As a professional, you build your skills on a very linear path: go to school, learn the skills you need, and apply them once you start working. But what happens after that? How do you keep advancing and living up to your true potential? How do you identify where you can still grow?
We can look to Atul Gawande, a renowned surgeon, for a great example of how this is done in a professional setting. He noticed that for the first five years of his career, his surgery complication rates kept going down, but then they stabilized. Basically, five years into it, he was not getting better anymore. He had reached the limit of his potential.
Instead of just accepting this, he went to one of his retired professors and asked him for feedback and coaching to identify his skills gaps. To Gawande’s surprise, instead of a glowing review, he got a page full of notes.
Gawande said that within a year of coaching, his complication rates dropped even further. The coaching helped him push through the limits that were artificially imposed on his potential.
He proved that even at the top of your field, the right feedback can help you close skills gaps you didn’t know you had.
In this article, we’ll show you how to identify skills gaps across your organization and use coaching grounded in real work to help employees close them.
How to identify skills gaps in your workforce
How do you, as a people leader, identify these skills gaps like Gawande’s professor did? We certainly can’t have managers observe people in their work all day, every day.
Previously, this was done with a large, very expensive team of consultants. They would come in, watch everyone work, and identify the skills you have and what’s missing. The problem with this is that it’s expensive and only captures one window of time.
This is where skills intelligence tools come in. The most recent developments, like CodeSignal’s agentic skills intelligence, measure people’s skills from the work they already do and compare them with what your organization needs.
The most important part of the equation at this step is the data. Gawande’s professor’s evaluation worked so well because he had:
- The proof of the quality of the work because he was watching a live surgery in front of him.
- The knowledge and experience needed to evaluate the work.
- The evidence that there was a problem that needed to be solved: complication rates stopped declining.
Gawande’s professor reviewed real work instead of relying on a self-rating or a job-title. That’s the difference between tools that measure skills and those that infer them.
How to collect the data you need to spot skills gaps
At CodeSignal, we translate collecting skills data into these three steps:
- Find observable evidence: Use skills intelligence tools that review real work artifacts, the proof of performance.
- Grade against task-specific standards: Don’t allow just any AI agent to review work. Use agents that score work artifacts against rubrics grounded in I-O psychology and built around your company’s competencies.
- Combine the evidence over time: Look for repeated evidence of a skills gap. After training, review new work artifacts to see whether the skill scores improve.
For our partners, this approach has been able to replicate a human reviewer. According to our research, 89% of work artifact skill scores align with ratings from human experts.
Curious to see how this would look for your team? See how CodeSignal’s agentic skills intelligence turns work artifacts like Gong calls, GitHub code, and Claude prompts into validated skills data.
How to address those skills gaps?
Identifying the skills gap is only half the problem. Once you have the data to define those skills gaps, you need to fill them.
Gawande had a page full of personalized coaching and feedback notes he could use to improve. How can you give this level of personalized training to your employees?
Give every employee a coach
Just like managers cannot spend their entire day watching the work their team does, they also don’t have the time to create personalized training for each and every employee.
This is where skills intelligence tools have really stepped up in the past few years, with the advancement of generative AI.
Some now give you personalized coaching options, like Cosmo, CodeSignal’s AI coach. So now, when skills gaps are detected, Cosmo builds personalized, scenario-based practice grounded in the person’s real work context.
Ensure training is useful for employees
There’s one part of this equation we haven’t talked about yet, and that’s why Gawande was able to see results from the changes he implemented. It’s because the coaching he got was applicable to the work he did day-to-day.
Again, how do we apply this at a larger scale?
We use three criteria to check whether training will work in practice. For training to be successful, it needs to be:
- Practical: Employees can realistically fit practice into their workday
- Attractive: Employees are willing to do the practice.
- Actionable: Employees can apply what they learn in their day-to-day work.
We surveyed employees from our partner study and found that Cosmo’s personalized training was 100% practical, 88% attractive, and 81% actionable.
That’s how you know when training is effective enough to close a skills gap.
How to align your skills intelligence to future workforce needs
Everything we’ve talked about so far was focused on how to identify skills gaps and close them based on skills already in your workforce.
But now we live in a world where human skills are being questioned and AI skills are said to be the future. So how do you prepare your workforce for that change?
There are a few steps you can start taking now:
- Track market data: See what new skills are coming up in job descriptions, where skills are gaining momentum, and what skills are cooling in demand.
- Set tracked skills within your organization: Define what up-and-coming skills you want your employees to have and start tracking those within your organization.
- See where your team stands: Do some employees already have those skills? Where are some teams getting ahead or falling behind? Use this data to push training to those who need it or use people with these skills in new roles and projects.
Because the truth is, yes, AI is going to displace a lot of jobs. It’s also going to create a lot of new jobs.
As CodeSignal CEO Tigran Sloyan put it at Beyond 2026, “AI is going to completely rewrite the entire equation of what it means for humans to be productive in the workplace. And that is exactly why we need to leverage that same extremely powerful technology to make it possible for all of our humans to thrive in that environment.”
Want to leverage AI to close skills gaps and advance learning? See how CodeSignal’s agentic skills intelligence prepares employees for the future of work.
This article is a written version of the Beyond 2026 keynote. Watch the full keynote here:
Questions about identifying skills gaps
How can organizations identify critical skills gaps in their workforce?
Start by defining the skills each role needs and the level of proficiency the work requires. Compare those requirements with evidence from employees’ real work, scored against consistent, role-specific rubrics. Prioritize the gaps that most affect your business goals, then use new work evidence to track progress after targeted practice and coaching.
What role does continuous learning play in addressing skills gaps?
Continuous learning helps employees build skills as their work and your organization’s needs change. Give people regular, scenario-based practice and coaching grounded in the tasks they need to perform, then check new work for signs of improvement. The aim is better performance on the job, so use demonstrated skills, rather than course completion alone, to judge progress.
What’s a data-driven method for identifying role-specific skills gaps?
Start with the tasks and skills the role requires. For a sales role, for example, score how well a rep handles objections in recorded calls against a consistent, role-specific rubric. Look for patterns across calls, use those findings to focus coaching and practice, and review later calls to see whether the skill improves.