Wharton partners with consulting company to model skill valuation for jobs in an AI era

The Wharton School’s AI and Analytics Initiative recently developed a benchmark to track the supply and demand of skills in the workplace.

Based on an analysis of millions of United States worker profiles and job postings, the index — developed in partnership with the technology consulting firm Accenture — calculates the value of specific skills by linking job-posted requirements to actual salaries. Researchers will regularly update the project to reflect changes in the evolving job market. 

“We know that AI is disrupting the workforce, but it’s very unclear how that disruption is happening,” second-year Wharton MBA student Christian Sitarz — who served as the business engagement lead on the Wharton-Accenture Skills Index — said. 

Sitarz explained how the increase in the use of AI tools results in “different supply and demand fluctuations,” which WAsX was designed to monitor — showing “companies how to construct their next recruiting cycles.”

“It’s really valuable information for a C-suite to have in order to deploy capital,” Sitarz said.

Wharton Vice Dean of AI & Analytics Eric Bradlow said the project “really came from Accenture’s vision,” and frames career recruiting as “skills-based” rather than “job-based.” 

“Let’s not worry as much about what your job title is — let’s worry more about what skills you have, and what skills are needed to do certain jobs,” Bradlow described.

The WAsX team used two major datasets to create the skills index — national worker profiles and job postings scraped across the U.S. economy. Researchers created LinkedIn-style profiles to compare with actual occupational distributions across the country.

Each skill cluster was evaluated by supply — how often workers listed the skill — and demand — how often employers required it. The model also added a third factor, which measured the correlation between skill and average salary. 

Wharton sophomore and WAsX Senior Analyst Christine Lam and Wharton and Engineering senior and WAsX Technical Lead Andrew Mao spoke about the methodology and techniques behind the index.

“We use Shapley additive values in order to actually extract average future impacts,” Mao explained. “This gives us the average dollar amount that skills either increase or decrease a salary by, and that’s how we got the economic value.”

Mao said if employers “multiply gaps by the economic value of skills,” they could see the benefits of training employees to "have certain skills.”

Lam spoke about WAsX’s “valuation model,” explaining the technology tells employers “which skills you are paying more for and which skills you are paying less for.”

“I think that’s super helpful to employers, because they can see that automating some of these things saves money,” Lam said. 

Lam suggested employers could use the index when hiring to find people with “much more niche” skills that could “create so much value for the company.”

Researchers also observed shifts in the hiring process amid the rising use of generative AI. 

According to the report, some writing-related skills appear to be declining in employer demand, while more advanced technical skills — such as machine learning engineering — showed strong and rising valuation.

Engineering senior and WAsX Senior Analyst Clarice Wang said the rise in people “putting down skills like using generative AI” on their resumes and applications calling for “use of generative AI or large language models as a skill” is related to the broader industry shift towards “adapting to using” AI in “day-to-day” operations. 

“Being able to accelerate and optimize specific processes within each industry is becoming more prevalent,” Wang said. “If I wanted to adapt my own resume and improve it, I would want to put more technical skills over communication and leadership skills.”

WAsX Senior Data Analyst and Wharton AI and Analytics Initiative Staff Scientist Russ Walters clarified that the model’s skill valuations are context-dependent. He described how a skill emphasized in one industry may have low marginal value in another. 

“The answer is not, ‘Everyone should get skill X.’ That’d be a misread of the data,” Walters said. “The next best incremental skill is unique to the individual and that individual’s career path, or desired career path.”

In the future, the team plans to refresh the index regularly as the labor market evolves. 

Bradlow said the index aligns directly with the mission of the Wharton AI & Analytics Initiative, emphasizing how the project’s educational component is core to its purpose.

“I’m pretty confident that at one of our meetings coming up, we will be talking about this skills gap report and what it means for us as educators,” Bradlow said. “I think every field should do this type of skills gap assessment.”

---