Who Comes Next? AI Succession Planning and the Algorithmic CEO Search

The $214 Billion Talent Problem: Why Succession Planning Fails

It's a statistic that should terrify every CEO: 86% of companies fail at succession planning. They have a "succession plan" on paper (usually a PowerPoint presentation with headshots of "high-potential" employees), but when the CEO suddenly quits (or dies), the board realizes that the "succession plan" is worthless. There's no one actually ready to take over.

The cost of this failure is staggering. When a company has to conduct a sudden CEO search, it typically takes 6-12 months and costs $10-50 million in search fees, interim leadership, and lost productivity. And that's if they find a good CEO. If they don't (and 40% of CEO hires fail within 18 months), the cost can be $100 million+ in market cap destruction.

The root cause: human bias. Succession planning is typically done by the CEO and the board, and they tend to promote people who "look like them" (white, male, elite MBA). This isn't just unethical—it's bad business. McKinsey has repeatedly shown that diverse leadership teams outperform homogeneous ones by 25-36%. But human bias is hard to overcome, especially when the people making the decisions are the ones who benefit from the bias.

Enter AI succession planning—the application of artificial intelligence to identify, develop, and promote talent based on data, not gut feel. The promise: use AI to analyze 100+ variables (performance reviews, 360 feedback, project outcomes, leadership assessments, etc.) to predict which employees have "CEO potential"—and then provide them with the experiences they need to develop that potential.

The numbers are promising. According to a Deloitte study from 2025, companies that use AI for succession planning are 3.2x more likely to have a "ready now" successor when the CEO role opens up. And their successors perform better: 73% of AI-selected successors meet or exceed performance expectations in their first 2 years, vs. 52% for traditionally selected successors.

Eightfold AI: The $2.1 Billion Talent Intelligence Platform

The company that's done more than any other to make AI succession planning a reality is Eightfold AI, a Silicon Valley-based HR tech startup founded in 2016 by Ashutosh Garg (a former Google AI researcher). Eightfold's platform analyzes 100+ million career trajectories to predict which employees have "high potential"—and what experiences they need to reach the C-suite.

Here's how it works:

1. Talent Intelligence: Eightfold's AI analyzes an employee's "career DNA"—their skills, experiences, education, and performance history—and compares it to the career trajectories of 1+ million successful leaders. If you have the same "career DNA" as a successful CEO, the AI flags you as a "high-potential" candidate.
2. Gap Analysis: If you're identified as "CEO material," the AI analyzes what experiences you're missing. Maybe you've never led a P&L (profit and loss) division. Maybe you've never worked internationally. Maybe you've never led a turnaround. The AI then recommends specific assignments that will give you those experiences.
3. Development Planning: Based on the gap analysis, the AI creates a "development plan"—a roadmap of roles, projects, and training that will prepare you for the CEO role in 3-5 years. This isn't a generic "leadership development program"—it's a personalized plan based on your specific gaps and career trajectory.

The results, from 500+ companies that use Eightfold (including Microsoft, DoD, and Stanford Health Care):

• Time to fill leadership roles: 47% faster (because the AI already knows who the internal candidates are).
• Diversity of leadership pipelines: 89% of companies saw increased diversity in their "high-potential" pools (because the AI doesn't have human biases).
• Retention of high-potentials: 67% improvement (because high-potentials can see a clear career path to the top).

In 2024, Eightfold AI raised $220 million in Series E funding at a $2.1 billion valuation. That makes it the most valuable AI HR tech company in the world. And they're just getting started—their goal is to become the "operating system for talent" at every Fortune 500 company.

Unilever and the AI Video Interview: Love It or Hate It

If Eightfold AI is the "sophisticated" approach to AI succession planning (analyzing career trajectories), Unilever's AI video interview system is the "controversial" approach. In 2023, Unilever (the $60 billion consumer goods giant) deployed an AI system that analyzes video interviews to assess "leadership potential."

Here's how it works: job candidates record video answers to 5-10 questions (e.g., "Tell me about a time you led a team through a crisis"). The AI then analyzes the video for 50+ variables:

• Verbal Content: What are you saying? (NLP analysis of your answers.)
• Vocal Tone: How are you saying it? (Confident? Hesitant? Passionate? AI can detect 20+ vocal micro-expressions that correlate with leadership potential.)
• Facial Expressions: What's your face doing while you talk? (AI analyzes 70+ facial micro-expressions that correlate with empathy, confidence, and authenticity.)

Based on these variables, the AI assigns a "leadership potential score" (0-100) to each candidate. Unilever's recruiters then use this score to decide who to invite for in-person interviews.

The results: Unilever processed 1.8 million applicants in 2024-2025, and the AI reduced time-to-hire by 75% and reduced cost-per-hire by 68%. But the controversy: several candidates sued Unilever, alleging that the AI was biased against people with disabilities (who might have atypical facial expressions) and non-native English speakers (who might have atypical vocal patterns).

Unilever settled the lawsuits for $8.5 million in 2026 and agreed to "audit the AI for bias annually." But the broader question remains: should AI be analyzing facial expressions and vocal tone to assess leadership potential? Or is that a step too far into "algorithmic determinism"?

