Most organizations roll out AI the same way. They buy the licenses, schedule the training, launch an internal campaign, and set up a dashboard to track active users. Then, a few months later, the questions begin. Why is usage flat? Why are only a handful of enthusiasts really using it? Why does the rest of the team seem to treat it as one more tool they never asked for?
The answer is rarely only technical. The tools work, and as we will see, the evidence that they can make us more productive is now substantial. What is often missing is meaning. We have spent a great deal of energy explaining what AI is and how to use it, and very little explaining why any individual should care.
This is where Simon Sinek has something important to teach us, even if his ideas were never meant for this conversation.
The Golden Circle, revisited
In Start With Why, Sinek describes a simple model he calls the Golden Circle: three concentric rings. On the outside is What: the products we sell, the services we offer, the tasks we perform. In the middle is How: the processes, methods, and differentiators that set us apart. At the center is Why: the purpose, the belief, the reason we exist beyond making money.
Sinek’s observation is that most organizations communicate from the outside in. They start with what they do, explain how they do it, and rarely reach the why. The organizations and leaders who truly inspire do the opposite. They start with purpose, and let the how and the what follow. His best-known line captures it well: “People don’t buy what you do, they buy why you do it.”
He grounds this in biology. The outer ring maps to the neocortex, the part of the brain that processes facts, figures, and language. The inner rings map to the limbic brain, which governs feelings, trust, loyalty, and, crucially, decision-making. This is why a spreadsheet of features can be entirely convincing on a rational level and still fail to move anyone. We decide with the part of us that responds to meaning, and we justify with the part that responds to data. Sinek built this framework to explain why some companies sell more and why some leaders earn deeper loyalty. But I believe its most useful application today has nothing to do with selling. It has to do with adoption, and specifically with the question every employee silently asks when a new AI tool lands on their desk: why should I?
The business case is already made
Let us start with what we know: the rational case for AI is strong, and it comes from some of the most respected institutions in the world.
In one of the most cited experiments to date, researchers from Harvard Business School, Wharton, MIT Sloan and Warwick worked with Boston Consulting Group to test GPT-4 on 758 consultants. On tasks within AI’s capabilities, consultants using it completed 12.2% more subtasks, worked about 25% faster, and produced work that human graders rated roughly 32% higher in quality. The study was formally published in the peer-reviewed journal Organization Science in March 2026.
A team from Stanford and MIT, led by Erik Brynjolfsson, studied AI in a real workplace rather than a lab. Tracking 5,172 customer-support agents, they found that an AI assistant raised productivity, measured as issues resolved per hour, by 15% on average. The paper appeared in The Quarterly Journal of Economics, one of the most prestigious journals in the field.
And at MIT, economists Shakked Noy and Whitney Zhang ran a controlled experiment published in Science. Among 453 college-educated professionals given realistic writing tasks, ChatGPT cut the average time taken by 40% and raised output quality by 18%.
The potential is proven. And yet adoption still stalls in many organizations. Why?
How we got AI adoption backwards
Look at how most AI programs are communicated, and you will see the Golden Circle running in reverse.
The What comes first: “We are rolling out a new AI platform across the company.” Then comes the How: “Here is the training schedule, here are the prompt libraries, here is the help channel.” And if a Why appears at all, it usually sounds like this: “This will make us more efficient, more competitive, and help us reach our growth targets.”
There is nothing wrong with that statement. It is true, and the research above shows it. But notice whose why it is. It is the organization’s why. It speaks to margins, revenue, and productivity metrics. It does not speak to the person sitting in front of the screen.
And when people hear only the organization’s why, they fill in the gaps themselves, often with anxiety. “More efficient” can quietly translate into “do more with less.”
“Competitive” can translate into “we need fewer of you.” Even when that is not the intention, it is frequently what people hear. Under those conditions, low adoption is not resistance or laziness. It is a perfectly rational response from people who have not been given a reason that belongs to them.
Not just a technology project
None of this means technology is secondary. Quite the opposite. No AI strategy, and certainly no agentic transformation, can succeed without strong foundations. We need secure, enterprise-grade platforms that people and organizations can trust. We need clean, well-governed data, because an AI system is only as useful as the context it can draw on. And we need a robust architecture that allows agents and tools to connect safely with the systems where real work happens.
The research confirms this too. In the MIT study, among participants who chose not to use ChatGPT in their real jobs, the third who said it was not useful mostly explained that the chatbot lacked the context-specific knowledge their work required. That is a data problem, not a people problem. Without solid foundations, even the most motivated employee will hit a wall.
But here is the trap many organizations fall into. Because the technology is advancing so quickly, and will only become more capable, it is tempting to treat AI transformation as an IT initiative: choose the platform, integrate the systems, switch it on. The technology makes transformation possible. It does not make it happen.
