Succession Planning in the AI Era: How to Identify and Develop Leaders for Roles That Don't Exist Yet

Every CHRO has a succession plan. In most cases, it looks the same: a matrix of high-potential talent mapped to specific roles two, five, and ten years out. The CFO's successor is identified. The COO's likely replacement is being groomed. The head of operations has a backup, and that backup has a backup.

It's a reasonable approach to a static problem. And it would be completely adequate if the world hadn't fundamentally changed.

Here's the core problem: succession planning assumes that roles stay the same. That a "Chief Marketing Officer" in 2026 will be roughly the same job in 2029, and very similar to the CMO role that existed in 2023. It assumes career paths are linear and predictable. A finance analyst becomes a finance manager, who becomes a director, who becomes a VP, who becomes a CFO. The role exists. You prepare someone to fill it. They fill it. The system works.

Except now, in the age of organizational transformation driven by AI, that assumption is spectacularly wrong.

We call this the "Futures Mismatch Problem." Organizations are identifying and developing leaders for specific roles that won't exist in their current form by the time the leader is ready to step into them. In some cases, those roles will be eliminated entirely. In others, they'll be transformed so completely that the competencies required bear no resemblance to the original role definition.

This article outlines a fundamentally different approach to succession planning, one that works in an environment of continuous organizational change. Instead of identifying leaders for roles, we identify leaders for capabilities. Instead of developing them for specific positions, we develop them for adaptability, learning velocity, and structural thinking. Instead of preparing them for a predetermined path, we prepare them to shape their own paths.

The Futures Mismatch: Why Current Succession Planning Is Failing

Let's start with a concrete example. In 2020, a Fortune 500 company identified a high-potential manager in their supply chain organization as the likely successor to the VP of Supply Chain Operations. They invested heavily in developing her: executive education, cross-functional assignments, exposure to the board, executive coaching. By 2024, when the VP retired, she was ready to step into the role.

Except the role had fundamentally changed.

The company had implemented AI-driven demand forecasting, AI-assisted supplier optimization, and autonomous inventory management across 60% of their supply chain. The VP role that existed in 2020, managing people who made complex supply chain decisions, no longer existed. The new VP role was about orchestrating human judgment around AI recommendations, managing the organizational change required to shift decision-making authority to AI systems, leading the transformation of team member roles from "decision makers" to "AI monitors and exception managers," and navigating the ethical and compliance implications of autonomous supply chain decisions.

The high-potential manager was well-prepared to be a traditional VP of Supply Chain. She was utterly unprepared for this role. She had never done organizational transformation. She had no experience with ethical AI governance. She didn't know how to redesign roles for human-AI collaboration. She had spent four years preparing for a job that no longer existed.

This is not an isolated story. This is the new normal.

The root cause is clear: we're using a 1990s succession planning methodology in a 2026 organizational environment. Traditional succession planning works when:

• Roles are relatively stable for 3-5+ years
• Career paths are linear and predictable
• The skills required for a role are well-defined and change slowly
• Organizational structure remains largely constant
• You can accurately predict which roles will be most critical in 5 years

None of these conditions exist anymore. Roles are being redesigned constantly. Career paths are non-linear and unpredictable. Required skills are shifting rapidly. Organizational structures are in flux. And the roles that will be most critical in 5 years are often roles that don't even exist today.

We need a fundamentally different approach.

The Capability-Based Succession Model

Instead of succession planning organized around "roles," organize it around "capabilities." Instead of asking "who should become the Chief Marketing Officer?" ask "who has the foundational capabilities required to succeed in whatever marketing leadership role exists in three years, regardless of what that role looks like?"

This requires identifying the deep capabilities that predict success across multiple possible futures, not the specific competencies required for a specific role.

Capability 1: Learning Velocity

Learning velocity is the speed at which someone can master a complex new domain with incomplete information and under time pressure. It's not IQ, it's something different and more important. It's the ability to move from "I don't understand this" to "I understand this well enough to make decisions" in days or weeks, not months.

In a traditional organization, a high-potential executive might spend a year in a rotational assignment learning a new business unit before being evaluated. In an AI-era organization, that's too slow. By the time you've learned the old structure, the structure has changed.

