IT leaders rethink talent strategies to cope with AI skills crunch

As head of transformation, artificial intelligence, and delivery at Guardian Life, John Napoli is ramping up his company’s AI initiatives. And a big part of that is scaling up AI talent.

But Napoli, like many CIOs, is facing a tough labor landscape for highly sought skills.

By most accounts, enterprise CIOs are rushing to hire for AI-related roles, putting them into fierce competition with one another — and with big tech companies and CTOs everywhere. And given the pool of those with experience in AI is shallow, many CIOs have recognized the need to build AI talent in-house themselves. Now, they’re racing to train workers fast enough to keep up with business demand.

“When it comes to how companies are getting talent, the word that comes to mind is ‘scrambling’ — they’re scrambling to get the talent they need. And CIOs are taking on the lion’s share of the quarterbacking,” says Saurajit Kanungo, president of the consulting firm CG Infinity and co-author of Demystifying IT: The Language of IT for the CEO.

According to a 2023 survey from Access Partnership and Amazon Web Services (AWS), 92% of employers expect to be using AI-related solutions by 2028 and 93% expect to use generative AI within the upcoming five years. The survey also found that 73% of employers have made hiring talent with AI skills and experience a priority. Yet 75% say they can’t find the talent they need.

Results from Foundry/CIO.com’s 2024 State of the CIO survey enforce this finding, with AI vaulting to the top slot of enterprise CIO’s hardest-to-hire roles.

Complicating matters is the nuance involved, Kanungo says. IT leaders need talent that knows how to enable and manage AI capabilities being built into enterprise applications. They also need workers who know how those capabilities can serve the business.

Moreover, many need deeper AI-related skills, too, such as for building machine learning models to serve niche business requirements. And they need people who can manage the emerging risks and compliance requirements associated with AI.

Here’s how IT leaders are coping.

Staffing strategies emerge

Despite the continuously tight labor market and complexity of the task, Napoli believes he has Guardian Life’s AI talent strategy under control. He has identified the skills required to move his company’s AI agenda forward, where those skills should reside in the organization, and how he will get them.

The approach will involve targeting a mix of existing and emerging skills, he says.

For example, Napoli needs conventional data wrangling, data engineering, and data governance skills, as well as IT pros versed in newer tools and techniques such as vector databases, large language models (LLMs), and prompt engineering.

He wants data scientists who can build, train, and validate models for use cases, and who can perform exploratory analysis and hypothesis testing.

He also needs talent to oversee the company’s standards around explainability and to root out unintended biases, both emerging competencies.

And he wants business-minded technologists experienced in AI-enabling technologies such as Snowflake, Databricks, Amazon Bedrock, and Amazon Sagemaker who put business benefit first.

“All existing jobs are basically changing to make sure we have people who are AI-capable, and they’re changing throughout the organization,” Napoli says. “We need to transition jobs to be ready to leverage AI tools.”

Balancing build vs. buy

To achieve his vision, Napoli is applying a tried-and-true IT talent strategy of combining building and buying tactics to create the AI-ready workforce he wants.

“There is a lot of reskilling and upskilling, and there are new hires because you want someone who has done it before,” he says, adding that new hires will be targeted for the company’s core AI delivery team — partly to bring workers experienced in AI and partly to get the skills quickly.

“You have to look for people who have really [worked with AI capabilities] in a way that makes a difference,” he says.

To supplement new hires, Napoli is training existing employees for this team, combining on-the-job learning, online programs, and certification courses, such as the AWS Certified Machine Learning — Specialty credential.

Napoli points to another critical piece of his AI talent strategy: retention, which he supports through robust compensation, engaging work, and a culture that emphasizes employee recognition.

Orla Daly, CIO at Skillsoft, maker of an online learning platform, is taking a similar approach to building AI talent. Like Napoli, Daly needs people with the technical skills to enable LLMs and other AI tools, as well as talent that can identify where AI can be useful, determine the right AI tools, and then implement those tools into workflows.

All workers are impacted by those needs, she says.

“Now every job description is changing because of AI, and so AI skills need to be reflected in every role,” she says. “It’s less about the machine learning skill set and more about how you adapt all your roles to take advantage of AI.”

Daly’s strategy is to hire for skills for work that needs to happen fast, although she admits she and other CIOs struggle to do so because of the high demand.

“If you’re trying to get going fast, then hiring can benefit you, but there isn’t the availability of skills on the market, and there’s also the issue of affordability,” she says, noting that data and cloud skills — both needed for AI projects — are in short supply, too.

Consequently, Daly is focusing increasingly on upskilling her company’s existing staff — in both IT and in other departments — so they’re able to handle AI-related work, from implementing the AI tools available from vendors to ensuring privacy and regulatory compliance.

“Everyone is learning,” Daly says.

Skillsoft has taken a multiprong approach to learning, starting with education and training to build AI awareness broadly among all workers and offering more in-depth training to those who need it. Case in point: Training data workers on AI bias.

Lunch-and-learns, mentorships, and cross-collaboration initiatives that pair IT and business workers together help further spread AI-related knowledge, Daly says.

“And for the rest of the organization, we’re trying to get them to lean in and to not be afraid to experiment with AI. A lot of this is about building the muscle. And you do that by trying it and seeing where you can go,” she adds.

Overcoming the talent crunch

Despite such strategies, the AI talent crunch seems to be slowing enterprise adoption of the tech.

The 2023 Digital Leadership Report from Nash Squared found that only 10% of the 2,000-plus technology leaders it surveyed said their organizations have large-scale implementations of AI. Meanwhile, 54% of respondents said skills shortages hamper change.

Additionally, the survey found that data analytics and data engineering are the most scarce skills today, with AI skills at No. 5.

“The top AI talent are working with the Magnificent Seven,” says Michael Goldberg, vice president of strategic partnerships for Harvey Nash USA & Canada, referring to the seven tech giants.

Only other top companies and innovative tech firms can hire that AI talent away, either because they have the brand-name clout, the big AI projects, and/or the generous compensation packages to compete with AWS, Google, and other tech giants for candidates with hard-core AI skills.

As a result, CIOs at most companies have a tougher time attracting machine learning engineers, prompt engineers, and other AI-specific talent, Goldberg says. That leaves many turning to AI consultants and training their existing data engineers, enterprise architects, and others so they can slide into those AI positions.

“They’re assessing workers within their own organizations and looking at those on their own teams who are high performers who could take on this work,” he adds.

That’s happening even at the top: Professional services firm Genpact in February announced that it expanded Vidya Rao’s role from CIO to chief technology and transformation officer to reimagine “internal processes, tools, technologies, and infrastructure with an AI-first approach.” Additionally, Rao was tasked with fortifying Genpact’s data capabilities in part to guide its AI initiatives.

As part of that work, Rao is creating an AI center of excellence for which she says, “We now have to hire, find, and upskill.”

“We need trilingual skills — data, domain, and core AI skills — and you can’t find many of these people,” she says.

Rao needs, among other roles, data scientists, solution architects, project managers with AI acumen, as well as AI developers — some of whom she has with more to be built or hired.

“They all need to know how to play with the data, read the data. Those skills existed in pockets before but demand for that [capability] is prominent now,” she adds.

Training, though, is getting the team to where they need to be.

The company has an Introduction to Generative AI course on its internal learning platform, and it registered 30 of its leaders for a spring 2024 online course on AI and business strategy offered by the MIT Sloan School of Management.