
Tom's LibraryRequirements catalogueRisk scenariosRisk moduleStandardsInternal guidelines
AI Support
Generate suggestionsPlausibiliseCheck consistencySupport groundsSpotting gaps
Basic principle
Support instead of replacement
ANYVA-KI does not make decisions. It supports professionals in making better and more understandable decisions.
Exam
Reviews checked for plausibility and completeness.
Suggestions
AI suggests TOMs, requirements, or actions – the user decides.
Understandability
AI suggestions are documentable and auditable as support notes.
AI support in practice
In a demo, we'll show how plausibility checks and RAG support are integrated into everyday work.