What AI Can and Can't Tell You About Your Future Leaders

AI has potential — but it isn't the succession planner

AI has a lot of potential when it comes to succession planning. It can organize large amounts of information, compare competencies, identify development gaps and find patterns that might take a leadership team hours to uncover. Give an AI system performance ratings, career histories, assessments, skills, experience and job requirements, and it can help you make sense of all that information.

That's useful, but there's something we need to be careful about. The person who looks best in the data isn't necessarily the person who will become your best leader. Sometimes the person with tremendous leadership potential simply hasn't had the opportunity to prove it yet. That's why I see AI as a tool for succession planning. I don't see it as the succession planner.

AI can help you look. Leadership still has to see.

Start by asking what the data actually tells you

Before using AI to evaluate leadership potential, I would start with a very basic question: What information are we giving it? Most of the information organizations have about employees tells us what has already happened. Performance ratings, promotions, previous positions, tenure, education, assessments, completed training, manager evaluations and project experience can all be useful.

But the data isn't the person. It's a record of the things the organization happened to document about that person.

Consider two employees. One has spent five years working for a manager who writes detailed evaluations, recommends people for high-profile assignments and actively looks for development opportunities. The other has spent five years working for a manager who writes minimal evaluations and rarely puts employees forward for anything. Which employee is going to have more evidence of leadership potential in the system? Probably the first one. Does that mean the first employee actually has more potential? Not necessarily. We may simply know more about them.

This is something I think organizations need to pay much more attention to as they begin using AI in talent decisions. Opportunity creates data. People who get stretch assignments, mentoring, executive exposure and high-profile projects get opportunities to demonstrate what they can do. That creates information that eventually finds its way into talent systems. Someone who hasn't been given those opportunities may have far less evidence attached to their name. AI can see the difference in the records, but it won't necessarily know why that difference exists.

Some of the things that matter most aren't in the system

Think about the people you've worked with over the years. Who does everyone turn to when something goes wrong? Who stays calm when everybody else is getting rattled? Who can disagree with a senior executive without making the situation worse? Who gives the team credit when things go well and takes responsibility when they don't?

Then there are the people who don't have the biggest title but somehow get things done across departments because people trust them. Those things tell you something about leadership. So does the technically brilliant employee who produces terrific results but leaves damaged relationships everywhere they go.

How much of that is sitting neatly in your HR system? Probably not enough. Judgment, influence, resilience, curiosity, accountability, emotional intelligence, self-awareness, the ability to build trust and the ability to develop other people all matter when you're looking for future leaders.

That doesn't mean human observation is perfect. Managers have biases too. The answer isn't deciding that humans are better than AI or AI is better than humans. The better approach is recognizing the limitations of both and creating a succession process that forces us to question what we're seeing.

Today's best performer isn't automatically tomorrow's best leader

This is an easy mistake to make, with or without AI. Someone can be exceptional in their current position because they work hard, have tremendous technical knowledge or understand the operation better than anyone else. That's valuable, but the next job may require something completely different.

A larger leadership role may require the person to think across the entire organization instead of one function. They may need to influence people they don't supervise, make decisions without having all the information and develop other leaders instead of solving every problem themselves. There's another question we sometimes forget to ask: Does this person even want the job?

Succession planning isn't simply about identifying who performs well today. We need to look at what the future position will require and whether there is evidence that the person can grow into those responsibilities. AI can help organize that discussion, but it shouldn't make the discussion unnecessary.

Be careful about recreating yesterday's leaders

If an AI system is looking primarily at historical organizational data, it's going to find patterns based on the people who succeeded in that organization in the past. But succession planning isn't about recreating the past. We're supposed to be preparing for what's coming next.

The person who was exactly right for a leadership role ten years ago may not be what the organization needs three or five years from now. Future leaders may need to be much more comfortable with AI, rapid technological change, distributed teams, changing workforce expectations, new business models and a level of uncertainty that previous leaders didn't face.

So instead of asking, 'Who looks like our successful leaders today?' I'd rather ask, 'What is this organization going to need from its leaders tomorrow?' That's a much better place to start.

If AI gives you a name, don't stop there

Suppose an AI-supported talent system identifies three people as strong candidates for an important leadership position. I wouldn't automatically accept the recommendation, but I wouldn't dismiss it either.

I'd want to know why those three people surfaced. What information was used? What wasn't included? Have those employees had opportunities that other potential candidates haven't had? Are we looking at future potential or simply rewarding strong current performance? I'd also want to know what the leadership team has actually observed. Does the person want more responsibility? How do they respond when things don't go well? How do other people respond to them? What would they need to develop before they were ready?

And there's one question I particularly like: Would we have reached the same conclusion if AI hadn't put this person's name in front of us? The recommendation should start a conversation, not end one.

AI might actually help you find the people you're overlooking

To me, this may be one of the more interesting possibilities. Instead of asking AI to select your top three successors, use it to help broaden the conversation.

Who haven't we considered? Where is information missing? Have the same people been getting all of the development opportunities? Who demonstrates leadership without having the title? Are there employees whose potential is known almost entirely through one manager's opinion?

That turns AI into an exploration tool instead of a selection tool. Good succession planning shouldn't simply confirm who senior management already thinks is talented. Part of the process should be finding capable people the organization hasn't paid enough attention to.

Strengthen the succession process before adding AI

If the succession-planning process isn't very good to begin with, adding AI isn't going to magically fix it. Start with the positions that really matter to the organization's continuity. Determine what those jobs are going to require in the future, not simply what the people holding them today happen to look like.

Then establish reasonable and consistent ways to look at performance, potential, readiness, development needs, career interests, critical experience and future capabilities. After that, compare what's in the system with what leaders actually know about their people.

Who hasn't received a stretch assignment? Whose potential isn't very visible? Who demonstrates informal leadership? Who has never really been discussed? Are there people we've labeled as successors who have never demonstrated that they're ready outside their current environment?

AI can help organize the information and give leaders questions they may not have considered. The leadership team still has to answer them.

The goal is a better succession conversation

I think AI can make succession planning better. It can help organizations work through information faster, find patterns, question assumptions, identify development needs and possibly surface people who haven't received enough attention.

What I don't want to see is organizations handing over one of their most important leadership responsibilities and asking an algorithm to tell them who should be in charge next. Leadership potential isn't something that lives entirely in a database. Sometimes you see it when there's a crisis. Sometimes someone surprises you on a difficult assignment. Sometimes you have to give a person an opportunity before you find out what they're capable of doing. And sometimes a future leader is already sitting in the organization and nobody has noticed them yet.

AI can help you look. Leadership still has to see.

Sources

National Institute of Standards and Technology (NIST). Artificial Intelligence Risk Management Framework (AI RMF 1.0). https://www.nist.gov/publications/artificial-intelligence-risk-management-framework-ai-rmf-10

National Institute of Standards and Technology (NIST). AI Risk Management and Human-AI Interaction. https://airc.nist.gov/airmf-resources/airmf/appendices/app-c-ai-risk-management-and-human-ai-interaction/

National Institute of Standards and Technology (NIST). Artificial Intelligence Risk Management Framework: Generative Artificial Intelligence Profile. https://nvlpubs.nist.gov/nistpubs/ai/NIST.AI.600-1.pdf