The Strategy Room
Strategy Fails When Human Behavior Is Ignored, Just Look at AI
Most strategies do not fail on paper. They fail when the plan meets the people who have to carry it out. Right now, artificial intelligence is the clearest example of that pattern.
Strategy is a human system
Goals get set, structures get drawn, timelines get approved. Then the work reaches real people with real assumptions, real fears, and real conflict patterns, and the plan bends around them.
My doctoral research kept pointing back to one finding: performance runs on human behavior. How people communicate, how they handle disagreement, and what they believe about change all shape whether a plan gets executed or quietly stalls. Skip that layer and even a strong strategy is fragile.
Exhibit A
Why so many AI rollouts stall
AI is where this lesson is playing out most visibly, and the numbers are hard to ignore.
The gap is in the work, not the code
Most organizations are not getting a return on their AI pilots.
In 2025, MIT’s NANDA initiative reviewed more than 300 public AI deployments and surveyed and interviewed business leaders. Ninety-five percent of organizations saw no measurable return from their generative AI pilots. The researchers did not blame model quality. They pointed to tools that were never built into how the work actually gets done.
A long-standing rule of thumb in change management holds that technology is the smaller share of any transformation, and people, process, and culture carry the rest. AI has not changed that math. It has confirmed it.
The resistance is real and mostly unspoken
Worry about jobs and confidence gaps quietly stall adoption.
Pew Research Center found that 52% of U.S. workers are worried about how AI will be used in the workplace, and about a third expect it to mean fewer job opportunities for them. Most of that worry never reaches a manager directly. It shows up as slow uptake, quiet workarounds, and tools that sit unused. That is conflict going underground, the same pattern that derails any strategy.
Those concerns do not disappear because no one asked. They come out as failure to adopt. The answer is a direct, two-way conversation about what is changing, why, and what it means for each person’s role.
The workflow was never mapped
Teams choose the tool before they define the work.
Last week I co-led a workshop for small-business owners and nonprofit leaders built around one question: which of your bottlenecks are strategy problems, and which are tool problems? One participant described losing a new client because a reply took two days. The first instinct was to find an AI tool for email.
When we mapped the process, the real gaps were plain. No one owned follow-up. Every quote waited on the one person who was always busy. The calendar lived in someone’s head. A tool dropped into that process would only have automated the confusion.
AI can only follow a workflow that exists. If the process says the team gets an email and the owner sends a text instead, the automation loses track of what actually happened. You have to decide the steps, the owner of each step, and the rules before any tool can carry them out.
The pattern is always the same
AI is the current example, but the lesson reaches well past it. Reorganizations, new systems, mergers, and strategic pivots fail for the same reason when they fail: the people side was treated as an afterthought.
What this means for leaders is simple to say and harder to do. Decide what the work is, who owns it, and who checks it before you decide which tool will do it. In my own firm, every AI-assisted draft goes through a review step before it reaches me, and I am the only person who approves anything for publication. The tools move fast. Judgment stays with people.
How to lead change so it actually lands
Map the work before you buy the tool
Write down what starts the task, the main steps, who owns each one, and what done looks like. If you cannot describe the workflow, no tool can run it.
Surface the resistance on purpose
Ask directly what people are worried about. You cannot address an objection that no one feels safe enough to say.
Make the conflict productive
Treat skepticism as data, not insubordination. The person raising the hard question is often protecting you from a failure you have not seen yet.
Keep a person at the final check
Before anything leaves your organization, someone should confirm it. Automation can speed up the work, but accountability still needs a name attached to it.
Communicate more than feels necessary
In any change, people fill silence with their worst assumptions. Over-communicate the what, the why, and what it means for each person.
Shorter plans, steadier follow-through
Strategic planning is moving in the same direction. More organizations are setting aside the five-year plan that sits on a shelf in favor of shorter, adaptive cycles: a focused one-year plan, reviewed at least quarterly and adjusted as conditions change. The pressure behind that shift is real. The Nonprofit Finance Fund’s 2025 survey of more than 2,200 nonprofits found that 85% expected demand for their services to rise, while 36% ended 2024 with an operating deficit, the highest share in ten years of the survey.
A one-year plan moves faster, but it also leaves less room for drift. The plan still has to reach the people who carry it out, and that is where most plans stall. That is why, at Grey & Associates, we recommend keeping us on board for at least the first full year of a new plan. We measure success by what gets implemented: quarterly check-ins, course corrections when conditions shift, and a partner who keeps the plan on track when the daily work pulls attention elsewhere.
A question for your next leadership meeting
Before you approve the next tool, ask: is this a tool problem, or a strategy problem we have not named yet?
If you want a place to start, our free Find Your AI Starting Point map walks you through five questions about one recurring task, so you know where a tool can help before you spend on one.
FAQ
Frequently asked questions
Why do most strategic plans fail at execution?
Because they are treated as technical problems and ignore the human systems that carry them out. Performance runs on how people communicate, handle disagreement, and respond to change, so a plan that skips that layer stays fragile.
Why do so many workplace AI initiatives fail?
MIT research in 2025 found that 95% of organizations saw no measurable return from their generative AI pilots. The main cause was how poorly the tools fit into real workflows, not the technology. Unmapped processes, unclear ownership, and unspoken resistance quietly stall adoption.
Is AI resistance a technology problem or a leadership problem?
Largely a leadership problem. Technology is the smaller share of any transformation, and people, process, and culture carry most of the weight. Resistance that never gets voiced is avoided conflict, and avoided conflict derails execution.
Where should a small organization start with AI?
Pick one recurring task and map it first: what starts it, the main steps, who owns each step, and who confirms it is done. Once the workflow is clear, it becomes much easier to see where a tool can help and where a person still needs to decide.
How long should a strategic plan cover?
Many organizations now favor a focused one-year plan, reviewed at least quarterly, over a traditional five-year plan. A shorter cycle makes it easier to adjust when funding, demand, or leadership changes, as long as someone is responsible for keeping implementation on track.
How should leaders manage resistance to change?
Surface the resistance by asking directly what people fear, treat skepticism as useful data rather than insubordination, and over-communicate what is changing and why so people do not fill the silence with worst-case assumptions.
Sources
References
- Grey, K. S. (2017). Exploring the Relationship between U.S. Jamaican Immigrants’ Cultural Beliefs and Conflict Management Styles as These Correlations are Related to Their Perceived Job Satisfaction and Job Performance. FIU Electronic Theses and Dissertations. digitalcommons.fiu.edu/etd/3181
- MIT NANDA. (2025). The GenAI Divide: State of AI in Business 2025. Summary: forbes.com
- Pew Research Center. (2025). U.S. Workers Are More Worried Than Hopeful About Future AI Use in the Workplace. pewresearch.org
- Nonprofit Finance Fund. (2025). 2025 State of the Nonprofit Sector Survey Findings. nff.org
About the author
Dr. Keisha Grey
Founder & CEO, Grey & Associates
Strategy and leadership consultant based in Lauderhill, FL. 20+ years across nonprofits, government, and corporate sectors. Wiley DiSC and Five Behaviors authorized partner. $10.5M+ in grants secured and managed.
Build a strategy your people will actually execute
Book a 30-minute discovery call. No pitch, just a focused conversation about your team and your strategy.
VISIT
6721 NW 45th Court
Lauderhill, FL 33319
Last updated: September 2026
Discover more from Grey & Associates
Subscribe to get the latest posts sent to your email.