IT leaders: What’s the gameplan as tech badly outpaces talent?

Dun and Bradstreet has been using AI and ML for years, and that includes gen AI, says Michael Manos, the company’s CTO. It’s a quickly-evolving field, he says, and the demand for professionals with experience in this space is exceedingly high. He’s seeing the need for professionals who can not only navigate the technology itself, but also manage increasing complexities around its surrounding architectures, data sets, infrastructure, applications, and overall security.

“Professionals with real experience in this space are rare and can command significant compensation expectations or pursue roles of their choice,” he says. Dun & Bradstreet can attract that kind of talent, he says, because candidates are looking for roles where they can expand their own reach and scale. “We’ve been innovating with AI, ML, and LLMs for years,” he says.

But not every company can say the same. According to a survey conducted by FTI Consulting on behalf of UST, a digital transformation consultancy, 99% of senior IT decision makers say their companies are deploying AI, with more than half using and integrating it throughout their organizations, and 93% say that AI will be essential to success in the next five years. But 76% of respondents say there’s a severe shortage of personnel skilled in AI at their organization, according to the August report.

Other surveys found a similar gap. In a November report by HR consultancy Randstad, based on a survey of 12,000 people and 3 million job profiles, demand for AI skills has increased five-fold between 2023 and 2024. Gen AI-related job listings were particularly common in roles such as data scientists and data engineers, and in software development. And a Red Hat survey of IT managers in several European countries and the UAE found that 71% reported a shortage of AI skills, making it the most significant skill gap today, ahead of cybersecurity, cloud, and Agile.

The talent shortage is particularly acute in two key areas, says Arun Chandrasekaran at Gartner. “There’s clearly a demand for more professionals,” he says. “Like someone who monitors and manages these models in production, there’s not a lot of AI engineers out there, but a mismatch between supply and demand.”

The second area is responsible AI. “How do you build privacy, safety, security, and interoperability into the AI world?” he says. “The demand is high, even as regulations evolve in that space and the supply of professionals is limited.”

To help address the problem, he says, companies are doing a lot of outsourcing, depending on vendors and their client engagement engineers, or sending their own people to training programs.

According to October data from Robert Half, AI is the most highly-sought-after skill by tech and IT teams for projects ranging from customer chatbots to predictive maintenance systems. And paying a premium isn’t out of the question. About 44% of managers are willing to increase salaries for people with AI and ML skills.  These are also areas where organizations are most willing to use contract talent. And to further address the skills gap, 58% of organizations upskill employees, 47% use contract talent, 47% have implemented mentoring programs, 47% pay for professional certifications, and 46% use managed services.

Hiring AI talent

US IT staffing firm Motion Recruitment found that AI engineers saw over a 12% growth in salaries compared to a year ago, and mid-level workers saw a 20% increase. Meanwhile, average starting salaries for tech jobs rose by just 1%. In their report, out of the current list of highest-paid skills in tech, half are related to AI and ML, with gen AI taking the top spot.

“Even outside AI-specific roles, those who boast a skill set that includes these new programs can see large salary increases, with some seeing a 50% boost in compensation compared to those who lack AI skills,” the report says.

Nick Kramer, leader of applied solutions at SSA & Company, suggests expanding gen AI talent searches beyond traditional recruitment channels by tapping into academic networks, attending specialized industry conferences, and engaging with AI community meetups.”

He also suggests looking for senior data scientists who are creative enough to adapt to new ideas and trends, or those more junior but with strong backgrounds in Python, ML frameworks like TensorFlow and PyTorch, and deep learning architectures.

Non-tech companies in particular have the hardest time finding talent with gen AI skills, says Brett Barton, global AI practice leader at Unisys, and the fast pace of change doesn’t help.

“Their pay scales often don’t afford top technical talent and often don’t have technical leadership that can accurately and adequately vet the people they’re interviewing,” he says. “They can certainly educate internally, but the technology is evolving so rapidly that by the time you finish a grad school course or program, the technology is different.”