Learning velocity has specific components:

• Pattern recognition: The ability to identify the underlying structure of complex domains quickly, to see which principles apply across contexts
• Intellectual humility: The willingness to say "I don't know" without it being career-limiting, which enables faster learning than pretending to know
• Rapid hypothesis testing: The ability to quickly form theories about how something works, test them, and update based on results
• Meta-learning: Learning how you learn, understanding your own learning process well enough to apply it in new domains

You identify learning velocity through experience, not interviews. How quickly did this person move from confusion to competence when assigned to unfamiliar domains? What's their track record on stretch assignments? How do they respond when they don't know something? Do they learn fast enough to be useful before the context changes again?

Capability 2: Adaptive Problem-Solving

Adaptive problem-solving is the ability to succeed in novel situations where the problem definition is unclear, information is incomplete, and multiple valid solutions exist. It's different from the systematic problem-solving that works in stable environments (define problem, analyze data, implement solution).

In the AI era, many leadership decisions happen in "ambiguous problem spaces", situations where you don't fully understand the problem, don't have complete information about the context, and multiple stakeholders have different definitions of what success looks like. AI governance decisions are a classic example. You don't fully understand how the AI system will behave in production. You don't have complete information about all the ways it could fail. Different stakeholders want different outcomes. And you have to make a call anyway.

Adaptive problem-solving is distinct from analytical problem-solving. It requires:

• Comfort with ambiguity: The ability to act decisively without full information, and to update decisions as new information emerges
• Stakeholder integration: The ability to create alignment across stakeholders with different mental models and different definitions of success
• Scenario thinking: The ability to hold multiple possible futures in your head simultaneously and design for multiple scenarios rather than a single predicted future
• Reversible vs. irreversible decisions: The discipline to identify which decisions are reversible and move fast on those, while slowing down on irreversible decisions

You identify adaptive problem-solving through how someone handles genuinely novel challenges. When they encounter something they've never seen before, do they panic, freeze, or escalate? Or do they stay calm, ask good questions, consult widely, and make a call? Do they move forward with imperfect information? Do they learn from what happens and update their thinking?

Capability 3: Cross-Functional Orchestration

Traditional leadership was often siloed: you were the VP of something, and you managed people within that something. Cross-functional orchestration is different. It's the ability to create alignment and drive outcomes across organizational boundaries where you have no direct authority.

AI transformation requires this capability intensely. An AI initiative touches data, technology, business operations, legal, compliance, HR, and ethics. Nobody has authority over all of those domains. Success requires orchestrating their alignment without controlling them.

Cross-functional orchestration requires:

• Systems thinking: The ability to understand how decisions in one part of the organization create ripple effects in other parts, and to design for those connections
• Influence without authority: The ability to persuade smart people who don't report to you, who have competing priorities, and who have incentives that don't align with yours
• Cultural translation: The ability to understand the mental models, language, and incentives of different functions and create connections between them
• Governance design: The ability to create decision-making structures that allow independent domains to act coherently together

You identify cross-functional orchestration capability by looking at people's track record on cross-functional initiatives. When they've had to coordinate across boundaries, how have they done it? Did they create alignment or just create more meetings? Did they drive decisions forward or create gridlock?

Capability 4: Ethical Judgment in Ambiguous Contexts

This is the capability that traditional leadership development almost never explicitly develops, and it's increasingly critical. As organizations deploy AI, make workforce decisions based on AI recommendations, and redesign processes around algorithmic decision-making, leaders face decisions with genuine moral weight and unclear precedent.

Should you deploy an AI system that improves overall efficiency but eliminates 200 jobs? Should you use AI to optimize hiring decisions knowing the training data has historical biases? Should you automate a compliance process if the automation is more accurate than humans but you can't fully explain why?

Ethical judgment in ambiguous contexts requires:

• Ethical reasoning frameworks: Not a single ethical code, but the ability to reason through difficult tradeoffs using multiple frameworks (utilitarian, deontological, care ethics, virtue ethics)
• Stakeholder empathy: The ability to genuinely understand how decisions impact different stakeholders, especially those with less power in the organization
• Moral courage: The willingness to raise ethical concerns even when doing so creates friction or delays a decision
• Consequence thinking: The discipline to think through second-order and third-order effects of decisions, especially unintended consequences

You identify ethical judgment capacity by asking about dilemmas they've faced. How did they handle situations where the "business-optimal" decision conflicted with what they thought was right? Did they find ways to do both, or did they compromise their judgment? Did they escalate ethical concerns? How do they talk about tradeoffs?

