AI-Powered Project Management: The Pros and Cons

In the tech world, it's common for the DevOps team to be responsible for the project management aspect of software development. DevOps workers have a strong reliance on analytics, so it's not surprising that there is a push toward using artificial intelligence — which does analyses of its own — in project management. This trend has both pros and cons — here are two of each.

Pro: AI could be ideal for handling large amounts of data

One task in which AI excels when compared to humans is processing vast amounts of data. There are countless examples of how teams applied AI to their work and achieved the desired results in days when it would otherwise take months.

“Using AI for data analytics and machine learning could help project managers hone their project plans to be more efficient based on past performance,” says Jayne Groll, CEO at DevOps Institute. “AI can also identify risks, indicators and trends that could threaten the success of the project.”

However, Groll cautions, “AI should in all circumstances serve as a ‘trusted advisor’ but not as the actual project manager.”

So, if a project is long-term or has an above-average amount of data associated with it for another reason, AI software could utilize the compiled data to make decisions and reach valuable conclusions after mining through the information. Additionally, an AI tool could find something in the data that causes project managers to make adjustments and see gains. It's best if project managers have questions in mind that they want an AI interface to help them answer through data. If that doesn't happen, people could spend too much time viewing data and not have concrete reasons for doing so.

Con: AI project management apps are in the early stages

Some tech leaders may have difficulty convincing their superiors to invest in an AI-based project management tool, since those options are less established compared to the apps and other project management methods that are more familiar.

For example, researchers recently made a framework for an Agile project management AI tool that seems promising and could have significant potential. But, even the team themselves refer to it as "a big, ambitious roadmap for future research and development of an AI tool suite for Agile project management." They do, however, believe their prototype could assist in almost every step of the Agile lifecycle.

In another instance, a firm in New Zealand called Psoda developed an app to synchronize physical and digital Kanban boards. The app, known as PsodaVision, uses AI for machine vision and works when a user takes a picture of physical Kanban cards, then views them in the app to confirm the correctness of the digital representation.

Psoda claims that people use its tools to handle more than $10 billion worth of projects around the world. If companies do not want to be early adopters, their tech professionals may have to wait for further evidence that AI project management tools are worthwhile and don't come with too many unforeseen pitfalls.

On a positive note, enterprise-level use of AI has gone through rapid growth over the past several years. A Gartner report revealed a 270 percent increase over the past four years. IT teams could bring up statistics like that to prove how AI is breaking into the mainstream.

Pro: AI could help keep project costs down

Project management professionals engage in a near-constant battle to ensure their teams don't go over budget. Research indicates that AI project management interfaces help keep costs at a manageable level in several ways.

Findings from Accenture showed that up to 54 percent of a project manager's time goes to administrative tasks, but analysts think AI will halve that figure. Moreover, AI can visualize data and illuminate bottlenecks in processes that would otherwise stay hidden. Another use of data is to input information about average completion time across teams and let the AI app figure out the likelihood of projects finishing on time.

Jayne Groll

Numerous issues with above-average costs occur when things happen that project managers didn't expect. Some of the problems will still be out of the control of the individuals responsible, but AI could be instrumental in helping project managers be more proactive and think of ways to minimize cost-related challenges before they become out of control.

Con: Data could lead to wrong conclusions if not properly trained

“If we start to rely too heavily on AI for decision making,” Groll says, “we may see a decrease in high level critical thinking skills that (at least for today) are unique to human beings.”

“Project management is as much art as it is skill,” she says, “so there should be a balance between the fast and smart analytics that AI can deliver and the skills and wisdom that talented project and product managers bring to the enterprise.”

Additionally, people often say that AI software is only as good as the data used to train it. Companies that are thinking about using AI for project management need to recognize that it may take a significant amount of time to teach the AI program, and doing so includes cleaning up the training data. Preparing the data used for the algorithms is often the most labor-intensive part of the process.

As such, the training segment of implementing an AI project management system cannot be rushed or overlooked. Otherwise, it's possible and even likely that the conclusions reached by the AI will be wrong or incomplete. Such a result could happen despite a company investing in the most advanced option for AI-powered project management available.

Project management still needs human involvement

Using AI for project management could cut down on some of the repetitive or administrative tasks that take project managers away from doing other things that drive progress in more meaningful ways. However, applying AI to project management is still a relatively new endeavor, and it could bring unintended complications.

In any case, humans cannot become too dependent on AI for project management because even a well-trained algorithm can still reach incorrect conclusions. When people monitor the data and evaluate it for possible problems before taking action, they increase the chances of positive results.