Is It Better to Hire People With AI Skills or Train Your Existing Employees?

Employers trying to adapt to AI have a practical problem that has little to do with choosing the right software.

They need people who know how to use it.

The question is whether those people should be hired from outside the company or developed from within.

There is no universal answer. The right choice depends on the role, the skills already available inside the company, and how quickly the business needs to change. But recent research suggests that employers should not treat AI skills as something that can only be acquired through new hiring.

AI is changing the skills required across many jobs, while training is struggling to keep pace.

The International Labour Organization said in August 2026 that AI adoption is increasing demand for digital, cognitive, and socioemotional skills, as well as AI literacy. The report also points to adaptability and human skills as increasingly important alongside technical knowledge.

That creates a choice for employers: hire people who already have the skills, or give current employees a way to develop them?

There are situations where hiring is the practical option.

A company building an AI product, for example, may need people with specialized technical knowledge that does not exist anywhere else in the organization. The same can apply to businesses that need machine learning engineers, data specialists, or people who can build and maintain AI systems.

The demand for those skills is rising.

SHRM's September 2026 analysis of job postings across 27 countries found that the share of IT and computer science postings mentioning AI skills increased in every country studied between June 2025 and June 2026, although the level of demand varied substantially between markets.

PwC's 2026 Global AI Jobs Barometer, based on more than one billion job advertisements across six continents, found that jobs requiring specific AI skills were growing much faster than the overall job market.

For an employer that needs a specialist immediately, hiring may be the only realistic way to fill the gap.

But that does not mean every employee needs to become an AI specialist.

One of the more useful points from the OECD's 2026 research is that the AI skills conversation is often too focused on highly technical roles.

The OECD found that fewer than 1% of workers need advanced AI skills. For most workers, the more relevant needs are digital skills, the ability to use and interpret data, problem-solving, creativity, and other workplace skills. The organization also found that a lack of skills is a major barrier to AI adoption.

That distinction matters for employers.

A marketing employee may not need to learn how to build an AI model. A customer support employee may not need to become a machine learning engineer. An accountant may not need to understand how a language model is trained.

They may, however, need to know how to use AI tools appropriately, check their output, protect confidential information, and recognize when human judgment is still required.

Those are skills an existing employee can often learn.

There is also a business reason to look at existing employees before opening another job requisition.

Employees already understand the company, its customers, and its internal processes. They do not need to learn the business from scratch.

The Conference Board's July 2026 research found that 55% of workers surveyed regularly use AI, but only 33% had participated in employer-provided AI training during the previous six months. Nearly 28% said their employer provided no AI training at all.

That gap creates an opportunity for employers.

If people are already using AI but the company has not given them structured training, the problem may not be a lack of talent. It may be a lack of support.

The OECD similarly found that many firms are investing in retraining and upskilling, and that workers who receive training are more likely to report positive outcomes from AI adoption.

There is a limit to the "train everyone" approach.

Some skills take years to develop. A company cannot turn a generalist into an experienced AI engineer with a few online courses.

Employers also need to consider the time involved. Training takes employees away from their normal work, and the business still has to pay for the programs, tools, and time required.

There is another issue: some employees may not want to move into substantially different roles.

That means workforce planning should come before deciding between hiring and training.

The better question for an employer is not:

"Who knows AI?"

It is:

"What work is changing, and what skills will we need to do it well?"

For each role, employers can separate skills into three groups:

Skills the company needs immediately.
If the business has a gap that is critical to current operations, hiring may make sense.

Skills existing employees can develop.
If the change involves using AI tools, analyzing information, or improving workflows, training may be enough.

Skills that will become important later.
These are the skills that should be part of longer-term workforce planning rather than an urgent hiring decision.

This approach also helps employers avoid putting "AI" into every job description simply because it is a popular term. The same planning should apply to compensation. Before hiring for a new skill, employers also need to consider whether the salary they are offering matches the current market. Our guide to how much employers should pay for remote talent in 2026 looks at current remote hiring and salary considerations.

Hiring and training do not have to be competing strategies.

A company might hire one person with deep AI expertise while training the rest of the team to use AI in their own roles.

That can create a different kind of workforce: specialists who understand the technology working alongside employees who understand the business.

PwC's September 2026 Global Workforce Hopes and Fears research shows why that balance matters. Across nearly 50,000 workers in 48 countries and regions, 64% reported using AI at work during the previous 12 months. At the same time, only 51% said they had access to the learning and development resources they needed.

The problem, then, may not simply be finding people with AI skills.

It may be helping more of the people a company already employs develop them.

For specialized technical positions, hiring may be necessary.

For many other roles, training existing employees may be the more practical starting point.

The best approach is often to do both: hire when the business needs expertise it does not have, and train employees when the required skills can realistically be developed.

AI is changing the skills employers need, but that does not automatically mean every new skill requires a new employee.

For many companies, the next hiring decision may begin with a different question:

Can we build this skill before we buy it?

Need people with the right skills for your growing team?