Pos Malaysia Delivers 10 AI Projects in 10 Weeks at Zero Cost

When most people think of AI transformation, they imagine expensive consultants and year-long timelines. Pos Malaysia just flipped that script entirely.

The 200-year-old postal institution recently delivered 10 AI proof-of-concept projects in just 10 weeks without spending extra money. Two are already live, four more launch by year's end, and they're already saving six figures.

Group Chief Technology Officer Shahrin Oli Mohamed credits an unusual approach: putting tech teams in the same room with employees who actually do the work every day. No ivory tower planning. No solutions searching for problems.

"Technology works best when there is deep collaboration and joint ownership," Shahrin explained. "The team experiencing the pain point should work closely with technology."

Take their medical claims project. With double-digit millions in annual healthcare costs, they needed to detect payment anomalies. But first, they had to figure out the real problem.

The team met every other day, sometimes so frequently they joked about seeing each other more than their families. They dug through seven possible approaches before finding the one worth pursuing.

That intensive collaboration revealed insights no technology team could have discovered alone. The employees processing claims daily understood the practical challenges and edge cases that would make or break any AI solution.

Why This Inspires

Shahrin brings an interesting mix to the role. She founded a startup in 2016 that failed, worked her way up at Shell from trainee to Global Head of Technology for the world's largest LNG trading business, and spent time at Gartner in Dubai.

Now she's applying startup speed with corporate discipline. The secret isn't just AI or cloud computing. It's reimagining how people work together.

"Never get attached to the idea. Get attached to the outcome," she said. "First invest in gaining clarity on the actual business outcome we're trying to achieve."

Amazon Web Services provided cloud tokens and extra resources, but Pos Malaysia's internal teams drove the work. That partnership model kept costs at zero while building internal capability.

The approach challenges a common assumption: that legacy organizations can't move fast. Pos Malaysia is proving that with the right culture and clear problem statements, even established institutions can match startup velocity.

Four more AI solutions go live before December, each co-created by the people who'll actually use them.