Building a Capability-Based Succession System

If capability-based succession is what you need, how do you actually build it? Here's the framework we use with organizations:

Phase 1: Capability Mapping

First, map the four foundational capabilities across your leadership pipeline. Not the capabilities required for specific roles, the foundational capabilities required to adapt to whatever future emerges. Create a simple matrix:

• Learning Velocity: How quickly does this person move from "I don't know" to "I can decide"? Rate 1-5.
• Adaptive Problem-Solving: How well do they handle genuinely novel challenges with incomplete information? Rate 1-5.
• Cross-Functional Orchestration: How effectively do they create alignment across organizational boundaries? Rate 1-5.
• Ethical Judgment: How thoughtfully do they reason through complex decisions with moral weight? Rate 1-5.

You're not looking for people who are 5/5 across all capabilities, those people don't exist. You're looking for people who are 4+ on at least two capabilities and 3+ on the others. That profile predicts adaptability across uncertain futures.

Phase 2: Development Path Customization

Once you've mapped capability levels, customize development paths. If someone is strong in learning velocity but weak in cross-functional orchestration, their development looks very different from someone strong in orchestration but weaker at adaptive problem-solving.

For learning velocity development:

• Assign them to genuinely novel domains they've never operated in (different business unit, different geography, different industry vertical)
• Give them 60-90 days to move from "I don't understand this" to "I can make decisions"
• Pair them with a mentor in that domain who can accelerate learning
• Have them present their learning to executives, which forces them to synthesize quickly

For adaptive problem-solving development:

• Assign them to lead initiatives with genuine ambiguity, where the problem definition is unclear and multiple stakeholders have different visions
• Coach them on scenario thinking and reversible vs. irreversible decisions
• Create space for them to make mistakes without career-limiting consequences
• Debrief heavily on what worked and what didn't

For cross-functional orchestration development:

• Give them ownership of initiatives that require coordinating across domains where they have no authority
• Coach them on influence without authority, systems thinking, and cultural translation
• Have them design the governance structure for the initiative, not just execute within it
• Get feedback from peer functions on their effectiveness at creating alignment

For ethical judgment development:

• Have them lead decisions with genuine ethical complexity, AI governance, workforce decisions with human impact, tradeoffs between efficiency and fairness
• Create forums where they're explicitly expected to reason through ethical frameworks, not just business tradeoffs
• Expose them to perspectives from people affected by their decisions
• Coach them on moral courage, how to raise concerns without creating gridlock

Phase 3: Context-Agnostic Advancement

In traditional succession planning, advancement happens within a functional silo. You become CFO by being progressively more senior in Finance. In capability-based succession, advancement happens through demonstrating capability to succeed in novel contexts.

Instead of promoting internally within a function, create advancement pathways that validate capability across multiple contexts:

• Advancement requires success in assignments outside your functional expertise
• Advancement requires demonstrated ability to lead in ambiguity and ethical complexity
• Advancement requires ability to orchestrate across boundaries
• Your next role doesn't have to be higher in the same function, it might be sideways into a different domain, but at a more senior level

This creates a very different talent marketplace. People advance not by being "the best person for the specific job that exists today," but by demonstrating they can adapt to whatever role emerges.

The Uncomfortable Truth: Your Best Current Leaders Might Not Be Your Best Future Leaders

Here's something nobody wants to acknowledge: excellence in traditional organizational contexts doesn't predict excellence in transformed organizational contexts. The VP of Supply Chain who was brilliant at managing people and optimizing processes might not be the person who's brilliant at managing the transformation to AI-driven supply chain decision-making.

This creates a difficult leadership situation. You have people who have been loyal, successful, and deserve opportunity. But they may not have the capabilities required to lead in a transformed organization.

The ethical and pragmatic answer is this: create three distinct paths for senior leaders.

