The retired ex-US-intel officer who started an AI business in Southeast Asia

Photo credit: Cyril Ng and SGInnovate

Entrepreneurs are often eager to talk about their startups. Drew Perez, founder of artificial intelligence (AI) startup Adatos, seems like an exception to the rule. As I asked about Adatos, he seemed to steer clear of discussing it.

“Talking about the history of the company – that’s so myopic, right?” he says, preferring to discuss the big picture.

But it’s hard to ignore him and his company. Started three years ago, Adatos is one of the rare AI startups based in Singapore that focuses on agriculture. As its sole founder, Perez served over twenty years as an intelligence officer in the United States Department of Defense (DoD) before retiring.

His work in the DoD involved espionage. He also conducted training – where he taught students from the top schools in the US to become intelligence analysts. Naturally, he deflected when asked to elaborate on his past work.

Starting an AI company isn’t a typical retirement plan for most, but Perez isn’t typical. His goal is to find the right use case for artificial intelligence within Southeast Asia.

To do that, his team has automated the process of creating machine learning models – essentially an AI that builds its own AI to solve the problems specified by a human.

The technology essentially takes on the role of a data analyst, expediting the generating and testing of hypotheses. Call it an analyst-in-a-box.

Google didn’t invent this

If you think you’ve heard of a self-building AI before, you’re right. Google’s AutoML program and the H2O Driverless AI have recently popularized the concept.

But the technology predates these private sector efforts, claims Perez. He says that he has used versions of this over a decade ago when the US Intelligence Community funded programs to deal with the lack of qualified intelligence analysts in the military by creating AI that can mimic how humans solve problems.

The US government “has been doing over two decades what the private sector has only discovered in the past six months,” Perez claims.

This analyst-in-a-box, which like a lot of AI tools is built on top of open-source tech, has allowed Adatos to offset the exorbitant costs of hiring data scientists and analysts.

With their technology, they have been able to accelerate their R&D. Within three years, Adatos has tested 120 use cases for AI across 15 industries. It knows which applications can be profitable or even legal, allowing the team to rule out AI solutions in retail and medtech, for example.

On its client list are some of the biggest food producers in Southeast Asia.

Making something that lasts

He doesn’t need to fundraise again