AI Help
Answers specific questions about using BCS and its features based on the documentation and displays relevant sources.
AI FEATURES FOR PROJECT MANAGEMENT AND ERP
Artificial Intelligence in BCS
Understand information faster. Reduce routine work. Directly within the work context.
The AI features in BCS support your employees wherever information is created and used: in the software help, tickets, tasks, projects, requirements, and contact histories. Instead of copying content into a separate AI tool, users receive summaries and context-aware answers directly within their project management and ERP processes. This reduces the time spent searching and reading, simplifies handovers, and makes existing knowledge available faster.
What if your employees could use BCS to …
- receive a direct, source-backed answer to a specific question about using the software?
- understand the current status of a long ticket history without reading it in full?
- identify more quickly what has been completed and what remains open during a task handover?
- ask questions directly about a task or contact history without explaining the context again?
- review the sources used and systematically evaluate AI-generated results?
- use AI without manually copying project and ERP content into a separate tool?
AI Features in BCS
An overview of which AI features are already available, which are currently being tested, and which Projektron is developing.
Available
Ticket Summaries
Summarize the subject, description, and comments to provide a concise overview of the issue, processing status, and previous history.
Task Summaries
Condense task descriptions, checklists, comments, and history information to simplify onboarding and handovers.
Task Assistant
Answers questions about an individual task and takes the designated task information into account.
Project-Related Summaries
Present selected project information in a concise format, including history logs, project manager notes, and project-related contact histories.
Requirement and User Story Summaries
Condense extensive requirements and provide faster access to the subject matter documented in user stories.
Summary of External Organization Contact Histories
Summarizes relevant customer contacts from the past twelve months and identifies the filters used.
AI-Powered User Interface Translation
Translates user interface text from German into supported languages and simplifies the provision of international work environments.
In Testing
Solution Suggestions for New Tickets
Searches for similar tickets that have already been processed, generates an initial response suggestion, and allows users to refine it in the chat.
In Development
Extended Task Chat
Multiple tasks will be selectable for joint analysis of dependencies, similarities, and open points affecting several tasks.
Source References in Chats
Answers will reference the underlying checklists, comments, history log entries, and other BCS information.
Contact History Chat
Adds targeted questions about agreements, discussed topics, developments, and outstanding feedback to the existing summary.
Ticket Chat
Will enable targeted follow-up questions about a ticket and the information documented within it.
Additional Summaries
Extend AI-generated summaries to additional lists, history information, and other work areas in BCS.
Further Chat Improvements
The operation and presentation of chats are being further improved. Planned enhancements also include support for voice input.
Queries via AI Agents, Tools, and MCP
A central AI agent will answer documentation questions and retrieve specific information from BCS.
Actions via AI Agents, Tools, and MCP
AI agents will support selected processes in BCS, such as submitting leave requests, creating risks, or reviewing bookings.
Local BCS AI Framework
An installable package with local AI components, documentation, and training options will support additional deployment models.
The plans, scope, and availability of features that have not yet been released may change during development.
1. AI Help: Answers to Questions About Using BCS
AI Help provides conversational access to the extensive BCS documentation. Users ask a specific question and receive an immediately usable answer with source references.
From the Question to the Right Documentation
Users open AI Help with a single click. The question is processed using a RAG approach: First, BCS performs a semantic search for relevant documentation content. The language model then generates an answer based on this content.
Answers to specific questions about using BCS and its features
Illustrated step-by-step instructions provided as an answer
Source references with links to the online help
Answers generated in real time
Option to rate the answer as helpful or not helpful
Easy copying of the question-and-answer pair, including sources
Easy creation of a support ticket if the answer is not sufficient
Option to open AI Help in a separate window for parallel work
Intentionally Limited in Scope
AI Help is not a general-purpose knowledge AI. It answers questions covered by the available BCS documentation. This keeps the support focused on BCS-specific topics and allows users to verify the answer using the cited sources.
2. AI Summaries: Understand Complex Processes Faster
AI summaries condense existing information into an overview that is quick to understand. The original content remains fully intact and can be opened at any time for professional review.
Take Over and Process Tickets Faster
The ticket summary condenses the subject, description, and potentially numerous comments from different participants. This makes it easier to understand lengthy support and service histories and enables smooth handovers.
