Artificial Intelligence in Today's Project Management

We used to talk a lot about the future of project management and a place of artificial intelligence in the field. However, the fact is that the future is already here, and AI is a significant aspect of today's project management process. Technologies are advancing by leaps, and their impact on this area can not go unnoticed.

Artificial intelligence in project management

What is AI in project management

Skeptics will say that modern AI is not intelligent enough. And they'll be right. Today, AI is often understood as any automation, that is not completely true. Automation presupposes clearly predetermined actions and rules that the system follows. It allows you to program the computer to perform tasks, the nature and sequence of which are well known to humans, to free us from manual work. Artificial intelligence, in turn, is useful when the software must be able to learn and make decisions about the appropriateness for certain tasks independently.

Both approaches have the right to exist. AI shouldn't be applied just for the sake of AI itself. When the task can not be explicitly programmed, since the entire range of initial data and conditions are unknown or too large, the capacity for thinking is necessary. In all other cases, simple automation is enough. Maybe that's why nowadays, project management makes extensive use of automation, while artificial intelligence is not so widespread. Until recently, automation was sufficient for most tasks. However, human needs are growing along with the development of technologies, volumes of processed data and complexity of workflows.

Somebody believes that automation is just the first phase in the evolution of AI in project management along with the following stages:

  1. Integration & automation
  2. Chatbot assistants
  3. Machine learning-based project management
  4. Autonomous project management

We think that at the moment we're at the early third stage: we still cannot replace a project manager with an autonomous system, but the first attempts to apply machine learning to the area have already been made. What about global integration of different project management tools and services, automation and chatbot assistants, they have now become an integral feature of the most project management programs.

Thus, project management AI is a system responsible for performing management and administration tasks without human influence. It aspires not only to do typical tasks instead of employees, but to foresee the necessity of the task itself, understand its essence, offer the best way to solve the problem, including in unconventional situations. Absence of undoubtedly stronger human thinking is compensated by good computational ability, the presence of project data for training and analysis in the process of work.

Why do we need project management AI

If we take a look at the latest statistical data on project management, we probably won't be too surprised: they simply demonstrate the facts that we come across many times in our daily practice:

  • Managing project costs (49.5%) was the biggest problem faced by manufacturing project managers in 2017. Hitting deadlines (45.8%) and sharing information across teams (43.9%) weren’t far behind.

  • 59% of U.S. workers say communication is their team’s biggest obstacle to success, followed by accountability (29%).

  • For every $1 billion invested in the United States, $122 million was wasted due to lacking project performance.

  • 75% of business and IT executives anticipate their software projects will fail.

  • Fewer than a third of all projects were successfully completed on time and on budget over the past year.

According to Business2Community and Capterra

The list can be longed, but it won't change the tendency. Project management is associated with a variety of risks and losses, and it seems that we just accepted it instead of looking for a solution. By the way, 44% of project managers even don't use PM software, though PWC found that the use of PM tool increases performance. When it comes to project duration, there is a tendency towards underestimation. Standish group's chaos report claimed that 52.7% of projects cost 189% of their original estimates. Compliance with terms and budget is the most common issue and AI, that proved itself well in forecasting and optimization, is one of the solution.

Use cases of AI in project management

The list of tasks that can be assigned to AI includes forecasting, estimating, making recommendations and optimization. By learning from our project management historical data and data collected by other companies, artificial intelligence can make assumptions about the future development of the project. For example, Riter AI will be able to:

  • reduce tasks estimation errors taking into account the individual characteristics of developers, such as productivity and accuracy coefficients, average number of worked hours per sprints, current load level, previous experience and gathered statistics;

  • automate sprints planning, calculating the exact terms of tasks and distributing them in an optimal way among developers, predict sprint success or failure and define real deadlines which can be met;

Effectivity of such AI will be increasing in the process of use as it will learn from each new project and real statistics.

In addition, even simpler examples of artificial intelligence such as chatbots can use achievements of machine learning to perform their duties. Semantic parsing, automated planning, speech recognition, natural language processing... The fact is that the potential of chatbots is huge, but we use their capabilities in a minimal way. If we direct all the power of AI to chatbots, the level of business profitability will grow before our eyes. For example, chatbots let developers plan and change tasks just from their messengers with colleagues without having to open the PM tool itself.

AI could also automate time tracking process by monitoring user's activity during the workflow and recognizing screen content. Thus, developers won't need to change task states manually, commit time spent on work, prepare endless reports on the project and do other bureaucratic actions they hate. Productivity of the whole team can be increased in a moment without breaking any methodology requirements and confrontation with managers.


However, there are certain obstacles to be overcome before we can enjoy the benefits of AI in project management.

