This article first appeared in CFOtech on Friday 6th March 2026.
In every boardroom in Asia, and even most tearooms, artificial intelligence (AI) is a common topic of conversation. No one knows exactly how it will play out, but three things seem all but certain.
First, agentic AI - meaning systems that can independently make decisions and collaborate with people and other agents - will be a true game-changer. In banking, for example, AI agents will soon be able to complete every step of a customer interaction, from answering initial queries to offering complex advice.
Second, the greatest potential is not from agents taking the place of people, but from both working together. AI agents can make decisions; indeed, they are being designed to do so. But people must always be in the loop to ensure the right questions are asked, and to make the call on the most important matters.
And finally, for organizations to get the most out of agentic AI, women need to be part of the process. If they are not, today's problems could be automated into tomorrow's systems. There is also a straightforward business argument. In 2018, McKinsey research found that companies in the top quarter of female representation in their executive teams were 21 percent more likely to outperform in terms of profitability compared to those in the bottom quarter, and 27 percent more likely to outperform in terms of value creation.
With AI rising, then, the case for women to be solidly engaged is strong. Even though women comprise half of AI end users, however, they are under-represented when it comes to its design and deployment. In 2024, they made up only 29 percent of AI-skilled workers; a UNESCO study found that in 2019, only one in eight AI researchers globally were female.
All this can have significant real-world consequences. AI agents use what they are fed. For example, if AI uses historical data, and that data shows that men hold most senior roles, AI may learn to associate leadership with male traits, affecting everything from hiring to performance reviews. In recruitment, AI tools can undervalue resumes with career breaks, which for women are often related to caregiving. As AI agents take on more responsibility, the question of who designs them - and therefore who decides what "good" looks like - assumes ever greater urgency.
Looking specifically at the Asian banking sector brings clarity to the risks, and the opportunities. This is an area where Asia could stand out. Agentic AI is already primed to reshape banking across the region. Moreover, the regulatory environment is broadly positive, with many national regulators encouraging banks to innovate with AI. Even so, McKinsey has estimated that 60 to 70 percent of workflows in Asia's banking operations still rely on manual processes; that is a situation that calls for reimagination.
If banks don't employ mixed-gender teams to build their agents, women's particular needs and experiences are likely to be overlooked in decisions, workflows, and customer services. That matters, because when it comes to money, the sexes are different. Compared to men, McKinsey research has found that women are more likely to prefer stable investments and to prioritize long-term financial security.
Moreover, women are a growing financial force. From 2018-23, the amount of global wealth controlled by women rose 51 percent - markedly faster than overall growth (43 percent). Ensuring that women are embedded in the development and rollout of agentic AI, then, simply makes business sense. In addition, women's strengths in communication, collaboration, and long-term thinking could be assets in designing systems that are seen not only as accurate but fair. In a field like banking, where trust is the most important currency, that could translate into value.
The theme of this year's International Women's Day is "give to gain" - a slogan that is relevant here: by giving women more opportunities, organizations stand to gain an edge. But inertia is also a powerful force. Getting more women into influential AI positions is unlikely to just happen. Instead, there needs to be real commitment. For banks and other organizations, that means embedding inclusion into how AI systems are designed. It can also be helpful to set clear metrics for model bias, and to do gender assessments in project planning.
For policymakers, the same principle applies: that it is better to get off on the right foot than scramble to recover. That means designing and enforcing anti-discrimination laws in algorithmic decision-making and supporting programs that boost women's participation. And for women, the imperative is to get in the game. AI presents a chance to "skill up." Don't settle for being passive users; instead, seek to be active creators and decision-makers.
Asia's banks are moving toward a system in which fleets of AI agents work with people to make decisions and complete tasks. That is a fundamentally different operating model from ten, five, or even two years ago. And it is an opportunity not only to correct the mistakes of the past that sidelined women, but to construct an agentic AI model that fosters values like fairness and equity, while also building competitive advantage.
Violet Chung is a senior partner in McKinsey & Company's Hong Kong office.