Overview of the issue, processing status, and key history
Notification when a summary may no longer be up to date
Update, delete, and regenerate summaries
Rating from “Very good” to “Very poor”
Optional comment on the rating
Support for handovers, escalations, and reopened tickets
Get a Structured Overview of Tasks and Project Histories
In complex tasks and projects, relevant information is often distributed across different areas. Summaries generated by BCS AI combine master data, descriptions, checklists, comments, and history information into a concise overall view.
Faster onboarding for new or reassigned tasks
Overview of open items and documented progress
Summaries of selected project-related information areas
Analysis of project notes and history logs
Update, delete, and rate generated content
Understand Requirements and User Stories Faster
Even extensive Scrum requirements in the Products workspace can be condensed using an AI-generated summary. This helps users understand the subject matter more quickly and provides easier access to the documented content.
Concise overview of extensive requirements
Faster onboarding for new or reassigned user stories
Support for review, coordination, and further processing
Summarize Customer Contact History at a Glance
The AI-generated summary condenses the customer contact history from the past twelve months. This enables account managers to identify documented conversations, feedback, and developments more quickly.
Summary of customer contacts from the past year
Meaningful filtering of contact history entries
Display of the filters used in the result
Replacement of the existing summary when it is regenerated
3. AI-Powered Translation of the User Interface
BCS can use AI to translate user interface text from German into supported languages. This simplifies the creation of international work environments and accelerates the translation of new or modified content.
4. Context-Aware Chats: Ask Questions Directly About BCS Content
Context-aware assistants combine selected BCS data with a conversational interface. Instead of receiving only a static summary, users can ask specific follow-up questions about the documented information.
Task Assistant
Already available: The Task Assistant answers questions about a task and takes the relevant task information into account.
Example questions
- Which items are still open?
- What was changed most recently?
- What information is still required to process the task?
Review Multiple Tasks Together
In development: Users will be able to select multiple tasks together to examine dependencies, similarities, and open items.
Example questions
- Which tasks are awaiting feedback?
- What dependencies exist before the next milestone?
- Which open items affect multiple tasks?
Contact History Chat
In development: In addition to the summary, a chat will allow users to ask specific questions about previous customer communications.
Example questions
- Which agreements were documented most recently?
- Which topics have been raised repeatedly?
- Which responses are still outstanding?
Roadmap plans, schedules, and feature details may change during development.
5. Solution Suggestions for New Tickets
For new tickets, BCS is intended not only to summarize existing information, but also to actively support users in finding a solution. The feature currently being tested searches for similar tickets and uses them to generate an initial response suggestion.
From a Similar Ticket to a Relevant Response Suggestion
- Find similar tickets BCS searches for comparable tickets that have already been processed.
- Generate a response suggestion Existing solution knowledge is used to generate a possible response.
- Refine the suggestion The generated text can be revised and improved in the chat.
Your Benefits
- Use existing solution knowledge more quickly
- Process recurring support and service cases more efficiently
How AI Makes Existing Knowledge in BCS Available Faster
The AI features reduce the time spent searching and reading, simplify handovers, and provide information directly within the relevant work context.
Less Time Spent Searching
AI Help answers specific questions based on the BCS documentation and displays the sources used.
Faster Handovers
Summaries make it easier to understand the status of tickets, tasks, and project histories, even when responsibilities change.
Better Overview
Important content, open items, and developments are condensed for meetings, reviews, and further processing.
Work Within the Context
The AI features are integrated directly into BCS objects. There is no need to copy content into a separate tool.
Traceable Results
Source references, the underlying original content, and feedback features make professional review easier.
Deployment Tailored to Your Needs
AI components can be provided in line with the edition, license, processes, and internal company requirements.
The BCS Context Makes the Difference
BCS connects AI with the information designated for the relevant work process.
AI Support Directly Within the Work Process
The AI features use the relevant BCS content, such as tickets, tasks, requirements, comments, checklists, and contact histories.
Summaries condense information, assistants answer follow-up questions, and AI Help provides access to the BCS documentation. This keeps the support within the relevant work context.
Use Existing Data More Effectively
AI provides an additional way to access information that already exists. Complex processes can be reviewed more quickly and prepared for handovers, meetings, or further processing.
Well-maintained descriptions, checklists, and comments remain the basis for helpful answers and summaries.
How Different Roles Benefit From AI in BCS
The AI features deliver value wherever employees receive, share, and evaluate information in their daily work.
Project Managers
Review project notes, contact histories, and history logs more quickly, prepare task handovers, and identify open items for status meetings. Summaries help users classify extensive project information more quickly.
