Using AI in Project Management

Introduction

AI is now freely available, and everyone is thinking about how to use it, but in most organisations, its use is still very patchy. This article will explore how to begin to implement AI in your project management planning and control, starting from a position of zero knowledge or experience.

How to Start Using AI in Your Project Planning? 

If we simplify the project management process into twelve steps, we can then look at how AI might be helpful (or not) for each of the steps. 

Although it is developing rapidly, at the time of writing, AI is very useful for steps 1, 2, 3, 7, 8 and 12 and less useful for steps 4, 5, 6, 9, 10 and 11. 

It is interesting to note that the areas where AI is less effective, for example, in the drawing of a Gantt chart or monitoring progress, are the areas where traditional project management software like Microsoft Project is strongest, so by using both, we can have almost everything well covered. 

The other important point to make at the outset is that AI is only a tool to help the project manager. It can't (yet!) replace the PM. There are lots of things it can't do and areas where it can't entirely be trusted, so the best way to use it is to combine the results from AI with your own knowledge and feelings and those of your team. You could either do the task yourself first (ideally involving your team) and then see if AI can add anything useful to it, or you could use the AI first and then have a look to see if you and your team agree with it, and whether you feel it has missed anything. Either way, you are getting the best of both worlds: the human(s) and the computer.   

Let's have a look at the most useful steps at how you might get started in your use of AI in your projects, in order of their position in the 12-step process:

Step 1: Defining the Project

Now that you have practised using your AI software, you can start to apply it to some more subtle areas. At the start of the project, there are two ways it can be helpful.

Assessing Stakeholders

First is assessing your stakeholders – you can ask them who they might be, whether they are likely to be for or against the project, and how much influence they are likely to have. Then, you can ask how you could describe the project to motivate them to support it. You can also ask whether they should be consulted before you take action or whether they should be informed after you take action, and even what kind of actions need to be consulted or informed. Apart from doing a great job on this, using AI will encourage you to think about these questions, which are often forgotten or omitted by project managers who are impatient to get started on their projects. 

Recording Kick-off Meetings

The second way that AI can help at the start is to record the kick-off meeting(s). You can use Otter or similar to record the meeting (it can do it as an extra attendee when you are using Zoom or just by placing a phone in the middle of the boardroom table), and then it will give you a full transcript and, more usefully, a summary of what was agreed and any action points. Recording what was agreed at the kick-off meeting is a vital step for any project, and AI makes this very easy and efficient. 

Step 2: Listing the Tasks

This is the easiest part of the process to use AI for, and it shows instant and impressive results, so I would start here. Just ask ChatGPT (or whatever system you are using) to list the tasks for your project. If necessary, you can specify the number of tasks you want listing or the maximum size of any given task (in terms of hours worked or elapsed time in weeks) so you can control the level of granularity that you get. Remember to always sanity-check the results with your team and your own experience. 

Step 3: Estimating

Following on logically from the listing of tasks, it is quick and easy to ask ChatGPT to take the list it has given you and estimate the cost (broken down into hours worked and items to be purchased) and the elapsed time for each task. You can ask for the results to be presented as a table, and you can ask for maximum and minimum estimates if you want. You can ask it to add a column for contingency as well. 

Whilst you can also ask it to give you an estimate for the time required for the total project, i.e. based on some sort of critical path, I haven't found this to be as reliable, so I personally would rather take the list of tasks and do my own post-it note network diagram (step 4) to find and calculate the critical path. 

Step 7: Resource Planning

I would implement this last, as it's the most difficult and potentially least useful, although it could still be well worth doing. The idea is to get AI to help with the question of who to put on each task in your project. There will have to be some work upfront because you will need to give information on each person: how skilled they are in each area, how much they enjoy each area, how much they want to improve their skills in each area, and how busy are they – and then, based on this information, (which you will only have to write once), for every project it will work out the best distribution of your people. 

The main benefits of doing this are that there will be no favouritism or bias and that it will consider the development of your team rather than just putting the best person on each job – and bias is inevitable, and skills development is often forgotten when humans allocate roles for projects. 

Step 8: Risk

The next part of the process in which to start using AI is the estimation of risk. Just ask ChatGPT or whichever system you are using to list all the potential risks and estimate how likely and how serious they are, and then give you suggested mitigating actions. As always, compare the output with your own feelings about the risks. As with the listing of tasks, the chances are there will be things the computer has thought of that you might have missed, and vice versa, so the combined list is the best plan. This is so quick, so easy, and so effective that you'll always use AI for it from now on! 

Step 12: Reviewing and Learning for the Next Time

The big weakness of the existing way of reviewing projects is that it's too difficult to compare your new project with all the reviews of all the previous projects to find out how they might relate and be helpful. There will perhaps be small overlaps between your new project and some of the previous ones, but how do you find those overlaps? The answer is that AI will search through every previous project and pull out anything that is useful – maybe they were also in Japan, or also in cold weather, or also involved lifting heavy equipment? It will find every overlap for you in seconds. 

Of course, you do need a written summary of the learning points, good and bad, of your previous projects, so you may have to start doing that now, but once you have that, AI will give you a huge benefit. At last, the dream of Knowledge Management becomes possible, and we can all learn from each other's experiences. 

Conclusion

Using AI in project management is easy to learn and can be introduced gradually. There is no need for a sudden, risky and disruptive shift to wholesale AI-based planning. The cost of AI is minimal, and the benefits can be significant in terms of quicker planning, more accurate planning, and the capability for more planning to be done by the same number of project managers. Every organisation ought to be starting to move into AI-based project planning right now, or they risk wasting time and money, getting worse results, and getting left behind by their competitors.