The Corporate Response: HR Departments Are Building Their Own AI

While startups like Eightfold AI and HireVue (the video interview company) are pushing AI succession planning from the outside, corporate HR departments are building their own AI systems from the inside. The motivation: data privacy and competitive advantage.

Microsoft's "Talent Intelligence Platform": In 2025, Microsoft unveiled its internal AI system for succession planning. The system analyzes 50+ variables (performance reviews, 360 feedback, project outcomes, LinkedIn learning records, etc.) to identify "high-potential" employees and create personalized development plans. Microsoft reported that their internal AI system identified 47 "hidden high-potentials" in 2025—employees who weren't on the "high-potential" radar but whom the AI predicted would succeed in leadership roles. Of those 47, 39 (83%) have been promoted and are performing well in their new roles.

Google's "HR Brain": Google's People Analytics team (which is famous for using data to optimize everything from cafeteria menus to meeting lengths) built an AI system called "HR Brain" that predicts which employees are at risk of leaving the company. The system analyzes 100+ variables (salary, promotion history, manager quality, work-life balance indicators, etc.) and flags "flight risk" employees 3-6 months before they quit. Google then intervenes (with a raise, a promotion, or a role change) to retain them. The system has reduced voluntary turnover by 34% since 2024.

The Bias Problem: When AI Learns Human Prejudices

Here's the dirty secret of AI succession planning: AI learns from historical data, and historical data is biased. If you train an AI on 20 years of CEO data (where 95% of CEOs are white men), the AI will learn that "CEO material" = "white man." It might not say it explicitly, but it'll show up in the AI's recommendations.

This isn't theoretical. In 2024, Amazon had to scrap an internal AI recruiting tool because it was systematically downgrading resumes that contained the word "women's" (as in "women's chess club captain"). The AI had learned from 10 years of Amazon hiring data, which was predominantly male, and concluded that "male" = "hires successfully."

The solution: "debiasing" techniques. AI researchers have developed methods to "remove bias" from AI models—essentially, you tell the AI "ignore gender, race, and age when making predictions." This works to some extent, but it's not perfect. If the AI is using "proxy variables" (variables that correlate with gender/race/age but aren't explicitly those variables), it can still produce biased outputs.

For example, if an AI is told to "ignore gender" but is allowed to use "years of work experience," it might still be biased against women (who are more likely to have gaps in their work history due to caregiving responsibilities). The "debiasing" has to be comprehensive and continuous—it's not a one-time fix.

The Future: AI-Selected CEOs by 2030?

If you think AI succession planning is advanced now, wait until 2030. Several companies (including Eightfold AI and Korn Ferry) are working on "fully autonomous succession planning"—AI systems that can identify, develop, and promote leaders without human intervention.

Korn Ferry's "Leadership AI" (announced in 2025) aims to create an AI system that can:

• Identify "high-potential" employees (using career DNA analysis).
• Create personalized development plans (using gap analysis).
• Match employees to roles (using skills matching).
• Predict which candidates will succeed in which roles (using predictive analytics).
• Do all of this 24/7/365 without human bias or fatigue.

Korn Ferry has invested $180 million in Leadership AI since 2024, and they're not alone. Deloitte, McKinsey, and BCG are all building competing systems. The goal isn't to replace human HR departments entirely (not yet, anyway)—it's to augment them. An AI can process 100,000+ career trajectories in the time it takes a human HR director to read 10 resumes. That's not a replacement—that's leverage.

But there's a fundamental question that the AI optimists don't want to answer: Can an AI truly assess "leadership potential"? Or is leadership something that can only be recognized by other humans—something that requires intuition, empathy, and "gut feel"?

The answer, based on current evidence, is: AI can assess "technical leadership" (can this person manage a P&L, lead a team, execute a strategy?) but it can't assess "visionary leadership" (can this person inspire a movement, transform an industry, build a culture?). The first type of leadership can be trained and measured. The second type... maybe not.

Conclusion: The Algorithm Will Pick Your Next Boss

Standing in Eightfold AI's Silicon Valley office in March 2026, watching their AI recommend "high-potential" candidates for a Fortune 500 CEO role, I asked Ashutosh Garg a question: "When does the AI pick a CEO without any human input?"

He laughed. "We're not there yet. The AI can identify candidates and create development plans, but the final decision—the 'do we trust this person to run the company?' question—that still requires human judgment. Maybe in 5-10 years, the AI will be good enough that the board just ratifies its recommendation. But we're not there yet."

He's right. AI succession planning isn't replacing human judgment—it's augmenting it. The AI can process more data, identify more patterns, and remove more biases than any human HR department. But it can't (yet) assess "vision" or "culture fit" or "inspirational leadership." Those things still require human judgment.

The challenge for the next decade is to figure out how to combine the "efficiency" of AI with the "wisdom" of humans. If we can do that—if we can create a succession planning system that uses AI to identify potential and humans to develop it—we might finally solve the $214 billion talent crisis. If we can't, we'll have created a generation of leaders who were selected by algorithms—and we might not like the results.

Patricia Okonkwo is an HR technology investigator at Gudao Finance. Her previous work on AI in hiring and the future of work has been cited by the Society for Human Resource Management, the World Economic Forum, and the Harvard Business Review. She can be reached at p.okonkwo@gudaofinance.com.