Team Members
Review task content, understand checklists, comments, and history logs, and ask specific questions about the processing status. The Task Assistant provides answers directly within the context of the relevant task.
Support Staff
Take over lengthy tickets more quickly, understand their previous history, and generate an updated summary after relevant changes. This makes complex support processes and handovers easier to understand.
Account Managers
Condense the contact histories of external organizations and use the Contact History Chat to ask specific questions about agreements, discussed topics, or outstanding feedback. This feature is expected to be available with BCS 26.3.
Administrators
Activate AI components as needed and adapt their use to the edition, license, user roles, processes, and internal security, compliance, and governance requirements.
Managers
Obtain an initial overview of extensive project, task, and communication histories more quickly. Key figures, original data, and the professional assessment of those responsible remain decisive for business decisions.
AI in BCS for Project Management and ERP: Technology, Data Protection, and Security
Projektron develops the AI applications and their integration into BCS in-house. The modular framework connects selected BCS data sources with suitable AI components and provides a reliable foundation for high-quality responses, controlled processing, and the needs-based development of the features.
The right BCS context for precise AI support
Each AI function uses the BCS content intended for the respective use case. This provides the language model with the appropriate specialist context for answers, summaries and follow-up questions.
AI Help
AI Help searches semantically for relevant content in the BCS documentation. The documents found are provided together with the question as the basis for the answer.
Summaries
Summaries take into account the content intended for the respective process, such as subject, description, comments, checklists or history logs.
Context-related chats
Chats refer to selected tasks, contact histories or other designated BCS content. Questions can therefore be asked directly about the respective work process.
High answer quality through targeted data preparation
The quality of an AI answer depends both on the language model and on the selection and preparation of the information provided. It is crucial that relevant information is found and processed in its professional context.
Projektron develops and tests various methods for this purpose. AI Help has, among other things, been tested using realistic, previously resolved support questions. Weaknesses observed in retrieval and data preparation are incorporated into further technical development. The methods used particularly support functions that semantically search extensive content and assemble it for an appropriate answer.
Parent Document Retrieval If a short text section is identified as relevant, the complete parent document can be provided. This gives the language model additional professional context.
Query Rewriting Search queries can be adjusted specifically so that relevant specialist terms receive greater weighting and suitable content can be found more reliably.
Adapted text splitting Configurable section lengths and overlaps help preserve semantic relationships in extensive and structured documentation pages.
Supplementary search documents Short document summaries can bundle key terms and provide additional search anchors. The complete original document remains authoritative for the answer.
Controlled AI support within the work process
Users deliberately initiate answers, summaries and chats within the respective work process. The AI functions described do not independently analyse all company data and do not make automatic changes to projects or master data. The results are provided for further processing and professional review.
Defined data sources Each function uses the content intended for its use case. Which data is included depends on the specific BCS process.
Activation as required AI modules can be provided selectively and adapted to edition, licence, roles, processes and internal requirements.
Original data remains available Generated content supplements comments, checklists, history logs and other original information. These remain fully available in BCS.
Supported quality review Source references, update notices and ratings make it easier to assess generated results professionally.
Data protection and responsible use of AI
Which information is processed depends on the AI function being used. Companies can additionally govern its use through their own requirements for roles, permitted inputs, review processes and responsibilities.
Use-case-related inputs Free-text input fields are intended for questions relating to the respective work process. Passwords and unnecessary sensitive information should not be entered.
Consider trade secrets Existing internal company rules and approval processes can be applied to confidential content.
Clear organisational rules Internal policies can define which functions may be used, which content is intended and how results are to be reviewed.
Traceable context Processing remains focused on the data sources and BCS content intended for the respective function.
Secure operation and a reliable framework
BCS has been developed in Berlin since 2001. The software can be operated within the customer's own infrastructure or in German data centres.
Projektron operates an information security management system in accordance with ISO/IEC 27001 and is TISAX-assessed. These structures provide the organisational framework for information security, regulated responsibilities and reliable processes.
The AI applications described currently use a language model from Mistral AI, which is operated on a server of Projektron GmbH.
BCS can be operated within the customer's infrastructure. The AI applications currently described use a language model operated on Projektron GmbH's infrastructure. Modular architecture and flexible AI components
The framework separates applications, BCS data sources and technical AI components from one another. The language model, embedding model and vector database can therefore be developed, supplemented or replaced independently.
