top of page
Writer's pictureErwan Hernot

AI In Project Management: Great On Demos...


AI will help project managers but may be not very soon
AI will succed if... There are numbers of "if"!

It's all in on AI! Take for instance project management. There are numerous applications for AI in the project life cycle. Demos are impressive … on paper. Let's dive in a few examples.


1. Automated Task Assignment. Example: In a software development project, AI automatically assigns coding tasks to developers based on their expertise in specific programming languages, current workload, and availability. If a developer is overbooked, the AI reallocates tasks to another team member, ensuring that work is evenly distributed and deadlines are met. This minimizes manual intervention and ensures optimal team productivity.


2. Resource Optimization. Example: In a construction project, AI constantly tracks the availability of labor, equipment, and materials. When a particular resource becomes unavailable (e.g., a crane is being used on another site), AI suggests alternative resources or reschedules tasks to minimize downtime. This ensures that no resources sit idle and that tasks are completed as efficiently as possible.


3. Data-Driven Scheduling. Example: Using AI-powered software like Motion, a marketing project manager sets up a content creation timeline for a product launch. AI evaluates historical campaign timelines and flags potential delays based on current dependencies and team bandwidth. The project manager receives real-time alerts about potential bottlenecks, such as delays in content approval, allowing them to adjust the schedule to stay on track.


4. Scenario Planning. Example: In an R&D project, the project manager uses AI to run different scenarios—what happens if a key researcher is unavailable for a month or if a particular technology doesn't pass testing. AI simulates various outcomes, helping the team forecast potential delays or budget overruns. Based on the simulations, the team can choose the best course of action and mitigate risks by planning contingencies.


Through these applications, AI is enabling project managers to enhance efficiency, reduce risks, and ensure smoother project execution across various industries. Great. Except that it's not as easy as it seems. Let's breakdown these project management topics. The devil, be it digital or spiritual, is in the details. For AI-driven project management to succeed in streamlining planning and scheduling, several conditions of success must be met. These conditions ensure that the AI applications are accurate, reliable, and impactful in delivering real project benefits:


1. High-Quality Data. The effectiveness of AI depends on the quality of the data it processes. This means having access to accurate, up-to-date information on team members' skills, workload, resource availability, and project timelines. Example: For AI to accurately assign tasks, data on team members' availability and skillsets must be well-maintained. Similarly, real-time resource data ensures that AI can optimize usage effectively. You need someone to be responsible for setting up systems that capture and maintain accurate, up-to-date data. Chances are that it will be the project manager. He/she ensures team members regularly input information on their progress, availability, and task completion. If a team member is overloaded but hasn’t logged this information, the AI may make inefficient task assignments. And team members are very busy producing for multiple projects!


2. Comprehensive Historical Data. AI relies on historical data to predict potential delays, generate accurate schedules, and simulate what-if scenarios. The richer and more detailed the historical data, the better AI can forecast and adjust for future events. To generate a highly accurate project schedule, AI needs access to past project performance data, including actual vs. planned timelines, typical bottlenecks, and completion rates for similar tasks. Managers and project managers are then responsible for building and maintaining a historical project database. They also ensure that past projects are properly documented, which feeds into the AI’s ability to predict future events. Example: An IT project manager must ensure that after each project, the team completes post-mortem reports that detail timelines, challenges, and key learnings. When you know that most of the teams jump from one project to another, you can have serious doubts on setting up these post-mortem meetings not to mention the reports…


3. Defined Project Objectives and Constraints. AI can optimize resources and prioritize tasks only when clear objectives, constraints, and key performance indicators (KPIs) are defined. These parameters help the AI align its recommendations with the overall project goals. Example: If the project's priority is to stay within budget rather than reduce time, AI can allocate resources accordingly, ensuring that decisions around costs and deadlines are aligned with strategic goals. The project manager sets clear objectives and constraints (e.g., budget limits, time constraints) for the project, which guide the AI in making recommendations that align with strategic goals. Example: In a construction project, the manager must define whether the priority is completing the project on time or staying within budget. If time is the priority, the AI will optimize for faster task completion, even if that means higher costs. Without clear guidance from the manager, AI might optimize based on incorrect assumptions. Problem is: changes are frequent, stakeholders are rarely all on the same page at the same time.