First of all, datasets. To start behaving in any reasonable way, AI needs to be trained on really large datasets. How long does it take to solve a task for a team of 10 developers if each of them is also engaged in N more projects? If one employee rates a task in 5 hours, the second one in 7, etc., when can it be finished if we know the accuracy of their previous estimates and level of competency? The answers to these questions can be found only on the condition that our AI has been trained on a variety of data. Having large statistical datasets and current project and team detailed information can help, but most likely this won't be enough. Some project management datasets and other helpful information are published at ISBSG, Openscience, Hindawi. We could also analyze the statistics of git repositories for this end. But for serious goals the problem is still relevant.

The second problem that we encountered, while developing the idea of Riter AI, is an extensive control over the team's activity. This mainly concerns our previous statement: "Riter AI could automate time tracking process by monitoring user's activity during the workflow and recognizing screen content". However, this functionality should make developers' lives easier, and not put pressure on them in the process of work. Data collected by the system will be used exclusively by AI to better understand and predict users' actions and perform some routine duties for them. We are convinced that the trust between developers and managers plays a key role in the success of the teamwork. Can managers use screen recognition feature to track developers? Such AI-based software already exist, but we believe that this's unnecessary in a good team.

The next issue is connected to changeable conditions of each project and a team. Modern AI is a highly specialized thing aimed at solving a specific problem. Universal systems which can easily navigate in the modern world and adapt to any conditions still remain in our dreams and science fiction. Even if you've prepared a giant datasets and spent hours and days teaching your software to cope with common project management problems, don't think that you have nothing to worry about. Changes in the team structure, market conditions, the emergence of new technologies, production processes, methodologies - and your model won't work anymore. Moreover, it may happen, that a neural network successfully implemented in one team won't work in another. The complexity of the learning process and code reuse, the high demands on resources push many companies away from the use of AI.

In addition, there're a lot of aspects in project management which can't be replaced even with intelligent system. Good communication skills and emotional intelligence are essential qualities for a project manager, but not every person can handle with them, not to mention computer programs. Artificial intelligence may be a good performer, but this is not a good team player. Thus, in many tasks which need collaboration between team members, AI is not the best solution. At least, it requires constant monitoring by managers.

The last obstacle is basically in our mind, since the real problem really does not exist. Some humans believe that AI is taking their job. For example, a recent Atlassian survey found that 87% of respondents said artificial intelligence would change their job in the next three years. Almost the same number said that some part of their job could be done by AI. However, we don't need to worry about this. Gartner predicts that by 2020, artificial intelligence will actually create more jobs than it eliminates. The best thing we can do is to study artificial intelligence before it enters our lives.

"Now is the time to really impact your long-term AI direction. For the greatest value, focus on augmenting people with AI. Enrich people's jobs, reimagine old tasks and create new industries. Transform your culture to make it rapidly adaptable to AI-related opportunities or threats." - Svetlana Sicular, research vice president at Gartner.

AI implementation in today's PM tools

If you don't want to delve into the subject on your own, there're a lot of ready solutions for any taste. You can take advantage of smart project management tools right now, here're just some examples of their capabilities.

  • syncs with meetings on your Google Calendar and automatically transcribes and summarizes what is said during them. The service supports Webex, Zoom, Skype for Business and other web-conferencing platforms. There is also for Slack – a bot that helps you track work items and start automating tasks like meetings, emails, and integrate with your project management tools.

  • coordinates team members, auto-identifies risks, sends reminders, tracks performance, and integrates different project management tools.

  • Dialogflow – a powerful virtual assistant with natural language processing that integrates users on many popular platforms and devices.

  • Knightspear – a gamified project management system with a built-in AI work coach Isabella. She reminds you about overdue tasks, upcoming events, unfinished conversations, and analyzes team productivity to suggest improvements.

  • Teodesk makes use of natural language processing to sort your e-mails. It automatically assigns the mail that arrives to the related project and people involved, so you don't have to spend your time doing that.

  • Products like Clarizen and Forecast use artificial intelligence to simplify and automate resource and project management. They learn from project history and create a regression model to provide future estimates of budget and task duration.

  • ClickUp is currently working on machine learning functionality to predict which actions a user is likely to take. This will enable task predictions and, over time, subjective task estimation.

  • is another AI-driven solution which sorts communication across various channels and consolidates relevant information from different apps.

  • automates recurring tasks, strives to identify risks and suggest measures to minimize them, helps prioritize to-do lists to reduce wait time.

In summary, we believe that AI still has much to do for project management. Repetitive and manual tasks, approximate calculations, long-term disputes on time estimates, broken deadlines, and exceeded budget must be left behind. On the other hand, we have a lot of issues to solve so that the technology doesn't cause for concern to any unexperienced user.

Riter development team