Laura TorresPMP| IndependentMexico City, Mexico City, Mexico
Feb 19, 2026 1:05 PM
Replying to Luis Branco
...
Great question.
When someone says “we should use AI,” the conversation is rarely about technology itself. It is usually about pressure for speed, efficiency, innovation, or competitive leverage. The first step is to clarify intent.
Three signals help distinguish what is really being asked.
First, decision proximity. Is AI automating a task, augmenting human judgment, or moving toward managing objectives autonomously? These are fundamentally different categories of work. The closer AI gets to consequential decisions, the stronger the need for governance, traceability, and explicit oversight.
Second, problem clarity. Is there a clearly defined business problem with measurable impact, or is AI being treated as the starting point? When the solution precedes the problem, misalignment and inflated expectations follow.
Third, accountability design. Who owns the outcome if an AI-driven recommendation fails? When responsibility becomes diffuse, risk scales faster than performance.
In many organizations, “AI” simultaneously means efficiency, experimentation, and cost reduction to different stakeholders. Misalignment becomes visible when decision flows and ownership are unclear. A common tipping point is when stakeholders use the same word “AI” but describe different success metrics.
The real shift is not from manual to automated. It is from “man in the loop” to “man in control.” Without deliberate design of responsibility, capability increases while accountability erodes.
Clarity of purpose, category of AI work, and ownership separates disciplined transformation from technological noise.
For sure is efficiency if we cover from my point of view accountability and missalignment Saving Changes...
My first thought process/queries are - 1. How to use AI more. efficiently to automate some manual time consuming work (statistics, graphs, etc..) that is required for C-Level Stakeholder presentations ? 2. When will AI produce more accurate output, so that I do not hv to spend time reviewing/word smith the info. , instead begin to rely on them ? Saving Changes...
Yes, I’ve definitely been in conversations where ‘AI’ meant different things to different people. I noticed it when we started talking about examples, one person meant simple automation like chatbots or autocorrect, while someone else was talking about advanced machine learning models. The moment I realized we weren’t aligned was when our explanations didn’t match each other’s assumptions. Saving Changes...
We had a stressful debate at work recently: Should we use AI to map out decades of our old codebase, or does that risk the exact IP that makes us unique? We were all saying "AI," but we meant completely different things. One person was thinking of a secure internal search tool; another was terrified our source code was being fed into a public model. Saving Changes...
When Someone Says "We Should Use AI" — Unpacking the Real Ask When a stakeholder says "we should use AI," that statement is rarely a requirement — it's an invitation to ask better questions. "AI" means something different to every person in the room, and a project manager's first responsibility is to close that gap before any solution is proposed.
The First Cut: Integrating or Implementing? Before anything else, establish the nature of the engagement:
Integrating — embedding AI into an existing process or tool
Implementing — building a new AI-driven solution from the ground up
This distinction alone changes the delivery approach, governance model, and risk profile of the entire project.
Three Questions That Must Be Answered 1. What kind of AI are we actually talking about? This defines scope. It forces clarity around:
What data is needed, and where does it come from?
What are the compliance and security considerations?
What output or outcome is the AI expected to produce?
2. Are we solving a real problem — or following a trend? This is the hardest question to ask, but the most important. Identify whether AI is addressing:
A core functional gap
An operational process inefficiency
A user interaction or experience problem
Each of these demands a fundamentally different approach. If the honest answer is "we just want to be seen as innovative," that is a valid business goal — but it needs to be named as such, not disguised as a technical requirement. 3. How do we define success — and for whom? Success is not universal. Align early on:
What does the business sponsor consider a win?
What does the end user need to feel value?
What does the operations team need to sustain it?
Misalignment here is one of the most common — and most costly — causes of project failure.
The Bottom Line Building an AI-driven system is not the hard part. The hard part is identifying the need, mapping the process, and agreeing on the outcome before a single line of code is written. That discipline is what separates a project that delivers real value from one that simply delivers. Saving Changes...
I’ve been asked this many times, and my first response is always: what do you want to achieve with AI? Once the outcome is clear, we can define the right approach, tools, and path forward.
Out of all the posts this one resonates with me the most. It's simple and to the point. We really just need to know how are you planning to use it. If it fits within our goals, policies and ethics then it's an option we can consider. AI is a tool and should be used accordingly based upon your organizations challenges, gaps and objectives. We all should use our skills to enhance these tools towards the purpose of our goals. Simply put, be familiar with how to use it and don't let it replace you but let it enhance you or your organization. Saving Changes...
Sarb LotaIT Project Manager| Own CompanyLondon, United Kingdom
I've worked in the insurance industry for sometime. A key part of this industry is actuarial analysis; building predictive models to determine future assets/liability performance. Predictive AI. Been around for literally centuries!
So the concept is not new. However, talking to people in my last firm a) they didn't realise they'd been using AI tools for along time (above, satnav rerouting etc) and b) that this 'new' AI would somehow revolutionise everything.
The general lack of understanding of AI and the speed with which it is evolving compounds the difficulty of getting everyone at a base level of knowledge. Saving Changes...
Eric BonseCobb County Medical ExaminerHiram, GA, United States
Coming from forensic investigations, I’ve seen firsthand that adding technology to a poorly understood process rarely fixes the underlying problem. My instinct when someone says, “We should use AI,” isn’t to ask which AI tool. It’s to ask where the friction exists, how success will be measured, and whether AI is the right answer at all. Sometimes the biggest gains come from redesigning the workflow before introducing new technology. Saving Changes...
1. Find the real problem first, not the buzzword. "We should use AI" usually masks a specific pain — slow turnaround, manual errors, repetitive work. I'd ask what's broken today and what metric we're trying to move (TAT, cost, error rate) before any tool gets named. 2. Check if AI is even the right fix. Sometimes a rules-based automation (RPA) or a simple process fix solves it cheaper and more reliably. I'd push back gently — "what have we tried, and why isn't that enough?" — because AI only earns its place if simpler options fall short. 3. Anchor on data and outcome, not ambition. No usable data means no AI project, period. So I'd nail down what data exists, what success looks like in numbers, and what the risk is if the model gets it wrong, before scoping anything further. Saving Changes...
I focus on the end-goal: Is it efficiency? Accuracy? Consistency? Using an agent lends consistency for me; I also use AI extensively to cross-reference and cross-check my documentation. Its expertise supports my planning; it is a sounding board for communications content. It has nearly doubled my productivity. Saving Changes...
PMI Team Member
