AI Is Helping Us Break Eggs Faster. But Are We Learning Anything?

AI is not just another technology rollout. It is an organisational-design moment.

By Raatha Ganesh - Black Dog Consultants

You have to break a few eggs to make an omelette.

That’s what we say when we want permission to experiment.
Try something. Make a mess. Learn from it. Make something better next time.
Fair enough. But AI changes the kitchen.

Across Asia-Pacific, AI is helping organisations break far more eggs, far faster.
Teams are generating more analysis, proposals, content and decisions in less time than ever before.

Which sounds useful. Until you ask the awkward question:

Is AI helping us learn faster or simply helping us make mistakes at speed?

AI makes it easier to produce at speed but harder to ensure learning keeps pace. That matters because speed alone is not capability.

We have been here before

This is not the first time technology has forced organisations to rethink how work gets done. It may just be the first time technology has landed in almost everyone’s work at once.
We have seen this pattern before.

Enterprise systems changed who owned data and decisions. Digital channels changed what customers expected and how teams had to work together. Cloud and collaboration tools changed where work could happen, and how leaders had to lead when people were no longer in the same room.

The organisations that got real value did not just install the shiny new thing.

They changed the system around it.

The technology is new. The leadership task is not.

AI belongs in that lineage. It is not simply a productivity tool. It is a prompt to revisit who does what, how decisions are made, where expertise sits, what gets escalated and which work should exist at all.

The difference is speed. AI is entering everyday knowledge work much faster than earlier technologies did. That makes the gap between producing and learning easier to create and harder to ignore.

APAC is moving fast

Across APAC, the story is no longer about experimentation alone. Aon’s 2026 Human Capital Trends Study found that 74% of organisations have already deployed or are piloting AI programmes.

But deployment is not the same as capability. Only 21% of APAC organisations believe they can recruit and retain enough AI talent to support the shift. The same research found that 84% expect AI to automate some tasks without replacing jobs entirely, while 87% expect it to create new roles requiring different skills.

This is not just a skills gap. It is a work-design problem, an organisational-readiness challenge. If the work is changing, but the roles, decision rights, career paths, governance and learning systems stay the same, the organisation is not really transforming. It is bolting AI into an operating model built for another era.

APAC is not short of AI momentum. It is short of the capability to convert momentum into better work.

The productivity trap

Singapore gives us the awkward bit in numbers. AI is helping firms move faster. But faster is not the same as smarter. Among firms already using AI, 70.7% report improved worker productivity. Lovely. The machine is humming.

But then it gets quieter.

Only 13.3% report better decision-making and just 11.9% report innovation gains.

So yes, AI is helping people do more.

But those numbers should give leaders pause.

The bigger question is whether it is helping them think better.

It is much less clear that AI is helping organisations make better decisions, learn from results or redesign how work gets done.

Productivity is what happens when AI speeds up a task. Capability is what happens when an organisation learns to redesign the task.

If the time saved becomes more emails, more documents, more meetings and more decisions, AI has increased output.

But it has not necessarily improved the work.

Singapore’s firm-level data shows that only 3.8% of firms have integrated AI into their core processes. This means many organisations are still using AI at the edges.

They are using it to draft, summarise and speed up admin.

That is useful. But it is not transformation.

If AI is layered onto yesterday’s operating model, it will accelerate yesterday’s problems.

A different leadership question

The challenge is not to slow AI down. It is to make sure learning keeps up.
Because right now, a lot of organisations are asking the easy question:

How much AI are people using?

But that only tells you whether the tool is spreading. It does not tell you whether the work is getting better.

Leaders need to ask better questions:

• What is AI helping us learn about our customers, decisions and workflows?
• Where are we seeing rework, recurring errors or weak judgement?
• What have we changed because AI showed us a better way?
• Are people equipped to challenge AI output before it becomes a decision?
• Where has AI made us faster at something we should probably stop doing altogether?

AI does not automatically make an organisation smarter.

It can simply put a rocket under the habits already there.

So the winners will not be the organisations with the most AI tools, pilots or productivity charts.

They will be the ones asking harder questions:

• What work needs to change?
• What decisions need to improve?
• What are we learning that we could not see before?

Because adding AI is easy.

Learning from it is the real work.

Coming next: The AI-bilingual organisation

In Part 2, we look at the AI-bilingual organisation: one that speaks both AI and business.

Not fluently enough to sound clever in meetings.

Fluently enough to make better decisions.

Because the future will not belong to the organisations with the most prompts. It will belong to the ones with the best translators.