For a long time, hiring has been built around a fairly simple assumption: past experience predicts future performance.

How many years have you done the job? Have you worked in our industry? Have you sold to our buyer? Have you used our technology? Have you managed a team this size? Have you solved this particular problem before?

Those are all reasonable questions, and experience isn't suddenly becoming irrelevant. But I think AI is changing the relative value of experience compared with two other attributes that have historically been much harder to measure: learning velocity and leverage.

When knowledge is increasingly available on demand, what someone already knows still matters, but how quickly they can learn what they don't know matters more. And when one person can use AI to research, analyze, code, create, automate, and execute work that previously required significantly more time or people, individual productivity isn't the only thing worth measuring. We also need to understand how effectively that person creates leverage.

That leads to what I think may become an increasingly important talent equation:

Learning Velocity × Leverage

Instead of only asking, “What does this person know?” companies should also be asking, “How quickly can they learn what they don't know?”

And instead of only asking, “How much can this person accomplish?” we should be asking, “How much can this person multiply what they're capable of accomplishing?”

Those are very different hiring signals.

Learning Velocity

Imagine you're hiring a Product Leader. Candidate A has spent the last eight years building products in your exact category. They understand the customer, the competitive landscape, the technology, the terminology, and many of the product decisions your company is likely to face.

Candidate B is an exceptional Product Leader, but comes from an adjacent industry and has never worked in your specific market.

Historically, Candidate A starts with an enormous advantage. They walk in on Day One with years of accumulated context and pattern recognition that Candidate B might take months, or even years, to develop.

That experience still has real value. AI doesn't instantly manufacture years of judgment.

But AI may dramatically compress the time required to acquire the knowledge that informs that judgment.

Candidate B can now use AI to analyze thousands of customer reviews, synthesize support tickets and customer calls, study every competitor, understand unfamiliar technology, interrogate market research, analyze industry trends, and develop a working mental model of the customer and market in a fraction of the time it once required.

Candidate A may still be the better hire. That's not the point.

The question I'm increasingly interested in is: How quickly can Candidate B close the gap?

And what happens a year from now when the product enters a new market, customer behavior changes, a new competitor emerges, or a technology shift suddenly forces both candidates into unfamiliar territory?

At that point, what they knew when they were hired matters less. Their ability to learn what they don't know matters more.

That's learning velocity.

And I think companies should start testing for it deliberately.

Give Candidates Something They Don't Know

Most interviews do the opposite. We intentionally look for candidates who have already encountered the problem we're hiring them to solve, and then spend the interview asking them to tell us about those experiences.

There's nothing wrong with that, but it mostly measures what someone has already learned.

If you want to understand learning velocity, introduce something unfamiliar.

If you're interviewing a salesperson, give them a company in an industry they've never sold into. Give them your product and 45 minutes. Ask them to figure out whom they would sell to, what problems those buyers probably care about, what they still don't understand, and how they would approach the account. Tell them they can use whatever AI tools they want.

You're not looking for perfection after 45 minutes. You're watching the trajectory.

What did they decide they needed to understand first? What questions did they ask? How did they distinguish useful information from noise? Did they blindly accept what AI told them, or did they challenge it? Did they discover something halfway through the exercise that caused them to change their initial hypothesis?

You can do the same thing with almost any role.

Give a Product Manager a product category they've never worked in, along with customer reviews, usage data, and a few competitors. Ask them to develop an initial hypothesis about the biggest customer problem and what they would investigate next.

Give a Marketing leader a market they haven't worked in and ask them to develop an initial point of view on the ICP, buyer pain points, positioning, and the first experiments they would run.

Give an engineer a framework or API they haven't used and ask them to learn enough about it to solve a real problem.

Give an operations candidate a business process they know nothing about and ask them to understand it, identify bottlenecks, and recommend how to redesign it.

In each case, you're testing something that doesn't show up very well on a resume: how quickly someone can move from unfamiliarity to useful judgment.

Experience Is a Snapshot. Learning Velocity Is a Trajectory.

This is why I think the resume may become slightly less predictive over time.

A resume is largely a record of what someone has already learned. It tells you where they've worked, what they've been exposed to, what responsibilities they've held, and what they've accomplished. Those are important signals.

But the faster work changes, the more important another question becomes: How quickly can this person become good at something that isn't on their resume yet?

The job you're hiring someone to do today may not be the same job they're doing 18 months from now. Their tools will change. Their workflows will change. Their market may change. AI capabilities will certainly change. Parts of their role may disappear while entirely new responsibilities emerge.

In that environment, there is enormous value in someone who can repeatedly close the distance between what they know today and what they need to know tomorrow.

The Second Signal: Leverage

Learning velocity is only half of the equation.

The second half is leverage.

AI is giving individual employees access to capabilities that historically required more people, more specialized expertise, or considerably more time. A salesperson can research hundreds of accounts. A marketer can generate and test dozens of ideas. An analyst can interrogate enormous datasets. An engineer can prototype much faster. An operations leader can build an automated workflow instead of repeatedly executing a manual process.

But using AI to complete the same work faster isn't necessarily the highest form of leverage.

The more interesting question is whether someone recognizes opportunities to change how the work gets done altogether.

Imagine giving two candidates the same assignment: Here are 100 target accounts. Determine which accounts the sales team should prioritize.

Candidate A uses AI to research the accounts faster. They produce an excellent list and explain their reasoning.

Candidate B takes a different approach. They first determine which signals should predict account quality. They build a repeatable research methodology around those signals, use AI to gather and synthesize the information, create a scoring model, test it against several accounts, and leave behind a process they can apply to the next 1,000 accounts.

Both candidates completed the assignment.

