AI automation and data science for controlling and finance

Module 1: AI basics, RAG and LLM applications (2 days, live online)

• Introduction to Large Language Models (LLMs) - with a focus on real practical cases from finance, controlling and risk management.
• From the basics to the application: What is a prompt? How does prompt engineering work? How does it create controllable AI workflows?
• Step-by-step implementation of modern RAG workflows (Retrieval Augmented Generation) for the financial sector:
  • Document upload, automated chunking, embedding, vector search and semantic search in your own data.
  • Practical application with invoice PDFs, contracts or other financial documents.
• Introduction and use of no-code tools such as LangFlow for fast, visual process automation.
• Visualization and comparison: No-code flows vs. Python code - understandable for everyone.
• Understanding how to apply AI-supported document analysis and Q&A systems to your own data.
• Group and individual exercises: From the initial idea to a functioning AI solution in your own context.
• Theory & practice of the automation framework:
  • Introduction to a tried-and-tested framework for identifying, evaluating and designing AI automation projects.
  • Workshop: Application and practice phase with realistic examples from finance/controlling.
  • Direct transfer: After the module, you will recognize suitable processes and design your first projects.

Goal: You not only understand the technology, but also build your first AI-supported financial workflows and have a real toolbox for practical use.

Module 2: AI automation & no-code/Python integration (1 day, presence)

• Hands-on with AI tools for finance, controlling & accounting:
  • Development of own reporting and accounting assistants (e.g. for double-entry bookkeeping, automated evaluations).
  • Reporting workflows with SQL agents, charting agents and writing agents (e.g. automatically generated reports with graphics).
  • Automated fraud and anomaly detection using AI - visually with LangFlow, but also "under the hood" with Python integrations.
• Combination of No-Code and Python:
  • Non-programmers can also take part - professionals can go deeper if necessary.
  • Live demo: Integration of Python nodes in LangFlow for more complex logic.
• Application in everyday business: typical finance automation projects.
• Automation framework (deepening & application):
  • In-depth exercises: Selection and outlining of concrete automation projects that can be implemented with agents or AI workflows.
  • "From idea to proof of concept" - developing your own roadmap for implementation with modern AI technologies.

Goal: You develop your own automations, combine no-code and Python solutions and are ready to optimize processes in the financial sector with the help of AI.

Module 3: Advanced Agent Frameworks & Self-Improving AI (1 day, presence)

• Modern multi-agent systems with LangGraph, AutoGen and Python:
  • Overview: What are AI agents and how are they used in the finance/controlling context?
  • Practice: Building self-reflecting reporting agents that check and optimize their own results and correct errors independently.
  • Forecast agent frameworks with self-evaluation: time series analysis and automatic validation of forecast quality (e.g. cash flow or sales forecast).
• Simulation & live tests:
  • Develop, test and improve your own AI agents.
  • What does "self-reflection"/"self-evaluation" mean in practice and how do you use it for robust AI automation?
• Practical benefits: Automated, robust AI workflows that not only perform tasks but also continuously improve themselves.
• Automation framework in the agent context:
  • Methodical roadmap: From process analysis to proof-of-concept for complex automation with agents.

Objective: You will learn about the latest agent technologies, build your own self-improving workflows and be ready for the next level of AI automation.

Module 4: Practical transfer & automation framework (self-learning phase)

• Direct transfer to corporate reality:
  • Analysis and selection of automation potential in your own company using a structured AI automation framework.
  • Step-by-step: Identification, evaluation, prioritization and conception of your own AI automation project.
  • Development of an individual AI use case including business benefits, resource estimation and implementation outline (functional & technical).
  • Application of the framework for concrete planning of the technical implementation, including interfaces, data preparation, integration of LLMs, machine learning and automation tools.

Goal: You develop an implementable automation concept that can create immediate value in your company.

Module 5: Presentation & optimization of your own AI projects (1 day, live online)

• Presentation of your individual use cases:
  • Presentation of your project, including cost-benefit analysis, workflow, technical and organizational implementation.
  • Peer and trainer feedback: discussion and fine-tuning with the group.
  • Joint sharpening of practical projects for maximum company benefit.
  • Implementation recommendations and next steps for implementation in everyday working life.

Goal: You will receive valuable feedback, strengthen your implementation skills and complete the program with a sustainable automation project that delivers real added value for your company.