The upside of AI upskilling
Companies can throw money at AI. But can they make it stick without spending on training?
Companies are pouring gobs of money into artificial intelligence—more than $1 trillion globally in 2026 alone, according to Goldman Sachs—but almost none of it goes toward training their workers in how to use it. According to a 2026 Deloitte report, most (93%) of enterprise AI budgets go to technology and tooling, with just 7% for things like governance, training, and job redesign—leaving just a sliver of a sliver for AI upskilling.
They’re missing out. Treated as a business objective, AI upskilling doesn’t just deliver higher productivity, it gives companies a distinct competitive advantage. It keeps workers around. Nearly 3 in 4 employees say AI training opportunities motivate them to stay at their company longer. Even better, organizations that provide workers with role-specific AI training are four times as likely to outperform their competitors financially, according to a 2025 McKinsey report.
“Most of the AI debate at the CEO level is about which model, which stack, how much budget,” noted Freshworks CEO Dennis Woodside in a recent LinkedIn post. “That debate matters. What determines whether AI shows up in the P&L is whether the people you already employ know how to use what you've already bought.”
But most upskilling is fairly generic, one-size-fits-all AI literacy. Only 38% of employees say their current AI training is specific to their job role, while 73% say role-specific training is exactly what would help them use AI more effectively, according to the TalentLMS 2026 L&D Report.
That “spray and pray” approach to AI skilling is exactly why most companies aren’t seeing tangible business outcomes from their initiatives, says Susan Youngblood, a former HR leader at Fortune 100 companies and operating advisor with Bessemer Venture Partners.
“Where the real value in AI comes is when you’re enabling employees to use it to do something they could never do before,” she says. “That’s where you get the bigger returns on investment.”
What determines whether AI shows up in the P&L is whether the people you already employ know how to use what you've already bought.
Dennis Woodside
CEO and President, Freshworks
The upside of upskilling: Innovation and talent retention
The companies using AI skilling to pull ahead proceed from the same starting point: It’s the work, not the tool. “Everybody has access to the same tools,” Youngblood says. “The smart companies are looking at where can AI actually add value to an existing workflow, and what’s going to be the end result that changes as a result of looking at that workflow.”
Orange Business—the enterprise arm of French telecom giant Orange, with more than 18,000 employees across 26 countries—has focused its two-year upskilling push on day-to-day AI use in employees’ workflows and on redesigning jobs to bake in AI. Role-specific training and certification are multitiered and targeted: Sales teams learn how to surface AI business insights for precision pitches; support reps adopt internal co-pilots to resolve more customer service cases faster; and HR managers gain proficiency with purpose-built AI tools to better navigate internal policies.
Orange says that its AI upskilling has created a workforce capable of new levels of innovation. At its 2026 Orange Business Summit, the company brought to market a set of new agentic AI products that its workforce leveraged their AI savvy to build.
How McLaren, New Balance & more turn service performance into a competitive edge
At Freshworks, reskilling in AI tools has also led to tangible innovation. At its recent company-wide “AI in Motion” event, held across North America and India, employees from teams including SecOps, IT, CX, sales and product produced 25 working prototypes to tackle real customer and business problems. Among the most promising prototypes came from employees in business functions that don’t write code for a living but use AI tools in their daily workflows.
“It’s not complicated,” says Dr. Michael Edmondson, founding director of the Learning and Development Initiative at NJIT. “If you want to keep your people, educate them,” he says. “People want to learn. If you’re a leader, you have an obligation to support that.”
Read also: The AI talent paradox
Where to start when the tech is always changing
The hard job of AI skilling is knowing which skills to train for when the technology is evolving so fast. “If you have a limited training budget, you want to fund the thing that compounds,” Youngblood says. “Human-in-the-loop judgment, problem framing, ethics—these things compound and they’re lifelong.”
As a business objective, upskilling can’t be a one-and-done line item. The tools, the use cases, and the skills are continually evolving. Companies that want to get real returns will budget training as a standing capability and empower their workers to evolve along with the technology.
