The Three Questions Every Leader Should Ask Before Adopting AI

Most AI adoption decisions follow the same sequence: someone in leadership sees a demo, hears about a competitor's initiative, or gets pressure from a board member who read an article on the plane. A budget gets allocated. A vendor gets selected. A rollout gets planned. And somewhere around month four, it becomes clear that nobody asked the questions that would have made the whole thing work.

Joel Salinas has seen this pattern across organizations of every size and sector. The frustrating part is that the failure is almost always preventable — not with better technology due diligence or a longer procurement process, but with three questions asked honestly before anything gets purchased.

These questions come directly from the AI Leadership Triad: the framework Joel Salinas developed to help leaders think clearly about AI strategy. One question per Triad dimension. Each question is deceptively simple. Answering it honestly is where the real work happens.

The Scale of the Problem

Rand Corporation research published in August 2024 found that roughly 80% of AI projects fail to deliver meaningful ROI. That is not a niche finding — it reflects what is happening at scale across industries, sectors, and organization sizes. And it is not driven by technology failure. The most common cause is strategic: organizations that define the technology before defining the problem. They know what tool they want to implement but not what outcome they are trying to achieve, whether their organization is actually ready for the behavior change AI requires, or whether there is a better path to the goal than the tool they have already selected.

The three questions below are a pre-flight checklist. They do not replace rigorous implementation planning. But they catch the strategic errors before those errors become expensive. Joel Salinas has found that leaders who work through all three before any vendor conversation is held are dramatically more likely to end up with AI that actually sticks.

Question One (Adaptability): "What Would We Do Differently If We Weren't Afraid of Change?"

This is the question that reveals whether your organization is ready to adopt AI or whether it is performing readiness while staying structurally the same.

AI requires behavior change — not just process change, but actual changes in how people do their work, how they spend their time, and what they consider their core contribution. Organizations that are not genuinely ready to support that change will neutralize even the best AI implementation. Teams find workarounds. Adoption stalls. Leaders blame the tool when the actual problem is that nobody was willing to examine the underlying behavior the tool was supposed to change.

The question surfaces this risk directly. When a leadership team sits with "what would we do differently if we weren't afraid of change?" and the honest answer is "not much," that is important information. It means the limiting factor is not AI capability — it is organizational culture and leadership posture. Solving a culture problem with a technology purchase is one of the most reliable paths into the 80%.

How to use it: Put this question in front of your leadership team before any AI conversation. Give them five minutes to write individual answers before anyone speaks. The gap between private answers and first public statements is where the real strategic intelligence lives. If multiple leaders say "we would restructure how the department works" but none of them have taken any steps toward that restructuring, you have identified your actual starting point. It is not AI — it is the change that needs to happen first.

The leaders who get the most from this question are the ones willing to sit with discomfort in the answer. If the honest response is "we are not actually ready to change," that is not a failure. It is a precise diagnosis, and it saves six months of misdirected effort.

Question Two (Innovation): "Does This Serve Our Core Mission, or Does It Just Look Innovative?"

This is the question that protects organizations from the prestige trap — the significant social pressure in 2026 to be seen as an AI-forward organization, regardless of whether the specific AI initiative serves the work that actually matters.

The Innovation dimension of the AI Leadership Triad is not about novelty. It is about asking whether a new approach genuinely serves the core purpose of the organization. The test is not "is this impressive?" or "will this look good in our board deck?" The test is "does this make us meaningfully better at the thing we exist to do?"

An example from the field: Joel Salinas worked with a church operations leader who was evaluating a sophisticated AI platform for member engagement analytics. When they applied the Innovation question — "does this serve our core mission or just look innovative?" — the honest answer was complicated. The analytics were genuinely impressive, but the church's core operational constraint was not understanding member behavior in aggregate. It was the administrative burden on pastoral staff that was limiting meaningful pastoral contact. The higher-mission AI investment turned out to be a simpler tool that automated meeting notes and follow-up task generation. Less impressive to announce. Three times the actual impact on the work that mattered.

This question requires discipline, especially for leaders who are excited about AI. Excitement is not evidence of mission alignment. The exercise is to slow down long enough to make that connection explicit and specific — ideally in writing, before any vendor conversation happens.

Question Three (Creativity): "What Connections Is Nobody Else Making?"

This is the question that unlocks disproportionate value, and it is the one most leaders skip entirely.

Most AI adoption in any sector follows a template: the obvious use cases, the tools everyone else is using, the implementation approach the vendor recommends. The result is AI that delivers incremental improvement — the efficiency gains that show up in case studies and look reasonable in ROI analyses. That is not nothing. But it is not the category of result that gives organizations a genuine competitive or mission advantage.

The organizations that get dramatically better results from AI are almost always doing something that does not appear in the standard playbook. They are applying a framework from a different industry. They are using an AI capability in a context nobody in their sector thought to apply it. They are making a connection between two organizational challenges that everyone had assumed were separate problems.

The Creativity dimension of the AI Leadership Triad is about cultivating exactly this capacity. Not creativity in the artistic sense, but the discipline of looking at your organization's real challenges through lenses borrowed from outside your immediate world.

An example: A humanitarian organization that Joel Salinas has worked with was using AI for donor reporting — the obvious, templated use case. A leader on their team who had previous experience in supply chain logistics suggested applying the same AI capability to their field resource deployment, using predictive routing logic the way freight companies optimize deliveries based on real-time conditions rather than predetermined schedules. Nobody in their sector was doing it. In the first quarter, they saw a 34% improvement in field resource utilization. The creative connection came from outside the sector entirely. The question "what connections is nobody else making?" was what made space for it to surface at all.

This question does not guarantee a breakthrough. But it consistently produces better options than going straight to what everyone else is doing. Give your team thirty focused minutes with this question before finalizing any AI strategy. The answer you are looking for is almost never the first thing anyone says.

How to Use These Three Questions as a Leadership Team

Joel Salinas recommends running these three questions as a structured leadership conversation — not an email thread, not a pre-read document, but a facilitated discussion with the people who will own the AI initiative and the people who will need to change their behavior for it to work.

Set ninety minutes. Put the three questions on a whiteboard. Give everyone five minutes to write their honest individual answers in silence before anyone speaks. Then discuss each question in sequence, spending thirty minutes on each.

The output of this conversation should be three things: a clear statement of what organizational change is actually required (not just what tool you want to buy), a specific mission-aligned use case with a measurable outcome, and at least one non-obvious connection your sector is not making. If you cannot produce all three, you are not ready to move to vendor selection. That is a feature of the process, not a bug.

As Joel Salinas covers in the 90-day AI roadmap guide, the decisions made in the first two weeks of any AI initiative determine the majority of its outcomes. These three questions belong in those two weeks — before any purchase, before any pilot, before any announcement to the broader team.

If you want to run this exercise with a facilitator who can help surface honest answers and translate them into a concrete AI roadmap, the right next step is a free discovery call with Joel Salinas. Ninety minutes of preparation now saves months of recovery later.