Brew Markets – “Prof. Karim Lakhani: Revisiting the Corporate Challenge of Adopting AI”

Podcast: Brew Markets
Host: Anne Berry
Guest: Prof. Karim Lakhani (Harvard Business School)
Date: December 26, 2025

Episode Theme & Purpose

This episode revisits Anne Berry’s October conversation with Professor Karim Lakhani, a leading scholar on digital transformation at Harvard Business School, to explore how businesses are really adopting AI—and the often overlooked challenges along the way. They break down Lakhani’s latest research—including real “clinical trials” of AI at major corporations—discuss the productivity frontier, what the “one-person unicorn” future could look like, and ask: Who should be policing AI’s ethics? Throughout, Anne and Karim get candid (and occasionally witty) about hype versus real organizational change, offering practical insights for leaders and investors wondering how to make the promise of AI pay off.

Key Discussion Points & Insights

1. AI as a “Drug”: Why Caution and Experimentation Are Needed

• Clinical Trials for AI: Prof. Lakhani likens generative AI (“geni”) to a new drug, arguing that we still don’t truly understand the “dose,” efficacy, side effects, or toxicity of this technology in organizations.
  • Quote: “We actually don’t know the right dose to take, the efficacy of the product, the side effects, the toxicity or even the right diet to follow when we’re using this drug.” —Prof. Lakhani (03:37)
• Lakhani’s team runs randomized trials in companies—mirroring clinical drug trials—to test where AI works, when it doesn’t, and how to measure causality.

2. Real-World AI “Clinical Trials”: Procter & Gamble and BCG

• P&G Trial Findings:
  • AI transitioned from tool to “teammate.”
  • Three primary effects found:
    1. Productivity boost (more work done, akin to adding teammates)
    2. Fills in gaps in expertise (e.g., marketers gain R&D knowledge via AI)
    3. Increases positive emotions at work—contrary to common fears.
  • Quote: “People feel great working with AI, unlike ... buzz about the fact that people feel depressed ... people have more positive emotions and less negative emotions.” —Prof. Lakhani (06:24)
• Boston Consulting Group Study:
  • Investigated both creative and analytical consulting tasks.
  • Jagged Frontier discovered: AI shines in creative tasks (massive ~40% productivity boost), but can harm in strictly analytical ones (e.g., spreadsheet analysis saw performance drop by 20% due to hallucinations).
    • Quote: “AI has this jaggedness ... In some cases it’ll be incredible, in other cases ... you’ll bleed.” —Prof. Lakhani (09:21)

3. The Elusive ROI, the J-Curve, and Workflow Overhauls

• AI’s Real Impact: Immediate ROI is dubious without deeper process changes; adopting AI typically causes performance to dip before recovering—a classic “J curve” common to new technologies.
  • Example: One company slashed marketing message creation time from 100 hours to 1.5 hours, with a 15% click-through rate improvement, but only after reimagining workflows end-to-end.
  • Quote: “If you come in blindly saying, ‘I’m going to expect ROI without changing workflows,’ you’re going to be in a load of trouble.” —Prof. Lakhani (11:55)
• Change Management Over Tech: 70% of successful AI adoption is change management and leadership; only 30% is technical.
  • Quote: “Tech for me is actually getting easier. And 70% change management, leadership, providing the right mechanisms, rethinking jobs, rethinking workflows.” —Prof. Lakhani (12:13)

4. Hands-On Leadership: Executives Must “Get Their Hands Dirty”

• Leaders—regardless of rank—must use AI directly to set credible direction and understand its nuances. Many executives posture as strategic but don’t engage directly with the technology.
  • Quote: “It’s only when you actually use the tool will you understand both its capabilities, its shortfalls.” —Prof. Lakhani (13:35)

5. The AI Boss, the One-Person Unicorn, and the Ant Group Case

• Algorithmic Management Is Already Here:
  • Everyday services (Uber, Instacart) are managed by algorithms from hiring to task allocation.
  • AI-powered agents could increasingly become “the boss”—potentially even more objective than humans.
• Ant Group Example (Alipay):
  • By 2019: 1.2B users, 10,000 employees—a business model organized for scale, efficiency, and minimal human intervention.
  • Mantra: “Three, one, zero”—3 min per transaction, 1 min approval, 0 human intervention.
  • Demonstrates how future firms can achieve enormous scale with far fewer humans.
  • Quote: “The basis for competition is not people doing tasks, it’s organizing machines to do all the tasks.” —Prof. Lakhani (18:09)
• One-Person Unicorn?
  • Not here yet—at least a decade away, says Lakhani, but the trend is empowering non-tech founders and delaying the need for technical hires.

