Minding the Gap

Welcome to our inaugural blog series, Workforce Evolved, presented by the cofounders of Zal.ai. In this series, we explore the evolving intersection of technology and strategy in the learning and development (L&D) landscape. Each post will feature insights from leading experts in the field, discussing how they integrate resources like AI and learning science, into next generation L&D strategies.

Every other week we’ll share a new post exploring the complex world of learning in an AI-age, and provide expert perspectives on navigating its challenges and opportunities. Subscribe below for exclusive content delivered directly to your inbox.

I first started thinking about this question two years ago. I was a first-time people manager, and had recently been asked to lead a small but mighty team. Not long after, one member at the start of their career – let’s call them “Chris” – asked what they could do to get a promotion during a weekly 1:1. At first, I was thrilled to hear this, because it signaled a high level of engagement, and I needed Chris to stick around, as our project was ramping up significantly. But then reality hit.

I flashed back to the times when I, as an ambitious junior employee, had asked my former managers the same thing. I could finally relate to the uneasy look in their eyes as they tried to navigate the delicate dance of wanting to signal mutual enthusiasm, but having, simply put, no good answer. If you’ve ever had a job, you’ll probably be able to relate to one, if not both of us in this situation.

That uncomfortable moment led me on a journey to create something that would meaningfully tackle this problem for organizations – from the entry-level employees hustling for their first promotion, to the managers trying to keep their teams motivated, all the way up to the CEOs setting next year’s strategic priorities. 

What if we gave managers the information and tools they needed to help employees grow? And what if we could measure that growth, to objectively understand the impact that skill development has on a company’s bottom line? Imagine how a tool like that could transform organizations.

Credit: https://www.mckinsey.com/featured-insights/sustainable-inclusive-growth/chart-of-the-day/mind-the-skills-gap

Knowing What We Don’t Know

My co-founder, Stephen Kosslyn, recently introduced me to the Dunning-Kruger effect. It’s a term coined by Justin Kruger and David Dunning in their 1999 paper published in the Journal of Personality and Social Psychology, to describe the specific cognitive bias in which people who lack a particular skill wrongly overestimate their abilities. In essence, the Dunning-Kruger effect highlights not a capability problem, but a self-awareness problem. We don’t know what we don’t know. And it turns out – ignorance isn’t bliss when it comes to efficacious learning.

This concept has interesting implications when thinking about the workforce. In the hiring process, we rely heavily on talent acquisition leaders to extrapolate a candidate’s value based  on their self-reported skills and experiences via resumes and interviews. Promotions, in my experience, are typically based on even less quantifiable data. These are some of the most important decisions an organization can make, but even the alleged “data-driven” are largely flying blind.

Let’s return to my colleague Chris for a moment. If I knew what skills our company valued, I could help guide them to develop those skills. And if I knew exactly how skilled he was, I could recommend hyper-specific learning opportunities to level him up. If he took advantage of those opportunities, I could make an excellent case for his rise at the company, especially if he had some way to demonstrate his new skills. 

Unfortunately, I had none of this.

This dearth of objective information on skill attainment cripples an organization’s ability to hire and promote the right people, form the most competent task forces for specific projects, and see ROI on L&D investments. 

So, how can we combat the Dunning-Kruger effect in the workforce?  It starts with identifying skill gaps. When we know our gaps, we can develop more tailored solutions to solve for them. 

Gap The Problem

In the spirit of combating my own Dunning-Kruger effect, I set out to better understand the landscape of assessments in workforce skill development. What solutions exist today, and what’s on the horizon?  

Many applied skill assessments in the workforce are either entirely qualitative (i.e. “survey your colleague”) OR entirely quantitative, and only narrowly applicable to certain measurable skills (i.e. programming scenarios like LeetCode). 

Some try to combine qualitative and quantitative measurements, but rely heavily or exclusively on biased data, like peer reviews. A friend of mine described having to rank 4 of her colleagues from 1-10 on certain soft skills for her company’s quarterly assessments. They gave each other all 10s as an unspoken quid pro quo. How was this effective? 

Then there’s the spectrum of time intensiveness. Think of a quick virtual quiz or survey on one end (i.e. the short quizzes that a Coursera course might offer after watching a lecture video), and on the other end – in-depth person-to-person interviews (perhaps done by an expensive consulting firm).

Our advisor, Stela Lupushor – founder of Reframe.Work and coauthor of Humanizing Human Capital, helps clarify the complexity in applying these concepts to the workplace:

“Skills assessments need to be nuanced and take into account the specific context in which the skills will be applied. A one-size-fits-all approach to assessing skills may not accurately reflect the unique requirements, cultural norms, size, industry, and goals of different organizations.”

So, an ideal performance assessment would integrate both qualitative and quantitative aspects, would pull from unbiased data sources, and would mimic the ease of a virtual quiz but with the nuanced and meaningful results of a time-intensive, in-person interview process. It would take into consideration the unique and nuanced standards of proficiency for each organization, and would be hyper-relevant to the assessment taker via personalization.

Enter: Smart Assessments

The solution to this problem became an obsession for our founding team. We spoke to hundreds of individuals in the workforce – and continuously refined our “best possible version” based on needs and constraints, all while becoming smarter about our own gap. 

Conveniently, we soon experienced one of the most exciting technological breakthroughs: the consumerization of generative artificial intelligence. Could this new technology be used to enable this aforementioned best possible version? We think so, and we call our solution Smart Assessments, which are a combination of generative AI and the latest in learning science. 

Generative AI acts as the mouthpiece of smart assessments—it’s what communicates the complexities of the assessment in real-time, personalizing questions and scenarios to the employee’s unique role and workplace, while dynamically tailoring responses to provide a sense of realism. Learning science acts as the brain—ensuring the assessments are not only adaptive, but also fundamentally sound and effective in nurturing growth.

Kosslyn notes that “an enormous amount is now known about learning and the issues that get in the way of it, but surprisingly little of that knowledge is systematically used to help people learn. Zal is building the science of learning into its DNA, and it permeates everything it does. This approach is unique, and sets Zal apart from its competitors.”

In the next blog we’ll dive into how these tools work, from broad screening to targeted diagnostics, and see how they can significantly impact individual career paths, using real world examples. Subscribe to our newsletter below to access exclusive content, and have each post delivered directly to your inbox.