It’s no surprise that, according to Unisys’ latest AI business impact report, 75% of executives surveyed say they don’t have enough people with the skills to use AI effectively. And with all the competition for AI talent, some companies are taking a different approach to recruiting. Dave Merkel, CEO at cybersecurity company Expel, says some people are always looking for innovative ways to break things, so the solution is to find employees who are always interested in learning new things.

“We ask, ‘When did you last learn a new thing? Tell us a story,’” he says. “The speed of the cyber technology revolution is very fast and attackers are always changing behaviors. So we’ve aggressively tried to hire for traits around curiosity, learning, knowledge-seeking, and initiative. Those traits, regardless of where you sit in the business, lend themselves well to the adoption of new technology, whatever it is.”

And the challenge isn’t just about finding people with technical skills, says Bharath Thota, partner at Kearney’s Digital & Analytics Practice.

“The most challenging aspect is identifying candidates who possess not only the technical skills but also the right mindset,” he says. “We need individuals who can apply gen AI with an industry sector specific and functional perspective, focusing on solving business problems rather than merely adopting a tool or technology-first approach.” And that requires a deep understanding of the business context, he says, and the ability to integrate gen AI into practical solutions. But the talent shortage is likely to get worse before it gets better. “Companies are offering substantial compensation packages to attract top talent,” Thota adds.

Training and development

Many companies are growing their own AI talent pools by having employees learn on their own, as they build new projects, or from their peers. In the Randstad survey, for example, 35% of people have been offered AI training up from just 13% in last year’s survey. Tim Beerman, CTO at IT services company Ensono, takes a RAG approach to gen AI — and there’s not a lot of people out there with years of RAG database experience, he says. “There are data scientists, but they’re expensive,” he says.

One option is to find employees competent in the general area and interested in learning gen AI, and get them trained or have them learn on the job. “Or bring in a consulting company that can help you build and train at the same time,” he adds. “We have a lot of passionate people so this is something new to learn. There’s no shortage of people wanting to learn.”

Ensono uses gen AI to generate everything from marketing materials, thought leadership pieces, ticket analysis, and summaries, to helping sales staff understand products and services and software development. “We’re continually finding ways to leverage it,” Beerman says.

To further get staff up to speed, he brought in about a dozen AI experts from the consulting side of the firm, in addition to another dozen outside consultants who helped build the gen AI functionality for the company while training employees at the same time.

Torc, a technology talent marketplace, took a similar approach to developing gen AI talent. Instead of hiring AI experts from the outside, it looked for existing software engineering staff who were interested in learning the new technology.

“It was more important to us that the person was familiar with what we were doing,” says CEO Michael Morris. “It’d be more work to train someone on the Torc platform. Our VP of engineering said, ‘These guys are interested in doing it, they’re already playing around with it, and had already built some stuff with it.’”

Staffers learn by trial and error, he says. They also go to AI events, like the recent AWS re:Invent conference.

“The landscape is changing rapidly but so is the pace of deployment,” he says. “If 10 years ago it took six weeks to do a prototype and three years ago it took two weeks, now it takes two days. So we do a lot of prototyping. If it doesn’t work, we throw it away.”

Torc has an advantage in looking at its own staff and upskilling for gen AI because it’s a technology-forward company. So is Indicium, a global data services company.

“We’re building a department of AI engineering, mostly by bringing in people from data engineering and training them to work with gen AI and AI in general,” says Daniel Avancini, Indicium’s CDO. “We currently have about 10 AI engineers and next year, it’ll be around 30.”

The company has open positions posted, but it’s not easy to find people, he says. “Data engineering and data science are also difficult to hire for, but gen AI is even worse,” he says. “There hasn’t been time for people to get experience.”

The first round of training was mostly trial and error, he adds, as well as external courses and a lot of reading. Now the company is building its own internal program to train AI engineers. “We already have a pretty big data engineering and data science practice, and we’ve been working with machine learning for a while, so it’s not completely new to us,” he says.