6. GenAI Eating Software: The Changing Value of Technical Skills & Academia

• The original kings of SaaS (Salesforce, Adobe) now face tough questions about whether generative AI will eat their thrones.
• Academia risks “brain drain” to industry as bleeding-edge CS research requires industrial-scale data, compute, and talent. However, institutions adapt by deepening industry ties, and Lakhani sees a future with differentiated research focuses.
  • Quote: “Increasingly the bleeding edge of computer science is in the companies, and not inside academia. And that creates a big, big risk for us.” —Prof. Lakhani (22:07)

7. Who Polices AI? Ethics, Incentives, and Market Forces

• Ethical Responsibility: Both business execs and academics must ensure AI is developed and deployed with care for human impacts.
• Is Anyone Listening? With OpenAI and others setting rules (e.g., user data opt-out), will academic voices carry weight?
• Incentives & Copyright:
  • Lakhani describes Cloudflare’s initiative to fairly compensate content creators for training AI (drawing analogy to Napster → iTunes/Spotify).
  • Market-based solutions are essential to avoid “AI slop”—unverified/generated content devoid of proper incentives.
  • Quote: “We just can’t steal the content. We have to pay for that. And I think opt out ... is not the most ethical way to deal.” —Prof. Lakhani (27:13)

8. Are We in an AI Bubble?

• Explosive capital expenditure (~$500B among big tech) indicates hype, but Lakhani judges the productivity frontier as genuinely shifting.
  • The risk: The tech’s capabilities are advancing exponentially, but businesses absorb them linearly—causing a widening gap.
  • Quote: “The technology will improve exponentially, we’ll be absorbing linearly, and there’ll be a big gap.” —Prof. Lakhani (29:22)

9. Bright Spots: Who’s Doing AI Adoption Right?

• J.P. Morgan: Jamie Dimon’s leadership, putting both CTO and Chief AI Officer on the operating committee.
• Procter & Gamble: Effective AI deployment in supply chain, marketing, and R&D.
• But: “The average firm executive team is mostly lost.” —Prof. Lakhani (31:23)

Notable Quotes & Memorable Moments

• “AI is transforming from a tool to a teammate.” —Prof. Lakhani (05:38)
• “AI has this jaggedness ... in some cases it’ll be incredible, in other cases ... you’ll bleed.” —Prof. Lakhani (09:21)
• “If you come in blindly saying, ‘I’m going to expect ROI without changing workflows,’ you’re going to be in a load of trouble.” —Prof. Lakhani (11:55)
• “It’s only when you actually use the tool will you understand both its capabilities, its shortfalls.” —Prof. Lakhani (13:35)
• “The basis for competition is not people doing tasks, it’s organizing machines to do all the tasks.” —Prof. Lakhani (18:09)
• “Increasingly the bleeding edge of computer science is in the companies, and not inside academia.” —Prof. Lakhani (22:07)
• “We just can’t steal the content. We have to pay for that. And I think opt out ... is not the most ethical way to deal.” —Prof. Lakhani (27:13)
• “The technology will improve exponentially, we’ll be absorbing linearly, and there’ll be a big gap.” —Prof. Lakhani (29:22)
• “The average firm executive team is mostly lost.” —Prof. Lakhani (31:23)

Key Timestamps

• 03:33 — Anne introduces Prof. Lakhani; “AI is like a drug ...”
• 04:21 — “Clinical trials” of AI explained
• 05:38 — Procter & Gamble study: AI as teammate, key effects
• 07:12–09:45 — BCG study: “Jagged frontier” of AI effectiveness
• 10:22 — Capex, J-curve, ROI and transformation implications
• 12:13 — 30/70 rule: Tech vs. organizational change
• 13:35 — Leaders using AI themselves
• 14:39 — Algorithmic management—“the AI boss”
• 16:51–18:31 — Ant Group: The scale-up with minimal humans; 3-1-0 mantra
• 19:39–21:17 — Prospects for a “one-person unicorn”
• 21:53–24:13 — GenAI’s impact on legacy software & academia’s brain drain
• 24:26 — Ethics and AI—who takes responsibility?
• 27:13 — Market incentives, copyright, “AI slop”
• 28:41 — Is this an AI bubble? Why absorption may lag technology
• 30:35 — Who’s getting AI adoption right?

Tone and Style

Anne Berry maintains a practical, insightful, and occasionally humorous approach, pushing Prof. Lakhani to offer clear, actionable advice and honest reflections. Prof. Lakhani’s tone is candid, data-driven, and pragmatic, frequently anchoring discussions with academic rigor but also keen real-world anecdotes.

Takeaways for Listeners

• Adoption of AI isn’t a straight line—expect bumps, process overhauls, and real work to capture value.
• Success with AI is more about organizational change than software contracts.
• Leading organizations are investing hands-on, changing workflows, and working from the top down; most firms are still far behind.
• Ethical innovation needs market incentives—not just goodwill or opt-out policies.
• The AI revolution is indeed real, but translating it into lasting advantage requires leaders to “get their hands dirty” and bridge the exponential gap between technology and business capabilities.