How to Lead with AI Without Falling Behind It

AI is now part of how work gets done. McKinsey's 2024 global survey found that 65 percent of organizations were regularly using generative AI, nearly double the rate from ten months earlier. The mistake is treating it as a shortcut that replaces thinking. The opportunity is treating it as leverage that amplifies the judgment, clarity, and leadership you already bring.

Ethan Mollick, who researches human-AI collaboration at the Wharton School, puts it plainly: you cannot know what these tools are good or bad at until you use them in your own work. Leadership through AI means staying accountable for outcomes while the tool handles repeatable tasks. You own the frame, the recommendation, and the conversation with stakeholders.

What the research shows

“My argument has always been to use it for everything, and that is how you figure out what it is good or bad at.”

Mollick's research with Boston Consulting Group found that consultants using AI on familiar tasks moved faster, but only when they stayed in the loop: evaluating outputs, correcting course, and knowing when to take work back. AI expanded options. Humans still owned the recommendation.

“As managers and leaders, you get to make these choices about how to deploy these systems to increase human flourishing.”

McKinsey's research on the human side of generative AI reinforces this. Organizations that set a people-centric talent strategy, where AI augments judgment rather than replacing it, gain a competitive edge as more work is affected by these tools.

Mollick's organizational framework is simple: you need leadership (to set incentives), a lab (to experiment safely), and a crowd (everyone using the tools and sharing what works). Leaders who hide from AI lose credibility. Leaders who paste AI output without editing lose trust just as fast.

What leadership through AI actually means

Leadership through AI means you stay accountable for outcomes while AI handles drafting, summarizing, researching, and structuring options. The tool assists. You decide.

This matters at every level. Early career professionals who use AI to think more clearly stand out next to peers who paste generic answers. Job changers who use AI to translate experience into new contexts move faster. Experienced hires who sharpen analysis without dulling their voice keep the edge that got them hired.

Mollick's co-intelligence principles in practice

• Always invite AI to the table. Use it on tasks you know well so you can evaluate the output.
• Be the human in the loop. Never send what you have not edited. Verify facts, numbers, and names.
• Treat AI like a person (but remember it is software). Give it role, context, and constraints.
• Ask it to be a skeptic. "What is wrong with this plan? What am I missing?"
• Make your recommendation visible. Leaders name a point of view. Tools do not.

What works in practice: meetings and decisions

Before the meeting

Use AI to stress-test your agenda, surface likely objections, and draft three questions that move the conversation forward. Then cut anything that does not serve the decision you need. Your job is curation, not volume.

During the meeting

Do not read AI output aloud. Use it to stay prepared: key facts at hand, options already mapped, tradeoffs named. When someone asks a hard question, pause and answer from what you know. If you need to follow up, say so clearly. That honesty reads as leadership.

After the meeting

AI can draft recaps fast. You add what matters: decisions made, owners named, risks flagged, and the one thing that still needs a human conversation. Send the version you would sign your name to.

Prompts that build leadership, not dependency

Weak prompts produce weak thinking. Strong prompts force clarity. Mollick notes that managers, teachers, and parents often prompt AI better than coders because they know how to give a person clear instructions. Try these patterns on real work this week.

• "Here is the situation and the decision we need. Give me three options with tradeoffs and a recommendation I can defend to my team."
• "Play the skeptic. What is wrong with this plan? What am I missing?"
• "Rewrite this for a busy executive: one paragraph, clear ask, no jargon."
• "Compare these two paths using criteria I care about: speed, risk, cost, and team capacity."

What to avoid when leading with AI

• Hiding AI use when accuracy or authorship matters. Transparency builds trust.
• Letting AI flatten your voice into corporate filler. Edit until it sounds like you.
• Using AI to avoid hard conversations. No tool replaces showing up in the room.
• Chasing every new feature instead of mastering one workflow that saves real time.
• Assuming AI output is correct. Mollick's research shows confident wrong answers are common.

Your weekly AI leadership checklist

• Pick one recurring task you own: status updates, briefs, client emails, or team agendas.
• Build a personal template: input you always provide, edits you always make, final check you never skip.
• Run it for four weeks and track time saved versus quality maintained.
• Share one useful workflow with your team (that is leading, not hoarding).
• Once a month, ask AI to challenge your own recommendation before you send it.

AI is not the competition. Professionals who combine tool speed with human judgment are. Leadership through AI is how you stay on the right side of that line.

Want more field guides like this? Join the waitlist for launch updates, early excerpts, and new practical resources on leadership through AI and the capabilities that still require a human in the loop.

Sources

• Ethan Mollick, Co-Intelligence: Living and Working with AI (Portfolio, 2024)
• McKinsey Global Survey, The State of AI in Early 2024
• McKinsey, The Human Side of Generative AI: Creating a Path to Productivity (2023)