But Candidate B did something else: they created leverage.

They didn't just solve today's problem. They changed the cost of solving tomorrow's version of the problem.

Test Whether Candidates Can Multiply Themselves

This doesn't need to involve a lengthy case study or take-home assignment. In fact, I don't think it should.

You can learn a lot about someone's instinct for leverage with a 10-minute problem inside the interview. Give them a familiar business challenge, allow them to use AI if they want, and ask how they would approach it if the constraint suddenly became much larger.

For a Customer Success leader, ask: “Your team manages 100 customers today. Tomorrow that becomes 500, but you can't add headcount. What would you change?” You're listening for whether they simply redistribute accounts and prioritize harder, or start thinking about automated risk signals, AI-generated account intelligence, proactive interventions, and where human attention actually creates the most value.

For a finance leader, ask: “Your team spends the first three days of every month building the same management report. How would you get that as close to zero as possible?” The goal isn't for them to design the solution in the interview. You want to see whether their instinct is to make the existing process slightly faster or question why humans are repeatedly doing it at all.

For a Product Leader, ask: “You have 1,000 pieces of customer feedback and two hours before a product meeting. How would you figure out what actually matters?” A strong answer might involve using AI to categorize themes, identify anomalies, compare feedback across customer segments, and surface evidence—while reserving judgment about what should actually be built for the human.

For a recruiter, ask: “You wake up tomorrow with twice as many searches. You can't work more hours, and you can't hire anyone. What changes?” Do they simply prioritize harder, or rethink research, sourcing, candidate evaluation, outreach, scheduling, and administrative work around AI?

You can make the same question even simpler for almost any role:

“If your workload doubled tomorrow but your team couldn't, what would you do differently?”

That's a question I think could become incredibly revealing in an AI-native interview.

You're not looking for the candidate who says “AI” the most. You're looking for the person who naturally decomposes the work, identifies where human judgment matters, and finds ways for technology to absorb or eliminate everything else.

You want to see whether their instinct is to work harder, or create leverage.

AI Usage Isn't the Signal

This is why I'm increasingly skeptical of interview questions like, “How are you using AI?”

Eventually, asking whether someone uses AI will sound a little like asking whether they use Google.

Almost everyone will.

Two employees can have access to exactly the same AI models and create wildly different amounts of value with them. One might use AI to write emails 30% faster. Another might use the same technology to rethink how an entire workflow operates.

The difference isn't access to AI. It isn't even necessarily technical proficiency.

It's the instinct to create leverage.

Some people naturally look at repetitive work and ask how to automate it. They encounter a problem once and immediately think about how to prevent anyone from having to solve it manually again. They build systems, create reusable processes, codify knowledge, and use technology to extend their capacity.

AI makes that instinct significantly more valuable.

A Better Interview Process

If I were designing an interview specifically to identify AI-native talent, I would build it around three stages.

The first stage would test learning velocity. Give the candidate a problem containing meaningful unfamiliarity. Don't evaluate them based on how much they know when the clock starts. Evaluate how far they travel by the time it ends.

The second stage would test leverage. Give them more work than they could reasonably complete manually. See whether they simply try to work faster or whether they redesign the problem, intelligently delegate work to AI, automate pieces of it, and create something reusable.

The third stage would test adaptability. Halfway through the exercise, change something important. Tell the salesperson the buyer they targeted isn't actually the economic buyer. Give the Product Manager new customer data that contradicts their original thesis. Cut the Marketing leader's hypothetical budget in half. Tell the engineer that a major requirement has changed.

Then see what happens.

Do they defend their original answer, or do they update it? Can they quickly determine which assumptions are now invalid? Can they redirect their AI tools and workflow without starting over?

Learning quickly isn't only about acquiring new information. It's also about letting go of old information when reality changes.

Don't Just Grade the Answer. Grade the Process.

At the end of the exercise, I'd spend as much time discussing how the candidate worked as what they produced.

Ask them what they needed to learn first and why. Ask what they delegated to AI and what they intentionally kept for themselves. Ask where AI gave them a bad answer and how they recognized it. Ask what they verified independently. Ask which part of their process they made repeatable.

Then ask two questions I think are particularly revealing:

“If you had to do this 100 times, what would you change?”

And:

“What do you believe now that you didn't believe when you started?”

The first reveals leverage.

The second reveals learning velocity.

Hire for the Slope

I don't think experience is going away. Deep expertise will continue to matter. Judgment will matter enormously. There will be jobs where years of accumulated domain knowledge remain an extraordinary competitive advantage.

But AI changes the value of being able to rapidly acquire new knowledge and multiply what you can do with it.

That's why I think companies should increasingly evaluate talent across two dimensions: learning velocity and leverage.

The employee with high learning velocity can enter unfamiliar territory and become useful unusually quickly. The employee with high leverage can take what they know and build systems, workflows, automations, or tools that multiply their impact.

The person who possesses both is particularly interesting.

They don't need to know everything when you hire them because they have demonstrated that they can figure things out unusually fast. And once they figure something out, they don't simply perform the work, they look for ways to make that capability scalable.

For years, hiring has focused primarily on determining where someone is today.

Their experience. Their skills. Their knowledge. Their accomplishments.

Those things still matter.

But in an environment where technology is changing what people can do every few months, I think we should become much more interested in the slope.

How quickly is this person becoming more capable?

How quickly can they close a knowledge gap?

How much can they multiply themselves?

And how much more capable could they be a year after we hire them?

The most valuable employee of the AI era may not be the person who knows the most on Day One.

It may be the person who can learn faster than the environment changes and create disproportionate leverage from everything they learn.

That's the new talent equation:

Learning Velocity × Leverage.