En cualquier problema existen variables exógenas que corresponden al comportamiento de entorno, y variables endógenas, que corresponden al flujo interno de procesos, roles y decisiones. la IA contribuye a visibilizar y tratar variables exógenas, pero las variables endógenas por lo general solo pueden gestionarse a partir de la experiencia interna. Por lo tanto, es necesario la participación de individuos que pertenezcan al proceso y puedan analizar el problema desde la expertice interna. Saving Changes...
David SandersonDirector Program Management| FujifilmThe Woodlands, Tx, United States
The answer to what AI really is differs across cultures and even age demographics. Like any program analyses of the requirements across all stakeholders is need to fully understand the AI implementation conundrum. Saving Changes...
Utilizing AI for problem solving feels the same as any technical solution - it requires a PM to have worked with the team to well understand the problem that the project will be solving. AI is not a silver bullet for projects and can only support project success when the problem has been defined so the AI solution can be applied appropriately. Saving Changes...
AyoOluwa EbireIT Business Analysis| IHS Towers LtdLagos State, Nigeria
When there is a conversation to solve a problem with AI. The initial discussion should be to clarify what the requester needs AI to do, what solution is required, what the current situation or problem is. This helps to know if the solution expected requires the application of AI and what sort of AI capabilities is required to meet the needs. The lack of knowledge about what AI means is usually the first step in the wrong direction in applying AI to solve problems. If the requester cannot clearly define the problem and expected solution (As Is and To Be), buiilding an AI-driven solution could be a misalignment to the expectation. Saving Changes...
I’ve run into this exact situation more than once. The phrase “We should use AI” often gets thrown around as a catch-all solution before anyone has clearly defined the actual problem.
In one recent project, the business stakeholders kept saying they wanted to “use AI to improve our service request process.” To some people in the room, that meant automating the intake and routing of requests. To others, it meant generating personalized responses to customers. And to a third group, it meant predicting which requests were likely to escalate so we could intervene early.
What tipped me off that we were misaligned was that everyone was nodding in agreement, but when we started talking about success metrics, the conversation fell apart. One group wanted to measure reduction in handling time, another wanted higher customer satisfaction scores, and another was focused on fewer escalations. Those are three very different problems that require very different AI approaches.
What tends to go wrong when everything gets lumped together as “AI” is that teams end up selecting tools or designing solutions that don’t actually address the core issue. We either over-engineer something simple that could have been solved with better process design, or we under-scope something that genuinely needs more advanced capabilities (and the data to support it).
Now, when someone says “We should use AI,” I try to slow the conversation down and ask three clarifying questions:
What specific outcome are we trying to improve?
Is this about automating a task, augmenting human judgment, or generating something new?
What would “good” look like in measurable terms?
These questions usually reveal very quickly whether we’re all talking about the same thing. Curious to hear how others have handled these kinds of conversations. Saving Changes...
The real question here, not why we should use AI? But can it will be reliable to use it without verify the output data especially when the data are massive and as human being need more resources and time to verify it Saving Changes...
When someone says, "we should use AI". it means signaling to integrate technology with human input to get speedy and efficient results in the project. However, the outcome and question posed to AI shall be clear to achieve the desired outcome. Saving Changes...
AI types are distinguished by data inputs, task complexity, and compute needs. Lumping everything together causes misallocated budgets, mismatched talent, unrealistic performance expectations, and massive technical debt. Treat generative creativity, predictive analytics, and automation as entirely different engineering disciplines Saving Changes...
PMI Team Member
