ChatGPT is having its moment in the sun. As organisations get excited about or grapple with the potential of generative AI, there are questions about how effective it can be in the project environment.
Artificial intelligence and machine learning is a hot topic for organisations, tech leaders, and project practitioners right now, as the launch of the large language model ChatGPT drives conversation around generative AI, its capabilities and fuelling the imaginations of many about what its possibilities might be in the future.
While the prompt engineers might be enjoying their moment of influence and others suggest that humans will become redundant in key professions like law, accounting, project delivery, manufacturing and more, the promise of the latest wave of AI is fascinating but largely still unknown.
There are lots of questions, thoughts, and concerns about what “AI” can actually do. The latest graphics generators can create some amazing imagery and other creative assets, but there’s some good reasons to stand back a little as the wake goes by. The more nefarious uses of AI deep fakes have resulted in images of Donald Trump’s arrest following his indictment when in fact he turned himself in peacefully and then there’s somewhat amazing Tom Cruise deep fakes have resulted in a new company dedicated to video-based AI generation and de-aging Hollywood royalty in new cinema.
If seeing is believing, how long before we could be fooled completely into believing anything is possible, including comedian and satirical commentator John Oliver marrying a cabbage? See that AI output for some truly hilarious and disturbing results.
Is this just the beginning of what’s AI could deliver? It’s possible we may end up in a place we can’t quite imagine yet.
And so it is in the project world. There are lots of conversations about what benefits or impacts AI and machine learning will have on the world of the project and where it might deliver the best benefits, so we got a little curious and decided to have a little late Friday chat with ChatGPT what it has to say about how it will influence the world of project management.
We put our seasoned Practice Lead, Jon Pascoe, in the prompt engineer’s chair to ask two questions about what ChatGPT would tell us about AI’s impact on project delivery.
Question: Tell me, how will AI impact the project management function and the project management office
ChatGPT: AI has the potential to transform the project management function and the role of the PMO, here are some ways AI may impact project management.
- Improved data analysis: With AI and ML, PMOs can more accurately analyze large volumes of data from various sources, allowing them to identify trends, risks, and opportunities faster and more accurately.
- Better decision-making: AI and ML can help PMOs make more informed decisions by providing real-time insights and predictions. This can help PMOs to identify problems earlier and make adjustments to projects quickly.
- Improved project forecasting: AI and ML can help PMOs more accurately predict project timelines, budgets, and resource needs by analyzing data from previous projects, as well as external data such as market trends and economic indicators.
- Enhanced project scheduling: AI and ML can help PMOs optimize project schedules by taking into account multiple constraints and variables, such as resource availability, task dependencies, and project risks.
- Increased automation: AI and ML can automate routine tasks, freeing up project managers to focus on more complex and strategic activities. This can help PMOs increase efficiency, reduce costs, and improve project outcomes.
- Improved risk management: AI and ML can help PMOs identify potential risks and predict the impact of those risks on project outcomes. This can help PMOs take proactive steps to mitigate risks and minimize their impact.
JP’s Perspective:
Projects often face ambiguous decision-making, laden with assumption and risk that goes with carving a new path. Most project professionals would welcome an ability to improve data analytics to identify trends, risks, and predictions and shore up critical decisions such as execution plans, business cases, cost estimates etc.
As the AI response suggests, it requires large volumes of data to make determinations better and the challenge that this presents an organisation is often projects are completed once. It is why organisations come to firms for specialist expertise. For example, most organisations will change an ERP once every 10 to even 20 years, a specialist firm might do hundreds of implementations in that time and have more insights.
So, in theory, an AI model would be less effective using data from 1 ten-year-old project to predict or improve decision-making verse a model that has seen hundreds of ERP projects.
However, an AI model could take all data from all an organisation’s projects to identify underlying trends that might be occurring in the project ecosystem. AI could greatly assist in better identifying resource constraints and patterns in estimations vs actuals in the past to better inform future project forecasts.
At a task level, it would be advantageous to know if, for example, the IT integration team is regularly 25% under quoting cost and time for supplying integration builds, or that other types of tasks are regularly impacted by resource contention, issues, risks or other drivers of cost and time impacts. As project forecast, resource and cost estimates improve, in theory so should project scheduling. We feel regardless of the nature of the project, waterfall, hybrid or agile, the principles could still apply given the AI’s ability to take data from multiple sources, if it’s captured digitally. For Agile projects, the original story points and velocity estimates can be matched against actual effort spent and actual velocity achieve to help improve agile project estimates too.
The big one that seems closer is automation, removing more and more manual and repetitive tasks. Robotic Process Automation (RPA) has been around for some time, one would assume that this type of tech could be turbo charged with AI.
This isn’t completely new, AI has already been used by one RPA technology to ‘learn’ processes by observing how people interact with systems and data and turns that learning into actionable process it can undertake. Most project professionals know of the time it takes updating cost spreadsheets, merging in time sheet data, running pivot tables etc etc. A world in which smart AI automation completes these important but tedious tasks accurately so more time can be allocated to critical thinking and stakeholder management.
Question: Do you think AI could replace humans in project management?
ChatGPT: While AI can assist project managers and improve their capabilities, it is unlikely to replace humans in project management completely. AI and machine learning are good at processing large amounts of data and making predictions based on that data, but they lack the creativity, empathy, and critical thinking skills that humans possess. Project management involves more than just data analysis and decision-making; it also requires interpersonal skills, leadership, and the ability to motivate and communicate effectively with team members.
Furthermore, AI is not a replacement for human judgment and decision-making. Project managers need to consider a wide range of factors when making decisions, and not all of these factors can be captured in data. Human project managers also need to be able to adjust to unexpected developments, make decisions in ambiguous situations, and use their experience and intuition to guide projects to success.
JP’s Perspective:
We agree, project management remains a deeply human-to-human collaborative effort to drive change and transformation. AI will bring with it a number of opportunities to improve projects with better tools but will still need a human to drive them.
Having tinkered with some AI over the past months, we have noticed gaps or poor-quality responses in the world of project management, for example, one AI bot told me the difference between agile and waterfall was that agile doesn’t perform risk management, which we felt was not a fair or accurate representation of the methodologies.
AI and the PMO: It’s not ready… yet
What we took away from this experiment is people still need to invest in education, credentials and learning and develop strong critical thinking skills to work with AI effectively. They also need to develop their ability to understand that a tool like ChatGPT can only call on information in its reservoir of information for now. Right now, perhaps the most likely application for AI is enabling process-driven or busy work to become more efficient.
It can provide some neat shortcuts and perhaps accelerate your work. Still, if you believe everything on the internet and without knowledge, theory, and experience, you may assume something to be true when it is not. Students be warned using AI to write a paper for them on a subject matter they don’t know. Australian universities are already launching ChatGPT detection software to detect cheats.
ChatGPT didn’t respond with any mind-blowing insights about project management to be fair, we didn’t think it would. Perhaps in a few years time when its reservoir is big enough, it’s possible the answers might be different if there’s a way to harness the data that we have in a more practical or aggregated way. Right now, we are at the beginning of exploring use cases, applying them, and learning from the outputs.
We do, however expect that this will change and that AI will accelerate in years to come.
We believe that quality thought leadership is worth sharing and encourage you to share with your colleagues. If you’re interested in republishing our content, here’s what’s okay and what’s not okay.