As Generative AI (GenAI) becomes part of our everyday world, it is important for a project manager to understan the basics. This includes how GenAI can help us and how it can be a risk. Where to begin? Maybe we should ask AI…
Go to https://infinity.pmi.org/chat and chat “How can I start using generative artificial intelligence in project management?”
In this case, the Generative Artificial Intelligence (AI) offers a “Top 4” list:
- Automating repetitive tasks
- Predictive analytics
- Risk modeling
- Natural language processing
Automate Repetitive Tasks
In our first option, we determine the repetitive tasks for automation. Pick a task, such as “Update Project Status”. Updating status is an important task but let’s automate it to save time, ensure accuracy, and provide transparency to stakeholders. First, we need to identify the historical and current data; inputting this information into a generative artificial intelligence model. Information needed might be project schedule, milestones, and task due dates. This input is an upfront cost for your project. If there is not an existing large language model to track your project management status, you must first train the AI model.
Predictive Analytics
A second option for early entry into AI is predictive analytics. This area is essentially predicting outcomes based on changes in scope, resources, and budget. By using generative artificial intelligence for predictive analysis, project managers can take actions early to mitigate risks and seize opportunities. An example is critical path analysis, managing tasks within the schedule and the entire system of the project. This allows the PM to move tasks seamlessly across teams to deliver quickly to the market.
Risk Modeling
Risk modeling is tied to to predictive analytics. Again, we leverage existing data from past projects and the current project to determine outcomes and more importantly, mitigate future risks. The data inputs are quantifiable, rather than qualitative descriptions of risk. Only with quantifiable data can we determine the probability and level of impact on our project. An example here might be a series of questions to our generative artificial intelligence model:
- “The schedule variance for the project shows that we are 10% behind schedule; what is the probability that we will deliver on schedule?”
- “What risk mitigation efforts are available to the project to get back on schedule?”
- “What alerts and indices are available to identify and more increases to schedule delays?”
Natural Language Processing
Natural Language Processing (NLP) allows a project manager to use generative artificial intelligence to analyze existing voice, text, or images. After the NLP use case is targeted for the project, you may need to train your AI model. This involves collecting and characterizing available data. Perhaps you will use social media posts and customer sentiment from your industry or product area as a baseline.
After selecting and training your AI model, as with all options, you may need to go back to verify model performance. Especially since you leveraged external data inputs, you will need to ensure extreme cases in the training pool are not over-emphasized. One final note about NLP, language is continually evolving! Continuous model monitoring and refinement will be mandatory when using NLP. Will your model be able to characterize “No cap!” as a positive or negative review for your product?
Conclusions: Generative AI for Project Managers
In summary, there are a number of easy entry points for leveraging and introducing generative artificial intelligence into your project. Some tasks will require a scanning of existing large language models and generative AI models to determine the best fit for your project. Some tasks will require a bit of data curation and early work for the project manager. But the benefits of generative artificial intelligence clearly assist the PM to automate simple tasks and free project resources.
This article was written by Frank Tank. Frank serves as the Military & Veteran Adviser to Peak Business Management. He is a certified PMP and a 30-year US Army Veteran, serving as an enlisted soldier and as an officer in the intelligence field. Frank guides military members and veterans who are pursuing PMI certifications.