AI upskilling for large organizations

Adit Jain at Leena AI learned the hard way that voluntary training doesn't stick. When he softly encouraged 30 employees to take a reskilling course, only two did. When he switched to a strict 90-day deadline with an exam, more than 85% completed it.

His experience, along with examples from DAZN, IKEA and Coursera, shows what AI upskilling for large organizations looks like when it actually works.

Why soft pushes for AI training fail

Leena AI builds AI colleagues for the G&A function and the back office, automating repetitive IT, HR, finance and business processes in large enterprises. Jain said automating back-office processes displaces 40 to 60 percent of the people doing that work.

In 2024, when the company moved to an entirely agentic large language model architecture, an entire team became irrelevant. Jain offered them a three-month course to become solution consultants for higher pay.

Now, what I did, and it was my mistake, is we were softly pushing those team members towards, "Hey, there's this course we have created. Go become a solution consultant," which is now called an FDE, but, "Hey, go become a solution consultant, and it'll be great. And you do this in three months, and it's a higher pay."

Only two of the 30 took the course, even though their roles faced immediate automation. The cost of that gap adds up: Sarah Franklin at Lattice cited a Gallup study showing about $435 billion in corporate productivity loss because people don't know what to do with AI.

Enforcing adoption through deadlines

Jain tried again with seven or eight people assigned to an AI knowledge cleanup project for customers. This time the course and certification were mandatory.

But we said, "You have 90 days. There's going to be an exam. If you fail, you're out. If you pass, you're going to be in, and you get a pay bump."

More than 85 percent completed the program and got the raise.

Connecting AI skills to daily goals

Rasmus Holst at ZENSAI described how DAZN connected strategy data with HR data and individual goals. Once employees saw the actual strategy against their own goals, they started taking responsibility for their own training.

They saw 5X take-up in training that were voluntarily taken.

Holst said workers took the training to improve and help the business, rather than for compliance.

Scaling AI literacy across a massive workforce

Parag Parekh, CDO at IKEA, admitted the company missed earlier technology shifts.

The fact that we were very good at retailing held us back, and we missed the trends and the changes in the customer journey when it came to the rise of the internet, but more importantly, the smartphones.

Parekh said there is a lot of anxiety about AI in the workforce, and his goal is to turn that anxiety into curiosity. IKEA applied AI to its Billy chatbot, raising the containment rate for customer queries from 47 percent to 74 percent. The customer service coworkers it freed up were reskilled to help customers design their homes.

Measuring the ROI of AI upskilling for large organizations

Greg Hart at Coursera described a large global technology company that used the platform across its employee base and retained engaged employees at a 50 percent greater rate than anticipated. Hart also reported that 91 percent of employers say hiring someone with a micro-credential reduces their training costs by 10 to 30 percent.

Andrew Ng at DeepLearning.AI pushed back on the idea that AI makes learning to code unnecessary. He recommended that everyone learn to code, since AI makes it easier to build custom software and makes the people who do it more effective. On the advice to skip coding, he said:

We'll look back on that as some of the worst career advice ever given.

Jain's results are the clearest lesson for AI upskilling for large organizations: the voluntary push got 2 of 30 people through, and the mandatory deadline got more than 85 percent.

To hear more from the people putting AI into production, see the HumanX 2027 program preview.