Creating value with AI upskilling

We’ve heard it once too often: AI is supposed to help us become more productive, but it’s not delivering. Is that true? What are the real issues?

Why are organizations not capturing the gains from their AI investments? How can organizations more effectively reap the benefits of AI that individual staff are already experiencing?

Dhun Davar (DD): Here’s the thing that often gets missed in this conversation: the problem isn’t reluctance. It’s not that organizations don’t believe AI matters or that workers don’t want to learn. In our AI for All report, which surveyed workers and business leaders across Asia-Pacific, we found that 91% of workers are eager to acquire new skills to work with AI, and 96% of business leaders acknowledge generative AI’s significant impact. The will is there on both sides.

So why aren’t the gains showing up? McKinsey has called this the new Solow Paradox, where almost nine out of ten companies have deployed AI in at least one business function, yet 94% of respondents report not seeing “significant” value from those investments.

When you dig into why, the answer isn’t about technology. It’s about what happens after someone learns something. An employee who completes a prompt engineering course returns to a workplace with no AI governance, no process mapping, and no leadership alignment. Their new capability has nowhere to go.

What we consistently observe, particularly with the MSMEs and community organizations we work with through AVPN, is that the gap between individual AI use and organizational benefit comes down to whether there’s any infrastructure for knowledge to travel. When there isn’t, learning stays at the desk. The AI Opportunity Fund: Asia-Pacific is designed around this diagnosis. We are working through local organizations that can build that structure from the inside by intentionally engaging with the leaders and decision makers and delivering training that’s immediately applicable to real work contexts and practical use cases.

How can organizations leverage the multiplier effect to turn one person’s upskilling into benefits for their team members?

DD: Most organizations are chasing the multiplier effect in the wrong direction. The instinct is to train one person well, then cascade by having them run internal workshops, share learnings in a team meeting, maybe write a best-practice guide. That model sounds logical but rarely holds. Knowledge transferred through formal internal cascades loses context, urgency, and relevance at every step.

The more durable multiplier isn’t hierarchical but relational. The AI Opportunity Fund is built on a scaled version of this logic. Rather than training individuals and hoping knowledge travels, AVPN works through local organizations that already have community trust and sector knowledge by equipping them to deliver training to the workers and small business owners around them.

One organization trained becomes a multiplier for hundreds. Phase 3 of the Fund takes this logic upstream to educators: when a teacher gains AI fluency and embeds it into daily lessons, the impact carries forward through every student in that classroom for years.

What else should organizations do to realize and measure the returns on AI investments?

DD: Two things stand out, both of which most organizations are underinvesting in.

The first is workflow redesign. Training builds capability, but capability only converts to returns when the way work gets done actually changes around it. Deloitte’s 2026 State of AI in the Enterprise report confirms that the most common organizational response to AI is more education and reskilling instead of role or workflow redesign. Those two things need to happen together. Managers and business owners need to be part of the conversation because they’re the ones who decide whether newly acquired AI capability gets applied or quietly shelved. AI doesn’t slot neatly into existing processes. It requires rethinking which steps genuinely need human judgment and which don’t. That’s a leadership question as much as a training one.

The second is measurement. This is where most organizations are still flying blind. They track inputs: licenses purchased, workshops delivered, completion rates and tokens used. None of that tells you whether anything changed. The questions worth asking are operational: Are decisions being made faster? Is output quality improving? Are people spending less time on tasks that used to consume them? For social sector organizations, there’s an additional layer: is AI capability actually shifting outcomes for the communities you serve and not just making internal operations more efficient?

This is how the AI Opportunity Fund approaches it. We track outcomes beyond completion rates – whether workers are applying what they learned, whether small business owners are making decisions with AI, whether the awareness gap is closing. Completion is an output. Behavior change is the outcome.