Many Southeast Asian organisations have adopted artificial intelligence (AI) but have seen limited value, focusing mainly on visible tools such as chatbots and image generation.

These early deployments are accessible starting points, but they rarely shift business economics or reflect the region's market realities and strategic constraints.

Shaped by low labour costs, fragmented markets, and a heavy reliance on trade, the region’s AI landscape challenges productivity-led strategies and demands new models of growth.

“Southeast Asia has requirements that are slightly different,” said Mohan Jayaraman, Senior Partner at Bain & Company.

“Labour costs are much lower than global levels—around 7% of U.S. levels—so productivity gains alone cannot add to the P&L value that you get from AI,” he explains.

Leveraging its experiences partnering with regional players, its early involvement with OpenAI, and experiences with hyperscalers like Microsoft, Google, and AWS, Bain & Company has formulated ‘The Southeast Asia CEO’s Guide to AI Transformation’, which distils six lessons to help leaders in the region turn AI potential into measurable outcomes.

“What we wanted to do was convert this into something that is easily usable… a fixed, simple set of messages that leaders could use fairly easily,” said Mohan.

Many businesses still treat AI adoption as a technology deployment rather than a business transformation, and this narrow approach often misses the sustainability and economic impact that leaders expect.

Rather than squeezing out the last bit of productivity gain, Mohan said leaders need to consider a more focused strategy by taking fewer, bigger bets.

“Pick up a few areas but double down on them and see if you can drive value. We’ve seen many organisations go too broad and end up with a bunch of proofs of concept that never reach production,” he explained.

At the same time, leaders must also consider the potential vectors of disruption … changing consumer behaviour, disruptive competitors or startups, and new business models from emerging technologies could quickly change market dynamics.

Mohan noted a regional Southeast Asian bank that identified its wealth relationship managers (RMs) as a disruption vector after finding that they spent over 40% of their time on administrative work rather than client interaction.

“The bank is employing RMs for customer interaction, but a lot of their time was taken up in administrative tasks. So, they deployed AI agents to reduce these ‘no joy’ activities and make sure that the RMs can be more focused on the customers,” he said.

As the agents took over automatable administrative work, the organisation saw a 50% increase in productivity, while also freeing up capacity to take on more clients and markets.

Mohan noted that early deployments among Bain & Company's clients took about six to seven months for the employees to get used to the new tools.

As employees worked through these new use cases, they eventually got a better handle on the new tools, and by the third or fourth use case, they were able to move and pivot much faster.

“This velocity [to adapt to new use cases] is valuable in itself, as the capabilities of technology will continue to evolve,” said Mohan, adding that it is key that the organisation takes this capability and uses it as a strategic differentiator.

Moreover, for long-term AI scale, companies need strategies that shift growth economics—decoupling costs from revenue to scale impact without inflating operating expense.

Mohan said companies need to find ways to use AI to scale operations without a proportional rise in expense—whether through automation, workforce leverage, or new low-cost business models.

He shared the example of a bank exploring branches with almost zero operational cost, with KYC, document verification, and back-office processes handled by AI agents - aiming to grow its business without expanding headcount.

This approach changes the bank’s unit economics and enables financial inclusion in locations that were previously not viable.

“Companies need to think about the longer-term value they can get from these use cases. Consider using AI to ensure your cost and revenue curves are separated. If you can make a fundamental shift in the way your P&L works, that is a great long-term objective to push for,” said Mohan.

While organisations increasingly rely on AI-driven customer touchpoints, Mohan emphasised that transparency and structured design are prerequisites for trust.

Companies cannot simply implement AI systems and assume consumers will be comfortable with them, Mohan said, adding that there is still a need for the human touch in some instances.

“In one of our more complex installs, we have had interactions that are evaluated live, and if at any point in time you feel that the consumer-AI interaction is not going in the right direction, you can automatically switch it to a human.

“We still want that ability to be able to bring in a human, either to handle an exception or to just have the empathy and the ability to interface with a consumer beyond the basic rules that the organisation has in place,” said Mohan.

Besides that, when a customer interacts with an AI system, there must also be upfront disclosure for the sake of transparency - customers must know if they are speaking or chatting with an AI agent.

Mohan adds that relevant information, including how the customer can seek recourse in the event of an unfavourable interaction, must also be clearly communicated.

“It’s a composite of using the technology with a clear business objective and solving for the consumer experience,” said Mohan “It’s not just about deploying a tool.”

While comparisons have been drawn between AI’s hype and earlier technology waves, Mohan believes AI’s impact is unfolding differently, with value emerging far earlier for adopters.

He noted the hype around blockchain about five to 10 years ago, and more recently, the excitement around “the metaverse”, pointing out that these technologies were unable to scale value quickly.

“In many of the hype cycles in the past, the technologies were not in a place where they could create value immediately. Not to say there is no value; it is just that the value can only be realised over time.

“AI is different. You have probably personally experienced ChatGPT, Gemini or Claude, and you can see the value it adds. At an individual level, this is not the same as, let’s say, blockchain, which is a great technology but without the ability to create value immediately,” said Mohan. 

“The technology itself works - we’ve all seen it. That is not in question.”

However, he emphasised that realising AI’s value ultimately depends on how well organisations apply and scale it.

“This value creation we are talking about remains theoretical, unless we can get enterprises to actually deliver - and that is going to be important to double down on,” said Mohan.

Looking ahead, he believes the defining edge for companies will be their leadership quality, their ability to sustain the journey as a business transformation, and their ability to build an agile organisation that can absorb evolving technology.

Click HERE, to read the full report to explore how visionary leadership can turn today’s AI momentum into tomorrow’s competitive advantage.