The architecture provides the basis for integrating additional models and operating modes when they offer advantages in terms of quality, data protection, security or cost-effectiveness.
A fully locally operated AI solution is part of the long-term development of the AI functions in BCS.More About AI in the Projektron Blog: Background Information and Insights
The articles follow the development of the AI features in BCS and explain both specific applications and the underlying technology.
How Can AI Enhance BCS?
An introduction to the development project: initial technical foundations, potential use cases, and objectives for integrating AI into BCS.
An AI Framework for BCS
The article explains the requirements, modular architecture, and interchangeable components behind the AI applications.
AI Help in BCS
How AI Help answers questions about the BCS documentation and how it was improved through retrieval, testing, and feedback loops.
Use Cases for AI in BCS
An overview of specific applications for AI support and the gradual expansion of the features in BCS.
Tokenization in AI
How text is divided into smaller units and why this segmentation affects how language models process it.
Vectorization in AI
How words and content are converted into numerical vectors, making them comparable for semantic searches.
Attention in AI
How language models identify and weight relevant terms and relationships within a text.
RAG in AI
How language models are connected to defined knowledge sources to provide subject-matter context and make answers more traceable.
Frequently Asked Questions About AI in BCS
Answers about the features, use cases, technical foundations, data protection, and availability of AI support in BCS.
What are the answers provided by AI Help based on?
AI Help uses the RAG approach – retrieval-augmented generation. The user’s question is first analyzed semantically. Relevant content is then identified in the BCS documentation.
The relevant documents are provided to the language model together with the question as context. This is intended to ensure that the answer is specifically based on documented information about BCS.
The sources used are displayed and can be opened directly in the online help for professional review.
Can the AI answer general knowledge questions?
AI Help is designed for questions about using BCS and its features. It does not replace a general-purpose chat or knowledge search.
Context-aware assistants process the BCS content designated for the relevant process. For example, the Task Assistant refers to the selected tasks and the information documented within them.
Are AI-generated answers and summaries always correct?
No. Errors, omissions, or incomplete interpretations can occur even when documentation- and source-based methods are used.
Answers, summaries, and suggestions should therefore be reviewed carefully. This is particularly important when they are used as the basis for decisions with professional, financial, contractual, or legal relevance.
Are the AI features activated automatically in the background?
Answers, chats, and summaries are deliberately initiated by the user. The user decides when a summary is generated, updated, or deleted and when a question is submitted to an assistant.
BCS can indicate that a previously saved summary may no longer be current because of new or modified content.
How is data protection addressed?
The AI features use defined data sources and activatable BCS components. The content processed depends on the feature being used.
The language model used is provided by Mistral AI. Companies can manage its use through components, licenses, roles, and organizational requirements.
Does the AI run entirely on-premises?
No, the AI does not currently run entirely on-premises. AI Help uses the RAG approach and accesses a language model from Mistral AI that is hosted on a server operated by Projektron GmbH.
In the long term, Projektron intends to offer a fully local solution in which no data is transferred to language models hosted by third parties.
Operating BCS on-premises does not currently mean that all AI components are automatically operated within the customer’s infrastructure.What data is used for a ticket summary?
For a ticket summary, BCS primarily processes the ticket’s subject, description, and comments.
The summary is displayed and saved directly in the ticket under “Summary (AI).” It can be updated, rated, or deleted.
How does BCS identify a potentially outdated summary?
For tickets, BCS can display a notice when professionally relevant content has changed since the summary was created.
This includes changes to the ticket type or ticket status, as well as new external comments or team comments. The user can then update the summary.
What is the difference between AI Help and the Task Assistant?
AI Help answers questions about using BCS and its features based on the BCS documentation.
The Task Assistant, by contrast, refers to information from specific tasks. It supports context-aware follow-up questions and can consider multiple selected tasks together.
Can I rate the quality of a summary?
Yes. AI summaries can be rated on a five-point scale from “Very good” to “Very poor.”
A comment can also be added to explain the rating. AI Help answers can likewise be rated as helpful or not helpful.
Is every AI feature available in BCS.light?
Availability may vary by feature. The relevant factors are the BCS version, edition, license, activated components, and system configuration in use.
The current release information and individual licensing should be reviewed to determine which features can be used in BCS or BCS.light.
Your Contact
Projektron Customer ServiceProjektron Customer Service is your point of contact for all matters relating to BCS and is available to answer your questions or provide further information about the AI features.
