Corporations usually look for their potential employees, their potential hires, to have a certain set of skills. That made a lot of sense before AI was as advanced as it is today. Before we had analyzed the careers of more than a billion people. Before we had used artificial intelligence to take a deep dive into the relationships among more than one million skills.

Now, there’s something called a capability matrix that can go beyond skills.

Let’s talk about the difference between skills and capabilities. A skill shows what an employee or job candidate can do now. A capability is what someone could do, even if they haven’t done it before.

When a company can understand the capabilities of every one of its workers—and each of those employees can also understand their own capabilities and potential—the company gains significant new ways to compete.

That means connecting people with the right new jobs internally. It means getting the right people together in teams, or for projects or understanding the capabilities you need to hire for the future. And it drives outcomes like reducing turnover, improving performance, and increasing diversity.

Instead of looking only for someone who knows a given computer language, for example, the company could hire someone who could pick up that language quickly. The AI can show that the person has this capability.

Or, take succession planning. Instead of hiring someone from the outside to take over from a retiring or departing employee, someone who has done the job before, a company can use AI to look within its own workforce to see who has the capability to be successful.

Capabilities shift your thinking from pedigree and from networking as a measure of potential. Instead, capabilities show “who has what it takes to do this job?” This is an important shift in mindset to increase diversity and inclusion.

Moving from skills to capabilities changes the way you hire. It changes succession planning. It increases internal mobility. And, diversity. Deep-learning AI is what makes it possible.