Validating Real Skills in an AI World

Artificial intelligence is changing how companies recruit, evaluate skills and develop their employees. For companies looking to implement AI strategies, where should they begin?

Mike Mosgrove, head of solution engineering at CodeSignal, addressed that challenge during a thought leadership spotlight at From Day One’s Austin conference. Rather than trying to apply AI everywhere at once, Mosgrove encouraged organizations to start with specific problems and use AI where it can provide a practical solution. 

“Don’t boil the ocean,” he said. “Figure out a legitimate problem that you are facing and apply AI to that specific problem.” Mosgrove focused on three challenges organizations are facing as AI becomes increasingly embedded in the workplace.

The Rising Challenges of Recruiting

With the help of AI, one immediate challenge is the sheer number of candidates entering the recruiting pipeline. AI makes it easier for job seekers to apply for positions at scale and produce polished resumes and applications. 

At the same time, many HR teams are being asked to manage more work with fewer resources. The result can be weaker signals for recruiters trying to determine which candidates are actually qualified.

Mosgrove described how one large cryptocurrency company addressed the problem by using AI interviewers for top-of-funnel screening. Candidates received an AI interview after applying, giving recruiters another source of information before deciding whom to engage further.

But AI should not take over the hiring decision, he said. “The idea is you should never be allowing AI to fully make decisions for your company.” Instead, AI can handle some of the repetitive work while humans remain involved in the decisions that follow.

Mosgrove also emphasized the importance of a good candidate experience. “Top of funnel activity does not need to drive your entire interview process,” he said. For high-volume positions, it can provide a way to give candidates an opportunity while collecting information that helps identify stronger candidates.

Validating Changing Skills

AI is also creating an ever-changing set of skills that employers need to evaluate. CodeSignal’s talent science research shows that the number of job categories has increased tenfold in the last 50 years compared with the previous 200 years combined. New technologies are creating new jobs and new skills, while skills that once stood out on resumes have become table stakes.

“We all used to put Microsoft Word, PowerPoint on resumes. Those became table stakes,” Mosgrove said. “Now people are putting prompt engineering and Claude [as a skill]. These are skills that didn’t exist a couple of years ago.” And skills will continue to change at a rapid pace, he added.

Mosgrove described a “peanut butter jelly sandwich model” for assessing technical skills. One part involves understanding stakeholders and gathering requirements. The middle involves completing the technical task, potentially using AI. The final part involves explaining the reasoning behind the work.

He compared that last step to learning algebra. Many students wondered why teachers asked them to show their work if they were just using a calculator. “It’s not about your ability to use a tool.” The tools themselves will change, he says. 

What matters is whether someone can critically think, apply the technology and explain what they did. That philosophy can also change the way technical assessments are designed. Instead of trying to prevent candidates from using AI, employers can create simulations that reflect how people will actually work.

The goal is to make “simulation-based hiring and simulation-based learning and development as real-world as possible.”

Developing Skills After Hiring

The need to validate skills doesn’t end when someone gets the job. Mosgrove said organizations are increasingly trying to determine who among their existing employees understands AI and where additional development is needed.

He described skills as a “golden thread” running through talent strategies: Companies recruit for skills, workforce-plan around them, assess them and then develop them after employees are hired.

At one pharmaceutical company, Mosgrove talked about how assessments were used to identify employee skill gaps and learning and development programs were then used to address those gaps.

Practice is important. Completing an online course doesn't necessarily mean an employee has mastered a skill. Mosgrove described using simulations himself to practice difficult workplace conversations, giving him an opportunity to work through challenging situations before having the actual conversations.

“Practice early and often,” he said, describing the approach he uses for learning and development. “We want to apply that same logic and technology to learning and development, so that we are actually giving people safe spaces and environments to practice what they have learned.”

There are a lot of AI tools to choose from that can help validate employee skills. “Using these real-world simulations for upleveling your employees on both the recruiting the right skill set and then building the right and then continuing to build and grow on that in your current workforce,” he said.

For Mosgrove, AI is a tool that organizations can use to solve specific problems. From an HR perspective, that can mean using AI to help manage high-volume recruiting without handing hiring decisions entirely to machines. It can mean creating realistic assessments that measure skills rather than relying solely on polished resumes. And it can mean continuing to assess and develop those skills after employees are hired.

At the center of all three is the same challenge: In a workplace where AI can change both the tools people use and the way candidates present themselves, employers need stronger ways to see the skills underneath the surface.

Editor’s note: From Day One thanks our partner, CodeSignal, for sponsoring this thought leadership spotlight.

Carrie Snider is a Phoenix-based journalist and marketing copywriter.

(Photo by Josh Larson for From Day One)