May 17, 2023 – Reading time: 5 minutes
Delays in development and engineering projects are not only frustrating but extremely expensive. However, identifying factors that can create project setbacks, as well as checking the status of every teammate’s task, are time-consuming yet necessary steps to ensure timely, high-quality project outcomes. This is why efficient controlling of project management is crucial. What if we could automate the painstaking controlling work of assessing risks and monitoring task progress?
Project controlling is a constant and important companion in project management. However, in stressful phases and extensive projects, the associated effort can quickly get out of hand and divert attention from value-creating tasks. Thanks to advancing digitalization and modern technologies such as machine learning, the effort can be reduced and the focus on the essentials increases. In this article, we show two methods based on artificial intelligence (AI) that can realize this potential.
When and how AI can provide useful support in project controlling
Classically, project controlling has a steering and control function, providing support throughout the entire project management process. Success in controlling is based on the goals and plans formulated during project initiation. This is followed by the recording of the actual status for comparison with the planned variables and reporting. In this way, projects are designed more economically, and timely completion is ensured.
Figure 1: INVENSITY’s “Monitoring & Controlling” process. The abbreviations M.1 to M.5 stand for the various subprocesses.
The use of AI is particularly suitable for tasks with high effort or high analytical complexity. In our cases, AI is used to identify deviations and perform risk analysis. The challenges we observed in these processes inspired us to develop solutions using AI to realize potential.
Use Case 1
3 Steps to Conducting Risk Analysis with the Help of Machine Learning
Choosing the right solution(s) is important for the successful use of AI. It must first be considered whether to rely on generic solutions, e.g., language models such as ChatGPT, or specialized, customized solutions. In this case, one cannot rely on standard solutions and must develop one’s own data infrastructure with appropriate data quality.
Step 1: Establish a suitable database of sufficient quality: All relevant data must be available in a usable and uniform form. In order to build up a complete data base, it makes sense that the release of the next project phases is only given after the risk analysis has been carried out. This approach ensures that the analysis is regularly updated and maintained.
Step 2: Train a risk model based on past risk data: INVENSITY relies on the use of neural networks based on supervised learning. Supervised learning is a part of machine learning and works by training the model with data that includes information about the inputs and associated outputs. In the case of risk analysis, it is information about the factors that cause an effect (input) and the effect itself (output).
Step 3: Introduce and validate the risk model: Once the implementation phase begins, the model is run in parallel with the manual process. This provides the project manager with helpful information from the model while continuously collecting data to improve the model. At the same time, trust is built by constantly comparing the AI’s risk analysis with that of project managers, verifying the plausibility of the AI’s decisions.
Use Case 2
Status Determination with the Help of Artificial Intelligence
The second AI solution revolves around more efficient status determination and associated planning adjustment. With the help of AI, the effort can be reduced, and the actual status can be continuously maintained. Chatbots like ChatGPT are suitable for this. This is linked to the project planning, keeping track of employees’ assigned projects and the latest status updates of their work packages. The AI then regularly queries the status of the work package throughout the project, thereby significantly increasing efficiency.
Thanks to AI, there is great potential for cost and time savings not only in project or resource planning, but also in project controlling.
Source: Full Article in Projektmagazin
Find the whole article in the projektmagazin here.