.svg)
Conviction Over Code
Why AI doesn’t create advantage, it amplifies who you already are
What 20 years of researching Leadership in the Age of Personalization reveals about how to thrive in the age of AI.
Download the PDF
Six patterns in the evidence
- AI amplifies who you already are.
In a Harvard Business Review study that surveyed 236 sales professionals at Stripe and Checkr, managers reported assigning higher-value tasks to 44% of top-quartile reps after AI adoption, versus 12% of bottom-quartile reps. The authors found the gains came less from trust in AI, which showed little correlation with results, than from top performers who taught themselves to use it as a thinking partner. The tool did not level the field. It rewarded the people who already knew how they learn and what they bring.
- Confidence is being borrowed, not earned.
Pluralsight’s 2025 AI Skills Report found 91% of C-suite executives admit to exaggerating their AI knowledge, compared with 79% of all professionals. Meanwhile, 75% of U.S. workers tell the University of Phoenix that AI increases their confidence at work. The question no survey asks is whether that confidence was built into them or borrowed from the tool.
- The repetitions that build judgment are vanishing.
IBM’s 2026 study of 8,800 employees in 28 countries found 60% worry AI is eroding their skills, most often critical thinking, and three in four of them say the erosion has already begun. BCG found that the 14% of workers experiencing what it calls AI brain fry report 33% more decision fatigue and 39% more major errors. When the tool goes first every time, it replaces the moment judgment would have formed.
- Adoption is mistaken for ownership.
Microsoft’s 2026 Work Trend Index, a survey of 20,000 AI users, found organizational factors account for 67% of reported AI impact, against 32% for individual mindset and behavior. Yet only 26% say their leadership is clearly aligned on AI, and 13% say they are rewarded for reinventing their work even when results aren’t met. Everyone gets the license. Very few are trusted to decide what it is for.
- Trust is falling while AI rolls out.
Korn Ferry found the share of non-executives confident in their senior leadership fell from 68% in 2024 to 53% in 2026. Edelman found trust in co-workers fell from 82% to 73%, a record low in its data series, and 49% of associates call their transformation experience less than positive, versus 17% of executives. People are being asked to change how they work by leaders they are less sure can see them.
- It begins long before work.
Common Sense Media rated ChatGPT for Teens an “Unacceptable Risk” after testing it with teen accounts, and found that with Study mode off it completed all 80 assignments tested. In a randomized trial at the University of Maryland, 73.8% of student requests to an AI tutor sought direct answers and 0.7% asked for feedback on their own attempt. We are raising a generation that can get any answer and is rarely asked what it thinks.
Read separately, these look like six problems. Read together, they are one.
AI is exposing how few people were ever helped to know who they are, and how few environments were ever designed to want what they think.
Assimilation, not replacement
The public debate about AI is mostly about jobs: how many survive and which go first. That debate matters, but it misses the quieter risk. The greater danger is not that AI replaces people. It is that people slowly become indistinguishable from what AI produces.
For 20 years I have watched accomplished people perform a version of themselves until the performance became the person. It often begins around 15, when fitting in feels safer than standing out. It continues in the conference room, where the polished answer is rewarded and the candid one is softened. AI is simply the newest and most powerful editor in that long chain.
A machine can only assimilate a self that was never claimed.
People who know what they naturally solve for use AI to extend their thinking. People who don’t use it to replace their thinking, and they rarely notice the difference until someone challenges the frame instead of the data.
The economic cost of a workforce that holds itself back is already visible. Gallup puts global employee engagement at 20%, its lowest level since 2020, and estimates the cost of low engagement at about $10 trillion in lost productivity, or 9% of global GDP. AI deployed into that environment does not fix disengagement. It automates it.
Borrowed confidence or earned conviction
The divide that will separate people and organizations over the next decade is not access to AI. Nearly everyone will have that. It is the difference between borrowed confidence and earned conviction.
| Compare | Borrowed confidence | Earned conviction |
|---|---|---|
| Where it comes from | A title, a tool, an approving room | Repeatedly acting on what you believe |
| How it uses AI | The tool goes first and sets the frame | Your position goes first; the tool tests it |
| Under challenge | Defends the output or folds | Explains the judgment behind it |
| What it produces | Motion: more output, the same thinking | Meaning: better decisions, original work |
| What it teaches others | Perform certainty | Say what you see |
Conviction is the earned capacity to act on what you believe before the outcome is certain. It cannot be installed in a training session, purchased with a license or mandated in a memo. It is earned one decision at a time, and AI has made the cost of skipping those decisions invisible until it is too late.