This is why AI transformation cannot belong to the CIO or the CTO alone. It is equally a responsibility for division heads, business unit leaders, and HR and People leaders. They are the ones who understand how work actually gets done, how roles will evolve, and what people need to make the transition with confidence rather than fear.
Training sits at the heart of this. Today there is an abundance of tools available, and one of the most common questions I hear is: which one is right for this use case, in this context? Knowing how to use AI well, and knowing when not to rely on it, is a skill that must be built deliberately. The evidence shows it pays off. In a large study across Denmark by Anders Humlum and Emilie Vestergaard, employer-led investments such as encouragement, in-house models, and training boosted adoption, narrowed demographic gaps in take-up, and increased how useful employees found the tools in their work. Employees also reported stronger gains in productivity, quality, and creativity when adoption came with active support from their employer.
If we map this back onto Sinek’s Golden Circle, the picture becomes clear.
Technology is the What: the platforms, the data, the agents. Training and change management are the How: the skills, the processes, and the support that help people use those tools well. But at the center remains the Why: what is the benefit, not only for the organization, but for each individual who is being asked to change the way they work?
We need all three rings. Strong technology without training produces expensive tools that nobody uses well. Training without a why produces compliance, not commitment. Only when the why is clear, and personal, do the other two rings truly come alive.
Because the truth is simple, and we tend to forget it. People are not motivated just because the company hands them more tools. People are not motivated just because the company will generate more revenue. People are motivated when they see a genuine benefit for themselves: in their work, in their growth, in their lives.
Finding the personal why
So what does that personal why look like? Here is what I find most striking. The same studies that prove the business case also contain, often in their less-quoted sections, the evidence for a personal one. Researchers measured not only output, but how people felt.
Reclaiming time for the work that matters. Almost everyone has a part of their job they love and a part that drains them. The report that takes three hours to format. The meeting notes no one wants to write. The tenth version of the same email. The MIT study showed what happens when AI absorbs part of that weight: ChatGPT shifted work towards generating ideas and editing and away from rough drafting, and exposure to it increased job satisfaction and self-efficacy. People did not just work faster. They felt more capable.
Growing, not shrinking. Used well, AI is not a replacement for capability; it is an accelerator of it. The Stanford and MIT study is remarkable here. Less experienced and lower-skilled agents saw a 30% increase in issues resolved per hour, and agents with two months of tenure using AI performed as well as colleagues without it who had more than six months of experience. The researchers also found that AI assistance improved customer sentiment and increased employee retention. For many people, the most compelling why is not “this will save me time” but “this will help me become better at what I do, faster.”
Feeling better at work, not worse. We often assume new technology brings stress. A field experiment led by Harvard and Wharton at Procter & Gamble suggests otherwise. Among 791 professionals working on real product innovation challenges, individuals using AI matched the performance of teams working without it. Even more telling for our purpose: those using AI reported significantly more excitement, energy, and enthusiasm, and less anxiety and frustration, than those working without it. That is not a productivity benefit. It is a human one.
Staying relevant with confidence. There is real fear around AI, and pretending otherwise helps no one. But one of the strongest personal whys is the sense of agency that comes from understanding the technology rather than being shaped by it. People who learn to work with AI tend to feel less threatened by it, not more.
When leaders start the conversation here, the dynamic changes. The question is no longer “How do we get people to use the tool?” It becomes “How can this tool serve the people who use it?” That is a fundamentally different question, and it produces fundamentally different results.
What this means for leaders
Starting with the personal why is not about better messaging. It requires a shift in how we lead adoption, and it involves every leader around the table, not only the technical ones.
It means asking before telling. Instead of prescribing use cases from the top, invite people to identify the parts of their work they would most like to change. The best use cases rarely come from a central team; they come from the people closest to the work.
It means honoring human judgment. The Harvard and BCG researchers coined the term “jagged technological frontier” to describe the uneven nature of AI’s abilities. When tasks fell outside what AI handles well, consultants using it were 19 percentage points more likely to arrive at incorrect solutions. This is not an argument against AI. It is an argument for people, and for training that teaches them where that frontier lies. Expertise, critical thinking, and judgment become more valuable, not less, and employees need to hear that.
It means creating psychological safety. People will not experiment openly if they fear that every hour saved will be turned into a reason to cut their role. Leaders need to be explicit about intentions, and then act consistently with them.
It means modeling the behavior honestly. When leaders share how they use AI, including what did not work, they give permission for others to learn without pretending to be experts.
And it means measuring what matters to people, not only what matters to the business. Active users and prompts per week are useful signals. But the more revealing question is: “What has this given back to you?” If no one can answer it, adoption will remain shallow no matter how many licenses we buy.