Path 1: Transformation Leadership, For leaders strong in the four foundational capabilities, put them on the acceleration path to executive leadership. These are your future C-suite. Develop them intensively in the capabilities they need for transformed organizations.

Path 2: Functional Excellence, For leaders who are excellent at what they do but don't have strong adaptive problem-solving or cross-functional orchestration capability, create a path that values their functional expertise while recognizing it may not lead to the C-suite. Some of the best roles in the future organization will be "deep functional experts who work within a cross-functional orchestration model." These aren't consolation prizes, they're genuinely valuable roles.

Path 3: Managed Transition, For leaders whose capabilities and expertise are genuinely becoming less relevant (e.g., leaders whose entire value was managing people in rote processes that are being automated), provide generous transition assistance to next roles, whether inside or outside the organization. This is hard, but it's more humane than keeping them in roles that won't exist in 18 months.

Succession Planning for Roles That Don't Exist

There's a deeper implication of capability-based succession that we haven't fully addressed: in some cases, you're literally planning succession for roles that don't exist yet.

Consider a retail organization five years into AI transformation. The role of "Store Manager" as it existed in 2020 largely doesn't exist anymore. Store managers no longer make inventory decisions (AI handles that), don't set staffing schedules (AI handles that), don't manage pricing (AI handles that). The store manager role has transformed into something closer to "store experience orchestrator", someone who ensures the AI-driven systems are working well, handles exceptions and escalations, manages the human experience for customers and staff, and makes decisions that require human judgment about fairness and context.

When you plan succession for "store manager," you're not planning succession for a 2020 role. You're planning succession for a 2025 role that requires entirely different capabilities from what store managers did in 2020. Your current store managers may be excellent at what they do, but "excellent at managing inventory, staffing, and pricing decisions" doesn't predict "excellent at orchestrating AI systems and managing human experience."

This requires a brutally honest assessment: which of your current leaders have the foundational capabilities required for what their role will become? And for those who don't, what's the ethical and pragmatic path forward?

Implementation: Starting Today

If you're a CHRO or chief talent officer reading this, you probably have two reactions: "This makes sense, I need to do this" and "But I don't have time to rebuild my succession planning system from scratch."

You're right. You don't need to rebuild everything. Start with this:

Week 1-2: Capability Mapping, Sit down with your executive team and HR leadership. Map your top 40-50 high-potential leaders against the four foundational capabilities. Rate them 1-5. This takes a day of focused work. You'll immediately see who has the capabilities for future leadership and who may be stronger in traditional contexts.

Week 3-4: Development Customization, For your top tier (4+ on two capabilities, 3+ on others), customize development plans. Don't create generic executive development. Create specific experiences designed to develop their particular capability gaps.

Month 2: Context Assignments, Assign high-potential leaders to at least one "novel context" assignment in the next 90 days. Not a rotation in the traditional sense. An assignment in unfamiliar territory where they have to demonstrate learning velocity and adaptive problem-solving.

Ongoing: Capability-Focused Advancement, Start evaluating promotions based on demonstrated capability in these four dimensions, not just performance in their current role. When you promote someone, make it explicit: "This person demonstrated high learning velocity and cross-functional orchestration, so we're advancing them into this more complex context."

That's the starting point. It doesn't require a massive transformation. It requires a different lens on how you identify and develop talent.

The Real Succession Problem

The succession problem in the AI era isn't really about succession. It's about whether your organization can survive continuous transformation. Because once you acknowledge that roles, structures, and skill requirements are going to keep changing, the only leaders who will succeed are those with the capability to keep learning, keep adapting, and keep leading through ambiguity.

That's not something you can prepare someone for by mapping them to a specific role in five years. That's something you can only prepare them for by repeatedly putting them in novel situations, letting them struggle productively, and developing the muscles of adaptability and learning.

The organizations that will thrive in the next decade aren't those with the best succession plan. They're those with the most adaptable leaders. And the way you develop adaptable leaders is completely different from the way traditional succession planning works.

So the real question isn't "Who should be the next CEO?" It's "Who has shown they can learn, adapt, and lead well in contexts that don't yet exist?" That's the leader you need. And that's the only succession plan that matters anymore.