The returns follow the people, not the tools. Orgvue’s analysis of 475 Fortune 500 companies found those investing in human-led growth grew revenue 12.2% year on year, nearly twice the 6.8% of companies trying to do more with less.
Conviction Over Code: the sequence that works
The answer is not to slow AI down. It is to get the order right. Organizations that see real returns tend to move through three steps in sequence, and the sequence mirrors how conviction is built in people.
Each step depends on the one before it. Skills, transformation programs, strategy and governance all still matter; they simply stand on a foundation most AI roadmaps never mention.
What leaders must do differently
Stop measuring AI by adoption. Ask what your people now decide differently because of it, and whether the people closest to the work have been trusted to find the opportunities.
Protect the repetitions. Design roles so early-career people still write the first drafts, run the analysis and sit in the hard conversations where judgment forms. Plan the workforce around what each person naturally solves for, not only around tasks.
Make it safe to say “I don’t know yet.” A room that punishes uncertainty teaches people to let the tool speak for them.
Grade the thinking, not just the answer. Ask students what they would have said if the tool were not there.
Notice who your child is beside the screen. Ask what they find hard, what they care about and whose ideas they are building on.
Decide what you think before you ask. Credit the people whose work you build on. Know the problem you want to be known for solving.
Glenn Llopis on survival, renewal and reinvention in the age of AI
Survival
Protect what the machine cannot copy- The machine will always know more than you. Your job is to know what you stand for.
- Form your position before you open the prompt. Whoever goes first sets the frame.
- If you cannot defend your best AI-assisted work without the tool, it was never yours.
- Speed is not a strategy. A thousand fast answers to the wrong question are still the wrong answer.
- The most dangerous sentence in an AI rollout is “everyone else is doing it.”
- Never trade your judgment for convenience. Convenience compounds into dependence.
- Polished is not the same as true. Learn to tell the difference before your customers do.
Renewal
Recover what was edited out- Most of us were not born without conviction. We were edited out of it, one reasonable compromise at a time.
- Exhaustion is often the cost of performing a self you no longer believe in.
- What you set aside to fit in is usually what the next chapter needs most.
- Saying “I don’t know yet” is not weakness. It is where real thinking begins.
- The people who tell you the truth early are the most valuable people you will ever lead. Thank them in public.
- A room that rewards the safe answer will eventually be full of people who have stopped thinking.
- Renewal starts the moment you ask what you would say if no one were keeping score.
Reinvention
Move on purpose, not by mandate- Reinvention is not becoming someone new. It is reclaiming who you were before you learned to hide it.
- Change done to people produces compliance. Change built with people produces commitment.
- Don’t ask what AI can do for your job. Ask what problem you want to be known for solving, then make AI serve it.
- When everyone has the same tools, the person who knows the work becomes the advantage.
- The future will not reward the most informed person in the room. It will reward the one willing to act on what they know.
- Every new tool is a test of identity before it is a test of skill.
- Skills can be trained. Tools can be bought. Conviction has to be earned, and it is the one advantage no competitor can download.
Glenn Llopis on leadership in the age of AI
Glenn Llopis is a leadership strategist who has spent 20 years studying the question AI has now put in front of every organization: what happens to people when institutions reward sameness over who they are. He named the Age of Personalization before most leaders saw it coming, and built Leadership in the Age of Personalization, a methodology for balancing the standards that hold institutions together with the individuality that moves them forward. Its throughline is conviction, which he defines as the earned capacity to act on what you believe before the outcome is certain.
His authority is earned in operating, strategy and leadership roles, not only in research. The son of a Cuban immigrant, he took his first executive role at 25, became the youngest senior executive in the history of a 100-year-old company and reached the C-suite in his early thirties. With his father he founded Luna Rossa Gourmet Foods in 1997 and grew it into more than 6,000 stores across North America. For nearly 30 years he has advised Fortune 500 leaders through transformation.
A Forbes contributor since 2010 and a past Harvard Business Review author, Llopis has written six books on opportunity, individuality, reinvention and conviction, most recently Earning Conviction (2026), along with a healthcare leadership manifesto. Since 2019 he has designed and moderated nine executive summits with institutions including City of Hope, Clemson University, Fairfield University and the University of Phoenix, and has recorded more than 500 hours of interviews with leaders. Today he writes about AI, identity and the human advantage in Conviction Chronicles and on Forbes.com.
Download Conviction Over Code
Get the designed 17-page PDF to read offline, share with your leadership team or bring to your next conversation about AI.
Your copy is ready
Open the PDF below and save it to keep. Thank you for reading.
Open the PDF
Conviction Over Code · Glenn Llopis · theglennllopis.com
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
