Overview zapAgent Team Werner + Wilma
Overview of zapAgent Team Werner + Wilma Functionality
1. General Introduction
The Agent Team Werner + Wilma combines two specialized agents to support SAP data analysis from data extraction to business-oriented interpretation of results.
Werner extracts SAP data via RFC, prepares it in a structured format, and performs indicator-based or AI-based analyses. Depending on the selected Content Bundle, these analyses can be executed as one-off or continuous runs. All data processing takes place within zapAnalytics. For details, see Product Description zapAgent Werner.
Wilma builds on Werner’s results and supports auditors in analyzing and interpreting findings within their business context. Using a LibreChat interface, users can explore results, ask follow-up questions, and generate insights based on predefined prompts and analytical tools.
Both agents operate exclusively within your locally installed zapHome instance. A prerequisite for using Team Werner + Wilma is the technical setup of zapHome OnPrem AI.
This article provides an overview of how to work with the agent team and links to all relevant content.
2. Technical Onboarding
The technical foundation for this agent team is zapHome OnPrem AI. It is deployed via Docker on a Linux-based virtual machine or server. zapHome OnPrem AI provides:
- Local execution of all agents within your IT environment
- Full data privacy (no external data access, no SaaS applications)
- SAP data analysis and seamless integration with your Azure OpenAI Services
The following article provides an overview of the platform and includes links to all relevant technical topics, such as VM/server setup: General Setup zapHome OnPrem AI
3. Functional Onboarding
This section shows how the agent team supports your day-to-day work - from data extraction to analyzing and interpreting results.
3.1 Project Setup (Depending on Content Bundle)
The specific project setup is defined according to the selected Content Bundle. Key organizational aspects include:
- Objective: What is the goal of the project?
- Data Scope: Which data needs to be analyzed (SAP client, company codes, fiscal years) and what analysis should be performed?
- Timeline: What are the start and end dates of the project, and which milestones are planned?
- User Access: Who needs access to which tools?
- zapAnalytics
- Power BI
- LibreChat
3.2 Data Extraction and Analysis
Werner performs the core analysis in zapAnalytics. Key steps:
See: Content Bundles
- User uploads Content Bundles that define extraction scope, analysis logic and result formats
- Werner executes data extraction and performs analysis
See: Content Bundles
3.3 Analysis Results
Once Werner has completed data processing, the results can be utilized in two ways: via structured Power BI dashboards for visual analysis, and through Wilma's interactive, chat-based interface for dynamic exploration.
Power BI (Reporting & Visualization)
Data is exported as source files (.parquet), which are used to create dashboards in Power BI. This allows you to explore results in a structured, visual format.
Wilma (Interactive Analysis)
Wilma connects directly to the result database via MCP. You can interact with the results via chat, e.g. to ask questions, explain findings, and derive conclusions. This allows you to explore and interpret data specifically within your business context.
3.4 Available Content Bundles
The Agent Team supports Content Bundles for the following solutions:
- Cash Recovery AI (CRAI)
- ICS Assessment
- Digital Audit Preparation (DAP)
- Finance Benchmarking (FB)
4. Governance Onboarding
General governance-related topics include:
- Data protection
- Information security
- Access management
Solution-specific governance information is available in the respective subsections.