Our recent webinar with Projektmagazin was a great success, and we thank you for your active participation and the many questions asked.
The discussion ranged from technical aspects of implementing our AI software, to training and application of neural networks, to practical scenarios in machinery and plant engineering. In order to continue providing you access to this valuable information, we have carefully prepared the Q&As in our knowledge base.
1. What is the difference between real and artificial intelligence in project management?
Real AI has the ability to learn and adapt independently, similar to human intelligence, while artificial intelligence is limited to specific tasks and simulates human-like intelligence only within a predefined framework.
→ Learn more in our Knowledge Base
2. What added value does real artificial intelligence provide?
The AI in Can Do offers users several benefits. It learns from past project experiences and can effectively assess and monitor current projects. Can Do provides real-time recommendations for efficient project management. Overall, Can Do significantly facilitates project management for users.
→ Learn more in our Knowledge Base
3. Are the alternative resource suggestions for replacing a critical resource prioritized by the AI, i.e., is the most suitable suggestion presented first, or is it random?
The AI's alternative resource suggestions in Can Do are prioritized and follow an order based on suitability, with the best fitting suggestion presented first, without the use of a random generator. → Learn more in our Knowledge Base
4. How does model training work in Can Do? Are local models trained with my "customer data" or do I contribute to training the general Can Do models?
Model training in Can Do works in a way that the system is pre-trained, but customer-specific customization is done through the customer's own training data. For example, the AI learns from decisions like resource replacement, such as when "Ms. Müller" is replaced by "Mr. Schmidt." This way, the model becomes increasingly tailored to the specific needs and circumstances of each customer.
5. Who are the specific users of the system? Do users need to know the project staff/teams to know whether to follow the AI's suggestions?
The users of the system are typically project managers and teams. Understanding the project staff and teams is helpful for effectively utilizing the AI's suggestions.
→ Learn more in our Knowledge Base
6. How secure are my data?
Data security and privacy are ensured as the systems are customer-specific and do not share data with external AIs.
→ Learn more in our Knowledge Base
7. How long does it take to implement the AI system for a medium-sized company?
The implementation of Can Do through the RampUp process is gradual and agile, with users working with real data from the beginning. Initial results are usually visible within 3-4 weeks. The total duration of the RampUp varies but can be completed within 2-3 months for some companies, while others continue to use it for additional topics or areas.
8. What other tasks in project management could be complemented by AI?
Artificial intelligence can be used in various areas of project management, such as resource planning, risk assessment, time management, budget monitoring, and workflow optimization. AI can help make better decisions, increase efficiency, and analyze complex data sets.
9. Can AI be introduced gradually or does it require a "Big Bang"?
The introduction of AI into project management can be done gradually. This allows testing the technology on a smaller scale, collecting user feedback, and making adjustments before it is fully implemented. A phased approach can minimize risks and promote acceptance among employees.
10. How far can artificial intelligence simulate human association?
AI can mimic many aspects of human cognition to some extent, especially when it comes to analyzing and processing information. However, it has limits in terms of creativity, empathy, and intuition. AI is based on algorithms and data and therefore cannot fully replicate the complex and often unpredictable nature of human thoughts and feelings.
11. How laborious is data maintenance for, for example, 100 employees?
The effort for data maintenance in Can Do depends on the quality of existing data and the specific requirements of the company. Can Do minimizes this effort by importing and exporting data from existing systems. In general, the maintenance effort decreases with the use of Can Do. → Ramp-Up: Time required for manual data entry
12. What data foundation in terms of format and volume is needed to train Artificial Intelligence (AI)?
In Can Do, the data captured and maintained in the system is sufficient. This data repository is built over time and does not need to be explicitly captured for AI.
→ Learn more in our Knowledge Base
13. Why could it be a problem that AI cannot explain how it makes decisions?
The problem of AI not being able to explain how it makes decisions is known as the "Black-Box" problem. It occurs especially with complex models like deep neural networks, where decision-making processes are not easily understandable. Can Do addresses this by explaining the circumstances that lead to a situation such as resource overload.
→ Learn more in our Knowledge Base
14. What types of existing data formats can be imported when starting with the Can Do system?
Can Do allows importing various file formats, including Excel, MS-Project, CSV, ASCII, and XML. There are also special apps that allow defining files for importing specific data such as master data, project data, and working hours.
→ Learn more in our Knowledge Base
15. How many historical data are required if there are currently 20 projects but no historical data?
The absence of historical data is not a problem for training AI in Can Do, as imitation learning is used in the area of resource alternatives. This allows the AI to effectively learn even without historical data.