‘The cost can equate the benefit’: AIA Indonesia CHRO on the risks of AI adoption

As companies across Southeast Asia invest billions of dollars in artificial intelligence (AI), a key question remains: will the productivity gains justify the cost?

AI-related investment in Southeast Asia is expected to surpass US$110 billion by 2028, according to the East Asia Forum. However, technology spending alone does not guarantee business value. Companies must also account for implementation costs, employee readiness and changes to how work gets done.

In an exclusive conversation with People Matters Global, Ichsan Adiwijaya, CHRO at AIA Indonesia, warned that unchecked adoption could erode the benefits of AI.

“If we don't control it, we don't understand the applications of this technology, the cost can equate to the benefit of implementing it,” he said.

Not every problem needs the most powerful AI tool

Organisations must identify the problem they want to solve before choosing an AI platform, Adiwijaya said.

“We need to find the right platforms for the right capability that we want,” he said, comparing AI selection to choosing a vehicle. A long-distance journey may require a four-wheel-drive vehicle, while a short trip across Jakarta may only need a sedan.

“If you don't really pay attention to what platform you're using for what purpose, the cost is becoming bigger,” he said.

The lesson for business leaders is to match a tool's capabilities and running costs to the task at hand, rather than automatically opting for the most advanced technology.

The real cost of AI starts before deployment

AI investment extends beyond software costs to data readiness, infrastructure and changes to business processes. Adiwijaya said AIA spent years modernising its databases before the recent AI boom.

“We spent a couple of years to basically update and modernise our database,” he said, referring to customer, employee, market and product information.

“Every data that we hold, we need to make sure that they are accurate, they're updated and they're updated in a consistent manner.”

Without reliable data, AI outputs can be compromised. Organisations must also prepare employees and redesign processes to integrate the technology into everyday work.

“If you don't get that right, the AI would not be giving you the right answers because they're not dealing with the right data,” Adiwijaya said.

For HR leaders, this means factoring organisational readiness into AI investment decisions, rather than focusing solely on the cost of acquiring tools.

AI can assess candidates, but people still make the call

AIA has explored AI across finance, operations, IT and risk management, while also piloting AI-led interviews in recruitment, Adiwijaya said.

The system assesses candidates' facial expressions, demeanour and intonation to help determine who progresses to the next stage.

However, AI recommendations do not determine the final outcome.

“For the pilot, we still actually don't use the AI interview as a decision-making tool. We still interview the person regardless of whether the AI recommends or not,” he said.

AIA also reviews candidates who were not recommended by AI but were subsequently accepted, using their job performance to assess the system's accuracy.

“We go back to the AI application to then recalibrate the algorithm,” Adiwijaya said.

He stressed that HR professionals must understand how algorithms influence decisions and provide feedback to improve their accuracy. In recruitment, human oversight remains essential to prevent automated assessments from becoming the sole basis for hiring decisions.

Workforce planning must move beyond headcount

As AI changes how work is performed, HR leaders must assess its impact on individual roles rather than rely solely on traditional headcount planning.

“First and foremost, we need to have a deep business acumen for us to understand what's going on in each of the functions,” Adiwijaya said.

AI tools can already help employees summarise meetings, draft emails and analyse reports. At scale, these productivity gains can reshape responsibilities and staffing requirements.

Some roles may become more analytical as administrative work declines, while others may require reskilling or redeployment.

“Maybe they can be more productive, so I let them be and then they can do more work. So I don't have to recruit more people if the business is growing,” he said.

For employees whose roles are significantly affected, organisations must consider training and internal mobility.

“Maybe if they cannot fit with other jobs, we train them, upskill them, or maybe the job will be eliminated together,” Adiwijaya said.

He added that workforce planning must become more detailed as technology changes the nature of work. “In the old days, it's just basically headcount, headcount, headcount. Now it's more granular.”

The skills AI cannot replace

As AI takes on more tasks, employees need the judgement to evaluate its outputs and apply them to business decisions.

Adiwijaya cautioned that unchecked reliance on AI-generated analysis can create risks.

“If they're creating analysis recommendations that are incorrect, there's a big risk,” he said.

For HR leaders, the challenge is to equip employees to use AI effectively while questioning inaccurate outputs and understanding how technology affects their roles.

This requires a focus on critical thinking, business understanding and the ability to validate AI-generated information, alongside practical training in the tools themselves.

As Adiwijaya put it, “AI will not replace humans, but humans with AI will replace humans without AI.”

Making AI investment count

AI adoption requires businesses to weigh technology costs against the value it delivers. Platform selection, data quality, employee readiness and workforce redesign all influence the outcome.

“We need to understand how these cost drivers actually behave as we embark on the AI initiative,” Adiwijaya said.

He reiterated that organisations must choose platforms based on the capabilities they need, rather than simply opting for the most powerful tools.

For CHROs, this means working with technology and business leaders to identify valuable use cases, prepare employees for changing roles and retain human oversight where decisions affect people.

Ultimately, successful AI adoption is not about how many tools an organisation deploys, but whether they solve the right problems and deliver measurable business value.