夜色直播

A Case for Talent-Based Competency Models (And How AI Can Help)

By Emily Lambert

Blog Home


The Future of Competency Models, Part 2. In our last blog post, we came down a little harsh on competency models. We aren鈥檛 going to apologize though; it was pretty well-deserved.

But we鈥檝e said it before and we鈥檒l say it again  competency models are on the cusp of reaching their prime. Why? Well, not only is there a collective shift from skills- and knowledge-based competency models to talent-based ones, but the ubiquity of AI is helping to automate the creation and adoption of these competency models. But before we dig into all that, you may be asking, what are talents?

Skills, Knowledge and Talents: What is the Difference?

In their book , Marcus Buckingham and Curt Coffman differentiate between skills, knowledge, and talents. 鈥淪kills鈥 refers to the 鈥渉ow-to鈥檚鈥 of a role; whether you know how to use MS Excel, Javascript, Photoshop, and so on. 鈥淜nowledge鈥 refers to literally knowing something, which can usually be quantified in a degree or designation, like a CPA, MBA, or PhD.

Buckingham and Coffman's model of skills, knowledge, and talents

鈥淭alents,鈥 however, are recurring patterns of thought, feeling, and behavior. Talents include innovation, persuasion, teamwork, adaptation, communication, and so on. Whereas you鈥檇 find skills and knowledge on a resume, talents are a bit tougher to decipher at first glance. But they鈥檙e definitely worth deciphering; after all, talents are 4X more predictive of how successful someone will be in a role than skills and knowledge! That鈥檚 why we argue a competency model that prioritizes talents (rather than, say, what someone went to school for or number of years experience in a particular job) is the key to predicting which talent will thrive where.

A competency model that prioritizes talents (rather than what someone went to school for or number of years experience in a particular job) is the key to predicting which talent will thrive where.

The AI Opportunity

A completely subjective approach to competency modelling (in other words, getting stakeholders in a room and asking 鈥渨hat do you think should make up this competency model?鈥) may result in competency models with an emphasis on skills and knowledge; however, consultants who bring I/O Psychology expertise into the competency model development process already prioritize a talent-based approach to competency modelling. That鈥檚 because an individual鈥檚 talents can be quantified by methods established by I/O Psychologists in the field, such as assessments that measure personality and cognitive ability.

However, as we mentioned in our last blog post, the big barrier organizations face when bringing in consultants to develop competency models is their drain on company time and resources. Therefore, talent-based competency modelling is often reserved for directors and above in Fortune 500 companies.

AI and automation disrupt this cycle.

But first, we need to clarify that AI is not magic; it cannot simply create a perfect competency model for you at the click of a button. AI needs to be fed data points (lots of them) in order to make decisions. If the data you feed AI is flawed, then the output will be flawed, too.

by training an algorithm on 10 years of its own hiring data. Reportedly, the algorithm became biased against female applicants; simply having the word 鈥渨omen鈥 (such as 鈥渨omen鈥檚 soccer league鈥) in a resume could cause applicants to rank lower.

By training AI with data inputs leveraged from I/O Psychology, rather than biased human decision-making (such as the Amazon example), suddenly the expertise of I/O Psychology can be made scalable and affordable.

But what does a talent-based competency model scaled by AI actually look like? At 夜色直播, for instance, competency models are created for jobs through a 6-8 minute survey completed by all job experts (hiring managers, HR professionals, top performers, etc.) outlining the behavioral needs of the role. 夜色直播鈥檚 AI engine, called Ultraviolet, is then able to aggregate and average the results of the multiple surveys to develop a competency model. Ultraviolet can then automatically match applicants鈥 and employees鈥 talent profiles (aggregated in a separate personality and cognitive ability assessment) to the competency model to determine fit.

Competency Models and the Future of Work

So why do we keep saying that competency models are soon to reach their heyday?

recently quoted, 鈥400-800M jobs will be displaced by technology by 2030. People think that software engineering is one of the biggest gaps in the US, but it鈥檚 actually soft skills.鈥

You鈥檝e probably heard all of the talk on 鈥渢he future of work鈥 or 鈥渇ourth industrial revolution鈥 aka an impending disruption in the workforce. Millions of jobs will soon be obsolete, just as net new roles that we鈥檝e never seen before will be springing up at an unprecedented rate. Needless to say, organizations in every industry will soon be preparing to face workforce planning and talent management challenges the likes of which have never been seen. 

Although a lot of media focus seems to predicate on the jobs that will be displaced by automation, not enough focus is on the that have yet to be invented. How can you create competency models for jobs that don鈥檛 even exist yet?

85% of jobs that will exist in 2030 have not been invented yet. How can you create competency models for jobs that don't even exist?

Skills- and knowledge-based competency models simply won鈥檛 cut it. After all, with emerging job titles including Augmented Reality Journey Builder, AI-Assisted Healthcare Technician, and Chief Trust Officer, how do you even begin to understand what kinds of education and work experience will lead to success in these net new roles?

In our last blog post, we talked about the shortcomings of gamified assessments to create role-specific competency models. These assessments require 50-100 current employees operating in one specific role in order to create a valid benchmark. But the roles that will emerge in the fourth industrial revolution have never existed before, let alone belong to 50+ employees within your organization. Game-based competency models just won鈥檛 cut it either.

All roads point to a talent-based competency modelling approach. But you must be intentional in preventing this approach from falling into the 鈥淚 think鈥 trap in other words, getting stakeholders in a room and going around the table, suggesting, 鈥淚 think communication should be ranked above decision-making.鈥 A talent-based competency model is only as valid as the tools that quantify it.

A talent-based competency model is only as valid as the tools that quantify it.

Psychometric assessments automated by AI measure role-specific and organization-wide competency models in a highly valid, efficient, and scalable way. When a brand new role springs up in your organization, you don鈥檛 need to wait weeks or months for a consultant to create a competency model, and you don鈥檛 need to spend a fortune to create competency models for the 85% of new jobs that will be erupting in your organization. You don鈥檛 have to leave hiring and talent mobility up to chance either with a data-driven competency model approach, you can generate competency models with confidence, knowing that individuals will thrive in their roles and in your organization.