The question I keep asking myself
There is a question I return to often, and I want to share it openly, because I do not think we discuss it enough.
If AI makes us more productive, do we end up working less? Or do we keep working the same hours, perhaps even more, simply doing more things?
History offers a sobering perspective. In 1930, John Maynard Keynes predicted that, thanks to technological progress, his grandchildren’s generation would work around fifteen hours a week. The productivity gains he imagined largely came true. The shorter working week did not. Instead, we used efficiency to produce more, consume more, and expect more. Email made communication faster, and we responded by sending far more of it. Smartphones freed us from the desk, and the desk followed us home.
The early evidence suggests AI is following the same pattern.
The same Danish study by Humlum and Vestergaard, carried out with the University of Chicago’s Becker Friedman Institute, linked AI adoption surveys to official labor records. Covering 25,000 workers in 7,000 workplaces, they found no significant impact on earnings or recorded hours in any occupation, with average time savings of just 2.8%. Where did the time go? The hours freed up by AI were redirected to other work tasks, including fixing mistakes the AI itself had created.
An even more direct answer comes from UC Berkeley’s Haas School of Business. Aruna Ranganathan and Xingqi Maggie Ye spent eight months embedded in a US technology company of around 200 people, and published their findings in Harvard Business Review in February 2026. Employees worked at a faster pace, took on a broader range of tasks, and stretched work into more hours of the day, often without anyone asking them to. AI use was not even mandatory. People filled the space themselves.
What struck the researchers most was the gap between two experiences. In the moment of prompting and experimenting, people described momentum and expanded capability; yet when they reflected on their work overall, they described feeling busier, more stretched, and less able to disconnect. It is a single company, so it should be read as a signal rather than a universal truth. But it is a signal every leader should take seriously.
Here is the critical link back to Sinek. If every hour AI saves is immediately filled with more tasks, more meetings, and more output, we will have built a faster treadmill, not a better way of working. And when that happens, the personal why collapses. People are perceptive. If the benefit of AI flows only to the organization’s output, and never back to their own time, energy, or growth, they will notice, and adoption will quietly stall. Or worse, it will continue, and burn people out.
So what should the direction be?
I do not believe the answer is simply “work fewer hours,” nor is it “produce as much as possible.” I believe the answer is to treat the time AI gives us as a dividend that must be shared consciously. Part of it can be reinvested in higher-value work: deeper thinking, better decisions, more creative problem-solving. But part of it should return to people: time to learn, time to recover, time to be present with colleagues, customers, and the people they love.
The Berkeley researchers point in a similar direction. They propose building what they call an “AI practice”: being deliberate about the rhythm and boundaries of AI-enabled work, instead of simply accelerating because the technology makes it possible. This is exactly where HR and business leaders have a decisive role to play, alongside their technology colleagues. The choice is ours, but only if we make it on purpose. Organizations that get this right will not only see stronger adoption. They will build something rarer: trust that technology is working for their people, not just through them.
Starting with why, together
Sinek reminds us that inspiration comes from the inside out. AI adoption is no different. The research from Harvard, Stanford, MIT, Wharton, Chicago and Berkeley tells a consistent story: AI can make us faster, better, and even more energized at work, but none of that happens automatically, and none of it lasts if people do not see themselves in the benefit.
We need strong technology: secure platforms, clean data, solid foundations for the agentic era. We need serious investment in training, so people know which tools to use, when, and how. But none of it will truly land until each person can answer, for themselves, the question at the center of the circle.
So I will leave you with two questions.
First: what is your why for using AI? Not your company’s, yours.
And second: in your organization, where is the time that AI gives back actually going?
Sources
- Dell’Acqua, F., McFowland III, E., Mollick, E., Lifshitz-Assaf, H., Kellogg, K., et al. “Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality.” Harvard Business School / BCG. Organization Science, 2026.
- Brynjolfsson, E., Li, D., Raymond, L. “Generative AI at Work.” Stanford / MIT. The Quarterly Journal of Economics, 140(2), 2025.
- Noy, S., Zhang, W. “Experimental Evidence on the Productivity Effects of Generative Artificial Intelligence.” MIT. Science, 381, 2023.
- Dell’Acqua, F., Ayoubi, C., Lifshitz, H., Sadun, R., Mollick, E., et al. “The Cybernetic Teammate: A Field Experiment on Generative AI and Teamwork.” Harvard / Wharton / Procter & Gamble. Organization Science, 2026.
- Humlum, A., Vestergaard, E. “Large Language Models, Small Labor Market Effects.” University of Chicago, Becker Friedman Institute / NBER Working Paper, 2025.
- Ranganathan, A., Ye, X. M. “AI Doesn’t Reduce Work—It Intensifies It.” UC Berkeley Haas. Harvard Business Review, February 2026.
