Artificial Intelligence (AI) will transform project management.  Gartner estimated that AI may replace 80% of project management activities by 2030.  A recent  article in the Harvard Business Review described “How AI Will Transform Project Management.”  

I say, “Bring it on!”  I look forward to the opportunity. 

AI is powerful, potentially disruptive, and will change what project managers do and how they work.  This is not a new story.

The 1950s “home of the future” promised to eliminate the drudgery of housework.  It did.  But the total time spent on household chores barely declined.  

Personal computers transformed the workplace in the 1980s.  Activities such as writing a report or creating a paper spreadsheet were incredibly arduous and error-prone.  PCs made those activities “easy.”  So we ended up working more hours rather than less.   

A project manager’s day is still consumed with drudgery.  AI and related technologies offer the opportunity to shape-shift the profession.  Mundane activities will be automated, allowing us to focus on more value-added work—such as engaging stakeholders.  Remember, “soft” skills distinguish great project managers, and that cannot be automated.

In this article, I consider how emerging trends may impact the profession.  The Jetson’s Rosey the Robotsparked our imagination. And Kubrick’s HAL-9000 sparked fear.

Sam.ai, the Scheduler

Project scheduling experienced its first transformational moment 40 years ago when Primavera and Microsoft Project were introduced.  Before then, project schedules were either hand-drawn or created on mainframe computers with limited access.  The ability to build a schedule, calculate the critical path, and manage baselines on a personal computer was a huge advancement.  Regardless, most schedules are poorly constructed, and half of the projects are delivered late.  

Tools are already available to analyze project schedules.  SmartPM is a construction schedule analytics tool that grades the schedule and identifies gaps and potential issues.  One could easily envision expanding those capabilities by applying generative AI to other industries and domains.  

In the future, a bot could take us through a series of interview questions and build a project schedule based on a vast database of learned performance.  Deliverables and activity lists could be generated. Dependencies could be automatically generated and alternative paths identified.  

Sam.ai could simplify creating the project schedule and documenting underlying assumptions, constraints, and risks.  However, the project manager, resource managers, and team members would still be required to review and validate the plan.  Deployed in an enterprise environment, Sam.ai would create greater consistency and (hopefully) repeatability across the projects.  

Effort-driven projects, like construction, are likely the first candidates for AI automation.  Knowledge-work projects tend to be more unique and require more work to model.  Generating the data to feed the AI engine may also be a challenge.  

Notes from Listen.bot

Taking meeting minutes, publishing them, and updating the issues, risks, and action item logs is pure drudgery.  I enjoy facilitating a meeting, but hate the follow-up work.  My rule of thumb is that an hour meeting generates at least 1.5 hours of documentation effort.  

Several companies have introduced bots that “listen” to a meeting and produce meeting minutes with action items and key points called out.  I have experimented with some of these applications.  They are good but still maturing.  Automating the note-taking reduces the drudgery, but…

• Effort is still required to review the notes, check context, and make revisions; and 
• Issue, risk, and action item logs need to be updated manually.

The biggest meeting management challenge cannot be automated—engagement and follow-through.  I am afraid that AI-generated meeting minutes will reduce engagement.  People will be overwhelmed by the documentation or will ignore it.  Or they may have their bots write summaries of Listen.bot.  Wouldn’t that be funny?  

Don.ai, the Document Generator

AI and Machine Learning tools can consume billions of documents and generate reasonably good encyclopedia-like entries on any topic.  I used an AI bot to create a standard course description.  Colleagues report using the tools for similar mundane tasks or idea-generation work.  

So imagine the future.  What could Don.ai do?

• The project charter is not a good candidate for AI.  The charter is unique to the project, describing the problem statement and proposed solution.  
• The scope and requirements documents could be AI-assisted.  The tools could generate a comprehensive list of standard requirements.  Stakeholders and product owners could then review and prioritize the items.  
• Testing could benefit the most AI and related technologies.  Imagine a tool that would take requirements, develop test conditions, and automatically execute the tests.  

AI can quickly generate volumes of documentation.  However, that does not solve the fundamental problem of producing good documentation.  AI tools hallucinate and produce inaccurate or unverifiable information one-third of the time.  

As an experiment, I asked ChatGPT to create a 500-word article on the impact of AI on project management.  While the article was “written” in seconds, it was vague with interesting but unelaborated ideas.  

Robb.ai, the Reporting Wizard

I long for the day when status reports and dashboards are automatically generated.  Collaboration tools with self-reporting capabilities could remove the project manager from the process.  But the weak link in the reporting chain still exists—the people.  Project team members must actively participate in the process and provide information, which they often do not.   

Effective communication requires understanding stakeholders and their information needs, which is a soft skill developed through experience.  Status is subjective.  Projects have Pooh-Bears (eternal optimists) and Eeyores (pessimists).  Good project managers reduce the signal-to-noise ratio to find the “truth.”

Executive dashboard proponents have promised self-service reporting and analysis.  The previously mentioned HBR article envisions a CEO checking on the company’s strategic initiatives through a smartphone app and making decisions.  Sounds great, but!  Do we want CEOs making tactical project decisions?  They should be focused on strategy.  Don’t we want to empower teams to make decisions?

Imaging the Possibilities

Projects generate a lot of data.  Data comes from heterogeneous sources.  Most is qualitative, and the quantitative data often requires analysis to put it into context.  It would be amazing if AI tools could parse and synthesize this information in the project’s unique context.  Imagine the possibilities and consider the implications.  

Stakeholder Engagement

Spook.ai, the Stakeholder Monitor, could generate stakeholder profiles based on their emails, chats, social media, and “observation.”  It could generate insights into the stakeholders’ interests, power, influence, and engagement.  The tool could proactively alert the project manager when something is amiss.  

Imagine this future.  AI tools could generate a stakeholder-specific analysis of concerns, engagement, and influence.  Personalized messaging could be developed, and the preferred method and frequency could be selected.  Changes in attitude and behavior could be monitored, and corrective actions automatically generated.  

Wow, that’s creepy.  Remember Tom Cruise walking through the shopping mall in Minority Report?  

Assumptions, Constraints, and Risks

Imagine the technology used to steer an autonomous vehicle was employed to manage our projects. Assumptions, constraints, and risks could be generated, documented, and resolved without human intervention.  

Assumptions could be generated and validated progressively as the project progresses.  

The risk register could be automatically generated.  Risk could be scored, prioritized, and response strategies documented.  The environment would be continually updated with multiple “sensors” reviewing the threat surface.  Updates would be made on “observed” project changes.  Automated warnings would be issues when action needs to be taken.

Imagine how cool that would be!  Everything could be updated automatically.  But will we (the people) be engaged?

© 2023, Alan Zucker; Project Management Essentials, LLC

To learn more about our training and consulting services or subscribe to our newsletter, visit our website: http://www.pmessentials.us/.

Image courtesy of:  https://getyarn.io