What challenges did MANZ face before AI implementation?

Legal research requires reviewing hundreds or thousands of documents and handling specialized legal terminology. MANZ needed accuracy, transparency, current sources, and support for iterative discussions in legal workflows.

What was the impact of the AI solution?

Beta testing with more than 200 experts produced 20% higher accuracy than traditional search tools. The system processed more than one million legal documents. The solution improved search recall and helped users find information faster.

Source

Evidence: Medium50/100

Transforming Legal Analysis: MANZ Integrates AI Research Tools

Use case typeLegal research assistantUpdated Jun 13, 2026

MANZ built an AI-powered legal assistant called MANZ Genjus KI with AWS Partner deepset and Anthropic. The system helps legal professionals conduct complex searches, answer intricate legal questions, generate summaries, and draft contracts. It uses Amazon S3, Amazon OpenSearch Service, Amazon Redshift, and Amazon Bedrock as part of a RAG-backed legal research workflow.

Organization: MANZ
Industry: Legal
Location: Austria
Published: May 2026

Reported outcomes

+20%

accuracyQuality & accuracy

Strategic outcomes

New product / capabilityBuilt an AI legal research assistantNew product / capabilityEnabled complex legal research workflowsBetter decisions & insightImproved search recall and accuracyScale & capacityProcessed over one million legal documents

Why do we believe this?Outcome claims, sources, and evidence checks

Normalized claim

Accuracy: 20% increase

AWS Case StudyMay 27, 2026Case studyInferred claimMedium evidence strength

Beta testing with more than 200 experts produced 20% higher accuracy than traditional search tools.

Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks

Customer: MANZ
Provider: AWS
Maturity: Unknown
Linked source: AWS Case Study

No explicit deployment-stage evidence found.

Customer identity supportedSource describes one deploymentMaturity evidence evaluated

Primary read

Use case focus

Showing 3 of 3

1Legal Research
2Document Search
3Retrieval Augmented Generation

Legal research requires reviewing hundreds or thousands of documents and handling specialized legal terminology.
MANZ needed accuracy, transparency, current sources, and support for iterative discussions in legal workflows.

MANZ and deepset created MANZ Genjus KI as a legal assistant built on the deepset AI platform with Amazon Bedrock as the LLM backend.
Publishing content is stored in Amazon S3, indexed into Amazon OpenSearch Service for vector search, and supported by Amazon Redshift for structured data.
The solution uses OCR, VLM captioning, metadata generation, re-ranking, extraction, summarization, and source verification to support a RAG chatbot.

Technologies

Amazon Bedrock Amazon S3 Amazon OpenSearch Service Amazon Redshift

Beta testing with more than 200 experts produced 20% higher accuracy than traditional search tools.
The system processed more than one million legal documents.
The solution improved search recall and helped users find information faster.

Architecture

MANZ stores publishing content in Amazon S3. deepset processes the content in an event-driven pipeline, indexes data and vectors in Amazon OpenSearch Service, and uses Amazon Redshift for structured analytics. The platform acts as the orchestration layer and calls Anthropic Claude 3.5 Sonnet through Amazon Bedrock for legal search, summarization, contract drafting, and source-verified answers.

Implementation partners2

deepset Anthropic

Sources & evidence1

Evidence: Medium50/100Evidence strength