4. Integration with Existing Tools and Workflows. AI tools must seamlessly integrate with the existing project management tools and workflows that teams are already using. This ensures that the transition to AI-driven management is smooth and that data flows between systems. Example: AI scheduling tools like Motion need to integrate with task management platforms (e.g., Jira, Asana) and communication tools (e.g., Slack, Teams) to gather data and provide real-time updates and alerts. Managers must oversee the integration of AI tools into the existing project management ecosystem. This includes selecting tools that work well together and ensuring data flows smoothly across platforms. Example: In an R&D project, the manager ensures that the AI scheduling tool integrates with the team's communication platforms (e.g., Microsoft Teams) and task management tools (e.g., Jira) so that data sharing is seamless and the team receives real-time alerts on task priorities. Data doesn't flow so smoothly and project managers can't help: legacy systems may resist to AI. Flows of data suppose a high quality of collaboration between different departments: silos are still strong.


5. Skilled Human Oversight. Condition: While AI can automate tasks and predict outcomes, human oversight is crucial for interpreting results, handling exceptions, and making final decisions. Project managers need to actively engage with AI-generated insights and validate their applicability to real-world contexts. Example: Even if AI predicts delays based on past performance, a project manager should assess whether the current team or external factors (e.g., market shifts) could affect those predictions and adjust accordingly. Will he be empowered to make the final decisions when AI recommendations need contextual judgment? On a very politically visible project, one can legitimately think that her/his boss would like to be part of the decision. Example: In a product development project, if AI recommends extending a deadline due to predicted delays, the project manager faces an alternative that is a classic in decision making : he/she needs to assess whether that’s feasible given other business priorities or whether it’s better to allocate more resources to meet the original deadline. But chances are that his/her boss asks for both: high priority AND no additional resources !


6. Of course for all these different AI applications to work, you need teams buy-in and AI literacy. For AI to be effective, the project team needs to understand how to work with AI tools and trust its outputs. This requires some level of AI literacy so that team members can confidently engage with the recommendations and insights provided by the AI system. Example: If AI suggests resource reallocation or task prioritization changes, team members should be trained to understand how these decisions are made and why they benefit the project. The manager is key in fostering team buy-in and ensuring that team members understand how AI tools work and how to use them effectively. But will he/she be able to train the team on how to interpret AI-driven insights?


7. Regular System Updates and Maintenance. AI tools need to be regularly updated with the latest algorithms and data to maintain accuracy and relevance. Outdated systems can lead to inaccurate predictions and suboptimal task assignments. Example: As new project management practices and data sources emerge, the AI tools must evolve to incorporate them and remain effective in the fast-changing project environments. Given the division of work and the fact that organisation is not often built as a customer centric system, project managers won't be responsible for ensuring that AI tools are updated regularly with the latest data, algorithms, and system improvements. Example: In a finance project, the manager needs to ensure that the AI system receives regular updates, especially if new financial regulations are introduced. This prevents the system from making outdated recommendations that could jeopardize the project’s compliance with regulations. But it implies that he/she works closely with IT or software providers to keep the system optimized.


I could go on with more examples. What appears is that managers in contact with AI will need to be empowered to make decisions otherwise companies could lost its added value. In real life, running smoothly AI applications also implies that collaboration between departments, divisions will always work. AI success implies a change in culture (beliefs, values, behavioral norms), in managers profiles and in organizational design of companies. And it is possible that it is asking for too much at the same time.


ClavaConsulting talks about AI in the following training modules :

  1. Basics for project managers.

  2. Managing a project.

  3. From one to many: portfolio manager.


0 comments

Recent Posts

See All

コメント


bottom of page