Another company building its own gen AI training is Booz Allen Hamilton, where 14,000 employees have already started upskilling. The goal there is to have the entire 35,000-person workforce be AI ready, with specialized training for AI engineers, AI consultants, and other specialists.

Like other organizations, Booz Allen Hamilton is also trying to hire people with gen AI skills, says Jim Hemgen, the company’s director of talent development. “But there just aren’t enough people. Because of the type of work we do, though, we might not only require that you have those skills but also the proper clearances to do work with government channels.”

Then there’s the pace of change problem, he adds. “The half-life of skills is ever decreasing, and that’s particularly the case in this area.” The solution is to focus on the culture of AI adoption and continuous learning.

“We want people to embrace AI, to have a growth mindset with healthy competition between teams to foster momentum,” he says. “We’re seeing employees’ careers evolve, and change, and be exposed to meaningful work, so they can see they’re making a significant difference.”

However, according to the UST survey, 31% of large companies can’t upskill their own workforce because they don’t have enough training capability. Other surveys support this, too, like Wakefield Research, which surveyed 1,200 executives and IT professionals on behalf of Pluralsight and found that 94% of IT professionals say AI initiatives will fail without staff who can effectively use AI tools, but only 12% say they have significant experience working with AI.

“Although executives and IT professionals agree in investing in talent and training in this area, only 40% of organizations have formal structured training and instruction for AI,” says Chris Herbert, chief content officer at Pluralsight, a tech skill training firm. “The first step in building AI skills within an organization is to look at the current AI skill level among employees to better understand where talent disparities lie.” According to the survey, 90% of executives don’t completely understand their team’s AI skill and proficiency. Another approach to upskilling employees is to learn from outside consultants brought in for AI projects.

Back to school

Some organizations are turning to outside educational institutions for help in gen AI training. For example, Boston Scientific and Blue Cross Blue Shield of Minnesota have turned to the University of St. Thomas, based in St. Paul, Minnesota, for gen AI training.

“Reskilling your existing employees is great ROI,” says Manjeet Rege, professor and chair of the Department of Software Engineering and Data Science at the University of St. Thomas. The university offers both graduate certificates and a full master’s program in AI, designed for working professionals.

The critical difference today is that AI isn’t just a separate thing anymore, limited to data scientists, he says. Gen AI in particular is rapidly being integrated into all types of software applications.

“If you’re hiring a software engineer today, that person has to have some knowledge of AI,” he says. “If you’re hiring an AI engineer, that person needs to have knowledge of software development.”

Offering this kind of training to employees could be a recruiting advantage for companies. According to Korn Ferry’s global workforce survey of 10,000 professionals, released in late October, development opportunities were the fourth most important factor in accepting a new job offer, after flexible working hours, generous compensation, and job security. However, only 32% of companies say they plan to focus on upskilling current employees to address skill gaps.

To help address the skills shortage, the public sector is also stepping up. For example, the District of Columbia has already invested $1.2 million in AI training programs for DC residents, including data science, Python, and other areas. And students don’t pay for these classes.

“We pay people to go into these trainings, and then connect them with an internship or apprenticeship where for six months we pay their wages, and they get their experience,” says Unique Morris-Hughes, director of the Department of Employment Services for the District of Columbia. “When they complete the program, they’re ready to go right into the field.”

The accelerating pace of change

What makes gen AI different from other major tech revolutions is that the AI itself can be used to help meet the challenges it creates.

“Unlike any other technology, you can talk to it like a person,” says Adam Paulisick, professor at Carnegie Mellon University’s Tepper School of Business. “This single shift has led to more receptivity, adoption, and faster training than any other technology we have.”

And the changes are cumulative and unpredictable, he adds. “In three to five years, people might just have an agent, and websites become unnecessary,” he says. “Those kinds of nonlinear changes are hard to understand right now.”