Yesterday I attended DPEX 2026, and Kevin Shepherdson from Straits Interactive delivered a talk that cut straight through the noise.
While everyone else is still talking about AI literacy and fluency, Kevin made a simple but critical argument: stop learning about AI, and start building AI capability.
It’s a distinction that matters more than most people realize.
Prof Jay Gonzalez III from Golden Gate University followed with observations from Silicon Valley that validated exactly what Kevin was saying and added a sharp edge about what’s actually happening in the market.
The Problem with “AI Literacy”
Here’s what’s happening right now across organizations: 80% of people are stuck in what Kevin calls the “pilot trap”. They’re running experiments, testing tools, maybe even getting excited about ChatGPT demos but nothing scales. Nothing repeats. Nothing becomes part of how work actually gets done.
The issue isn’t awareness. Everyone knows AI exists. The problem is lack of execution capability.
Kevin framed it with a concept called the AI Capability Ladder, which breaks down into three levels:
AI awareness: You know AI exists
AI fluency: You know how to use some tools
AI capability: You can build, govern, and repeat AI-driven processes that deliver measurable value
Most training stops at fluency. But fluency doesn’t differentiate you. Capability does.
What Silicon Valley Is Actually Hiring For
Prof Jay’s perspective from the heart of Silicon Valley reinforced this shift with concrete evidence.
Golden Gate University sits at 536 Mission Street, surrounded by Salesforce, EY, Deloitte, OpenAI, and over 200 AI startups. Just months ago, Nvidia moved in and bought four buildings. This is the epicenter of AI transformation and the hiring patterns tell a clear story.
Two to five years ago, companies were chasing business analytics graduates: people who could do “content storytelling” and combine data with narrative.
Now? Business analytics programs are struggling to fill seats.
Who are they hiring instead?
People with AI backgrounds combined with business knowledge, what Minister Josephine Teo calls “AI bilingualists”. But not computer science or engineering grads. Companies want people who understand:
The business side: marketing, accounting, operations, communication
The AI side: governance, prompt engineering, process integration
How to bring AI capabilities to market
As Prof Jay put it: companies like Salesforce, Databricks, and the 200 AI startups in their ecosystem aren’t looking for pure coders. They’re looking for translators and orchestrators. People who can bridge domain expertise with AI execution.
And here’s the kicker: even Stanford and Berkeley haven’t figured out how to build hybrid programs that cross computer science and business schools. The egos won’t let it happen. But the market is demanding it anyway.
The AI Factory Framework
Kevin introduced something called the AI Factory, a structured approach to building repeatable AI capability rather than one-off experiments. The framework focuses on three critical layers:
People and culture: Getting teams aligned on AI-driven workflows
Governance and compliance: Ensuring AI usage is safe, ethical, and auditable
Technology and integration: Building systems that actually work in production environments
This isn’t about buying more tools. It’s about creating organizational muscle, the kind that turns AI from a demo into a competitive advantage.
From Bilingualism to Capability
Kevin referenced Minister Josephine Teo’s term AI bilingualism, the ability to speak both business and AI. But he pushed it further: bilingualism is just the start. The real goal is capability, being able to translate domain expertise into functioning AI systems that scale.
This resonates deeply with my own work. Over the past year, I’ve been building RAG systems, custom AI agents, and domain-specific workflows for clients. And what I’ve learned is this: tools come and go, but capability stays.
Knowing how to prompt ChatGPT is useful. But knowing how to architect a RAG pipeline that handles critical documentation? That’s capability. And that’s what clients actually need.
Silicon Valley Is Competitive and Unforgiving
Prof Jay called Silicon Valley “a valley of life, but also a valley of death”. In that ecosystem, you either build competitive advantage through capability, or you disappear.
The companies that win aren’t the ones with the best tools. Everyone has access to the same models. The winners are the ones who can orchestrate humans and machines, who own their processes, their content, and their judgment.
That’s not a Silicon Valley phenomenon. That’s the future everywhere.
What This Means for You
If you’re in operations, consulting, or any role where workflows matter, the question isn’t “Should we use AI?” It’s “What capability can we build that actually works in real life?”
This means moving beyond:
Random ChatGPT experiments
Proof-of-concept demos that never leave the sandbox
Tool-hopping without clear outcomes
And moving toward:
Repeatable workflows that solve specific problems
Governed systems that meet compliance and quality standards
Measurable value that justifies investment
Both Kevin and Prof Jay reinforced something I’ve been arriving at independently: the next advantage isn’t knowing AI exists. It’s building capability that actually works.
A Question for You
If you were at DPEX or thinking about AI transformation in your organization:
What capability could you build this year that moves beyond pilot and into repeatable value?
Not what tool could you try. Not what demo could you run. But what capability, what repeatable, governed, valuable system, could you actually put into production?
That’s the shift Silicon Valley is making.
That’s what Singapore organizations need to hear.
And that’s where the real opportunity is.
