When people say "Let's use AI" they usually mean that they need work done faster, efficiently and in a cost effective manner. However this does not take away the need to define the problem clearly first, deciding on whether the use of AI is the ideal approach and which type of AI should be used. Saving Changes...
When someone says, “We should use AI,” the first step is to define clearly what is expected and what problem is actually being addressed. AI can be helpful in organizing, prioritizing, and segmenting information, but it should support decision-making rather than make decisions on its own. Any insight it provides should be checked carefully against facts and business reality Saving Changes...
What is your expectation when you use AI? What is your plan? Do you have a clear road map? if all answers are yes, then we can use AI tools which will help us to understand the overall project and make a big difference between old school and the AI school. Saving Changes...
When a project says, "We should use AI," it often comes with high expectations. People expect the solution to be highly intelligent and capable of solving complex tasks efficiently. While AI is evolving rapidly and becoming increasingly accurate and capable, the most important step is to align expectations with the actual problem that needs to be solved. In the near future, I believe AI may exceed human capabilities in areas such as processing speed, data analysis, and calculation. However, the real challenge is not whether we should use AI, but how we can ensure that AI remains aligned with human values and helps people grow, improve, and contribute positively to society. The goal should be to use AI as a tool that empowers humanity, not simply to create something "super" for its own sake. Saving Changes...
ABDALLA HANAFIProject Management| ADWAA ALABRAR CO.ltdARAR, 8, Saudi Arabia
I would first step back and clarify the actual operational problem we are trying to solve. In facilities management and maintenance projects, the need may relate to repeated equipment failures, slow service-request routing, weak prioritization, resource constraints, or limited visibility of performance data. I would then identify the specific outcome expected from AI, such as predicting failures, classifying maintenance requests, recommending priorities, analyzing inspection data, or improving communication with users. I would also review the available data, the required level of accuracy, who will act on the output, and where human review is still necessary. For me, the key is not to introduce AI simply because it is available, but to confirm that it addresses a real problem, integrates with the existing workflow, reduces risk or effort, and delivers measurable value to both the organization and the service beneficiaries. Saving Changes...
I recently used a structured AI-assisted evaluation prompt for judging nominations — effectively an eAssessment task: scoring submissions against a defined framework (Prepared in advice criteria) rather than free-form grading.
The distinction that mattered here was "AI as a consistency layer" vs. "AI as the judge."
The tool applied the rubric uniformly across nominees and flagged where evidence was thin — but the actual scoring judgment stayed with the human evaluator.
In education/assessment contexts specifically, that line (rubric-consistency-assistant vs. decision-maker) is the one that determines whether you have a fair, defensible eAssessment process or a liability. Saving Changes...
Elena HoesleyManagement| US ArmyFort Novosel, AL, United States
AI is a great tool to enhance productivity but it shouldn't substitute the expertise of team members for the sake of efficiency. AI has served me as an analysis tool for our products or as generative AI to expand on ideas and what course of action would be more helpful. I like to think of AI as the antidote to "writer's block" for project managers. Saving Changes...
A statement like “We should use AI” is a signal, not a requirement. As project professionals, our first responsibility is to clarify the underlying problem and the value for the project. The desired outcome has to be defined. From there, I focus on identifying the actual need — whether it’s automation, augmentation, insights, or user interaction — and validating whether the existing data is reusable. Only once feasibility, constraints, and governance considerations are understood does it make sense to me to define scope.
Project & PMO Manager | Research & Enterprise Mentor| GFB HoldingSouth America, Brazil
The clearest signal that distinguishes different types of AI work comes from the nature of the problem being solved: automation of well-defined processes, decision support requiring human judgment, content generation at scale, or conversational assistance (each demands completely different scopes, premises, timelines, and deliverables). When everything is lumped together as "an AI project," what goes wrong is scope inflation (stakeholders each expect something different), conflicting premises (marketing wants content generation, IT builds a chatbot, sales expects lead qualification, all under one initiative), and unrealistic schedules (confusing weeks of automation with months of predictive modeling). The moment that tips you off that "AI" means different things to different people is when, in the same alignment meeting, the sponsor talks about cost reduction, the end user talks about workflow agility, the technical team talks about APIs and models, and legal talks about risk and compliance — and no one realizes they're actually discussing four different projects. The antidote is treating AI as a tool that serves the method, not the other way around: scope the problem first through project management discipline, define premises and constraints, and only then ask whether AI is the best answer or if a simpler rule, a process redesign, or better training would deliver more value. Saving Changes...
PMI Team Member
Community Champion
