Part Three – Roadblocks to AI and what can we do now to Help.
So we have the pieces to make an awesome AI tool that can make life as a project manager more efficient, so why are we not there yet? While the technology is almost there, here are some roadblocks that still need to be tackled.
- Data Scarcity
- AI needs a large amount of data to learn and predict accurately.
- Data is the Raw Material needed for Artificial Intelligence to work – But it needs to be clean, and it needs to be formatted in a way that algorithms can read and learn from (labeled Data)
- It needs a lot of unique datasets (different patterns for it to learn from)
- the data received needs to be unbiased.
- Data Privacy
- While data privacy is good (and necessary) for humans – it is not so good for machine learning. This specifically applies to Project Management. Companies do not want to share details of their projects with other organizations (for obvious reasons); however, this makes it difficult for machines to get unique patterns due to the data within your organization possibly being biased.
- Trust Deficit
- There is a huge trust deficit because of the unknown nature of how machine learning models will predict the output. There is currently too little usage and not enough data to gain human trust.
- Most people don’t understand AI and how it is integrated into our everyday life. We all use some kind of map on our phones to get to work every day and check traffic. But most of us have no idea how it gathers the data and is able to predict your arrival time.
- Different Data Source
- Organizations use different software to capture project data – all the way from PowerPoints, Excel sheets, Kanban Boards, and project schedules.
- All of these tools use different standards to capture data. In fact, even if organizations use the same tools, each organization will have a unique set of standards for capturing data, making it difficult for machines to learn effectively.
As you can see, data is king when it comes to AI. The key to getting AI working correctly is to get huge sets of data for it to learn from, but, not just any data, accurate and labeled data that is captured by following a standard method.
Challenges with Capturing Data Accurately
- There is a need to have standard processes not just for the project delivery lifecycle, but also for how data is captured and what tools need to be used to capture this data.
- Most Project Managers do not have enough time and feel the only reason they are capturing project data in a tool is for management to run reports (but do not see the value in it). For most Project Managers it is an administrative duty of their job.
- Training Project Managers can only go so far. They either forget the tricks they learned during the tool workshops or the trainer only taught them from the “press this button” perspective, and no real-life scenarios were discussed.
- Project Managers change jobs and take their knowledge with them to other organizations, making it difficult to have consistency.
A Potential Solution
- Separate the duty of project data entry from actual project management. Have dedicated staff, “Project Coordinators”, that are “experts” in the tool you use, but do not have the necessary skills to manage a project.
- Have standard processes on how data should be captured. What labels and metadata need to be captured? Measure performance based on how accurately data is captured and not just on whether or not the project was successful. Provide continuous training on the tool since most tools provide enhancements on a regular basis.
- You could also outsource this activity to outside vendors such as EPMA.
Immediate results you will see even without AI
- Since these “Project Coordinators” will be working on multiple projects and with multiple Project Managers, you will start seeing some consistent data across your organization, even if you do not have very solid processes laid out for them.
- They can be the eyes and ears of the Project Manager by collecting data from Team Members and inputting it into the tool. They can also be looking at the effects of delays on the project and provide a what-if analysis to the Project Manager.
- With consistent data input, you will have more trustworthy reports and be able to make more accurate decisions on your projects.
- With accurate and timely schedule data, you will effectively be able to assign resources to the correct tasks without over-allocation and also be able to adjust quickly if things don’t go as expected.
- Be able to not only prioritize tasks within projects but also prioritize projects based on facts rather than being a Project Manager that “shouts the loudest”.
So I want to leave you all with this
In the ideal world, we want AI to give us, essentially, a brain superpower. As Project Managers, one of our functions is to deliver a project on budget, on schedule, and within scope. We work to that based on our knowledge and experience. We use our past experience to help us analyze the risks and make better decisions. AI is going to be able to not just analyze a project schedule, but actually analyze the entire portfolio, across all resources, critical paths, deadlines, and unlike us mere mortals, the AI is going to be able to continuously run millions of scenarios to determine the most efficient prioritization to deliver the project because that’s its mission.
AI is not superhuman. AI-driven applications just have a higher speed of execution and higher operational ability and accuracy while tackling tedious and monotonous jobs compared to humans. On the other hand, we often overlook that Human Intelligence is related to adaptive learning and experience. Human intelligence does not always depend on pre-fed data like the ones required for AI. Human memory in itself is the computing power, and although the human body may seem insignificant compared to the machine’s hardware and software infrastructure, the depth and layers present in our brains are far more complex and sophisticated. Machines still cannot beat it! At least not in the near future.
AI is high in IQ, but it is low in EQ and we should never underestimate the spirit of human ingenuity and persistence in the face of overwhelming odds. Therefore, AI will support us Project Managers by removing the tedious and monotonous functions in project management and schedule analysis, while giving us the time to focus on the more strategic side of Project Management.
Thanks for joining us on our three-part journey of how AI can revolutionize project management as we know it! Missed Part One or Two? Click the links below to see what you missed!
AI and the Transformation of Project Management – Part One – Project Management Best Practices & Microsoft Project Tips (epmainc.com)
AI and the Transformation of Project Management – Part Two – Project Management Best Practices & Microsoft Project Tips (epmainc.com)
EPMA is a global solutions company focused on delivering projects better. Our unique proposition that embodies the full ecosystem and portfolio of any organization has enabled us to deliver projects better and make a significant difference. Our unique and proactive approach of having over 20 developers and solutions architects in the business acts as a true enabler for impact. We are more than ever before positioned to add value, advise, and impact your organization.
Interested in hearing more? Please contact EPMA at 832.772.3762 or email us at [email protected]