12/09/2025 - Articles
Use cases for the AI assistant in BCS: From help to automation
Artificial intelligence is not only changing the way we develop software, it is also changing the way we use software. For BCS, our project management and ERP software, this means that AI should make everyday work noticeably easier, reduce routine tasks, guide users to answers more quickly, and increase internal efficiency. While many AI initiatives focus on one-off solutions or experimental prototypes, Projektron pursues a systematic, long-term approach.
AI in BCS: Where it brings real benefits
After an intensive research phase on the fundamentals of AI, RAG (retrieval-augmented generation), prompt engineering, fine-tuning, and framework design, one thing was clear: to integrate AI meaningfully into BCS, concrete, recurring use cases are needed in which AI provides real benefits.
RAG is particularly crucial here, as AI first finds relevant documents or data and uses them to generate answers, thereby minimizing hallucinations. We have documented more about our experiences with RAG in a separate blog article.
The AI use cases in BCS can be roughly divided into three categories:
Help and information provision
Support in ticket and project management
Automation of tasks via AI agents
These applications meet key criteria: They occur frequently, are well structured, have clearly definable requirements, and can be implemented in a security-compliant manner.
Even in the early stages of development, it became clear that it is not the language model alone that determines quality, but above all how tasks are defined and what context is provided. For this reason, Projektron relies on a modular framework that can be operated both locally and scalably. Individual components such as the language model, embedding model, or vector database are easily interchangeable in order to be able to respond flexibly to rapid developments in the field of AI.
An overview of the use cases
Application | Status | |
|---|---|---|
| 1. | KI user help / FAQ: Answering the user’s exact question. | productive |
| 2. | Summarize Tickets: Summarizing problem descriptions spread across (possibly) many comments from (possibly) multiple people. | productive |
| 3. | Solution Suggestions for New Tickets: Search for similar tickets, generate a response proposal, and refine it in chat mode. | in testing |
| 4. | Summaries of Additional Lists: Entries in the sales history for customers or persons can be summarized, as well as other lists such as the project progress log. | 2nd Quarter 2026 |
| 5. | Queries via AI Agents, Tools & MCP: In the future, user questions will be routed to a central AI agent that uses “tools” to answer them. The AI software assistance is one such tool; various BCS functions are also tools. These are accessed through the MCP interface. If the user asks how to create a vacation request, the AI agent answers based on documentation. If the user asks how many vacation days they still have this year, the AI agent retrieves the data directly from BCS via MCP. | in development, approx. end of 2026 |
| 6. | Language Versions: Based on a single-language dataset, many output languages are possible. | productive |
| 7. | Local AI Framework: To operate AI functions fully locally for high security requirements, customers must install the BCS AI framework on-premise. Additional components such as the vector database, the embedding model, and local language models (Gemma, Llama, Qwen, ...) are included. We are working on packaging to make this system environment manageable, including documentation and training. | in development, approx. mid-2026 |
| 8. | Actions via AI Agents, Tools & MCP:
| in development, approx. mid-2026 |
A growing AI ecosystem
The use cases show that the AI assistant is much more than a search function. It forms a growing ecosystem of
assistance
automation
intelligent system queries
This ecosystem will take on more tasks step by step. BCS users benefit from
faster responses,
less routine work,
better overview,
higher automation, and
full data sovereignty.
At the same time, the assistant is growing modularly, can be operated locally including the language model, and is constantly being optimized through testing.
For more in-depth information about the development processes, optimizations of the RAG process, parent document retrieval, query rewriting, and our experiences with language models, please refer to our other blog posts on the topic of AI.
All blog articles on the main topic of AI: AI knowledge (1-4) and AI at Projektron (5-8)
About the author
Maik Dorl is one of the three founders and remains one of the managing directors of Projektron GmbH. Since its founding in 2001, he has shaped the strategic direction of the company and is now responsible for sales, customer service, and product management. As product manager, he is the driving force behind the integration of innovative AI applications into the ERP and project management software BCS.
More interesting articles on the Projektron blog
Product management at Projektron
How does software remain successful for 25 years? Projektron BCS shows that continuous updates, user feedback, and modern technologies ensure long-term success. Learn how product management works at Projektron.
AI-Help in BCS
Since version 25.3, the new BCS AI user help has been providing precise answers to questions about Projektron documentation. The article shows how iterative optimizations in retrieval and splitting have significantly improved the quality of responses.
AI-Framework for BCS
Projektron is developing a flexible AI framework for BCS that can be operated entirely locally and meets the highest standards of precision, data protection, and transparency. The article shows the technical and organizational foundations required for the framework and how the architecture is structured.
Comparison: ERP for Service Providers
For service providers, choosing the right ERP solution determines efficiency, transparency, and revenue. Our comparison highlights the best systems, explains which features truly matter, and provides valuable tips for making the right choice. Discover the ERP that will really drive your business forward!
An AI bot in front of symbolized BCS applications – visualizing Projektron's long-term AI strategy and the growing range of productive AI functions.
