Colombo — Before an organization can govern AI, it has to understand it. That conviction brought Sen & Co to the Sri Lanka Customs IT Directorate, where founder Sahan Perera led its senior officials through the foundations of modern AI — from first concepts to the inner mechanics of a large language model.
The workshop set the buzzwords aside and went to the machinery underneath: what these systems are actually made of, how they learn, why they answer the way they do, and where their limits sit. Step by step, the room followed a sentence through the layers of a language model — and saw for the first time what happens between a question going in and an answer coming out.
The gap nobody talks about
For many in the room, it was an eye-opener. These are the officials who run the IT backbone of a national customs service — and like their counterparts in organizations everywhere, most had never been shown how modern AI actually works.
That gap is no accident, and it is not unique to government. Tools like ChatGPT are free, capable and one browser tab away, so people simply start using them — long before anyone explains what happens to the data they paste in, or which tasks these tools can be trusted with. For a critical government organization, that combination of easy access and low understanding is a real vulnerability. Knowing how AI works is no longer a specialist's luxury; it is part of operating safely.
“Everyone is using AI. Very few have been shown how it actually works — and in a government organization, that gap is a risk in itself.”
Sahan Perera · Founder, Sen & Co
From one workshop to national policy
The session was warmly received — and it did not end in the boardroom. Following the workshop, Sahan Perera was invited to advise the presidential committee appointed by His Excellency the President of Sri Lanka on AI-driven use cases for customs operations.
This is how Sen & Co works: build the systems, and build the understanding that lets organizations run them with confidence. The impact starts with a room full of people finally seeing what is inside the box.