The first thing leaders should do with AI | Daphne Koller

Big ThinkBig Think
Education3 min read1 min video
Mar 3, 2026|917 views|33
Save to Pod

Key Moments

TL;DR

Leaders must play with AI first to grasp its power and guide change.

Key Insights

1

Leaders gain crucial understanding by hands-on experimentation with AI, not just reports or hype.

2

Having a tech-savvy guide or AI tutor helps translate AI capabilities into practical opportunities and guardrails.

3

Basic prompt engineering is a valuable, accessible skill that unlocks significant early value.

4

Appreciating AI's potential enables leaders to design effective change management and governance.

5

An AI-driven feedback loop accelerates learning and adoption by teaching leaders how to prompt and optimize outputs.

EXPERIMENTING WITH AI: A HANDS-ON START

Leaders should begin by getting hands-on with the technology to understand what it can actually do. This means moving beyond hype and into practical experimentation—asking the AI to draft outlines, summarize data, simulate customer conversations, or brainstorm options. By using AI directly, leaders gain firsthand insight into strengths, limits, and edge cases, which is essential for credible decision-making and change strategy. If you rely only on secondhand reports, you risk underestimating the speed of disruption or misjudging what’s actually feasible in your domain. In parallel, bring in a technically adept colleague or advisor who can translate observed capabilities into concrete opportunities and guardrails for your organization. Take notes on what tasks feel effortless and where the tool struggles, then map those patterns to potential initiatives and experiments that could move value forward.

PROMPT ENGINEERING AS A CORE SKILL

Prompt engineering is a core, learnable skill that can be rapidly deployed across teams. Learning how to phrase prompts, set constraints, specify formats, and cascade outputs dramatically changes results. Leaders don’t need to be machine-learning experts; they need to practice basic prompts and encourage teams to iterate. The goal is to build a lightweight internal capability so teams can achieve high-quality outputs without requiring a data science degree. Start with simple prompts for decision summaries or project briefs, then expand to multi-step tasks, role-playing scenarios, and reusable templates that can scale across functions, creating consistent, repeatable results.

THE AI TUTOR AND FEEDBACK LOOP

There’s a powerful feedback loop in which AI helps you become better at using AI. You can ask it to teach prompting fundamentals, demonstrate best practices, and critique your drafts, steadily raising the standard of outputs. As you iterate, you identify more efficient workflows, reduce cognitive load, and surface blind spots in your thinking. To realize these benefits, you need disciplined experimentation, clear objectives, and governance to prevent overreliance or uncritical trust in automated results. The loop works best when paired with human judgment and explicit decision criteria, ensuring that automation amplifies judgment rather than replaces it.

EMBRACING THE MAGNITUDE: WHY LEADERS NEED TO UNDERSTAND POTENTIAL

Without actually using and analyzing what AI can do, leaders cannot appreciate the scale of disruption or design effective change programs. This technology can automate routine work, augment decision-making, and unlock capabilities across functions, but only if leaders understand the landscape well enough to target value. That means investing time to learn, setting a vision for AI-enabled workflows, and translating capabilities into measurable outcomes, risks, and governance structures. The first step isn’t a grand strategy; it’s an informed, iterative exploration that scales through practical wins, lessons learned, and the credibility that comes from seeing real improvements.

ACTIONABLE STARTS FOR ORGANIZATIONS

Turn exploration into action by prescribing small, concrete experiments: use AI to draft plans, summarize research, or draft communications; test prompts with real tasks; and measure impact. Pair leaders with a tech-savvy partner who can guide prompt design, keep a shared playbook, and translate findings into initiatives. Create lightweight governance—data usage, safety checks, and escalation paths. Build a library of prompt templates and case studies that demonstrate value, and set a regular cadence to review results, scale successful experiments, and retire ineffective ones. This approach grounds AI exploration in tangible outcomes and creates a sustainable path to broader adoption.

AI Leadership Quick Start Cheat Sheet

Practical takeaways from this episode

Do This

Start by playing with the AI to understand what it can do.
Bring in a tech-savvy person to help you explore basics.
Practice basic prompt engineering and let AI tutor your learning.
Watch for the feedback loop between using AI and improving your own prompts.

Avoid This

Don’t assume AI is magical without hands-on exploration.
Don’t neglect the potential and magnitude of AI when planning changes.

Common Questions

Start by playing with the technology to understand its capabilities, and consider having a tech-savvy guide to walk you through what AI can do. This helps you grasp its magnitude and potential for change.

Topics

More from Big Think+

View all 6 summaries

Found this useful? Build your knowledge library

Get AI-powered summaries of any YouTube video, podcast, or article in seconds. Save them to your personal pods and access them anytime.

Try Summify free