
Two factors which are the sophistication of fraud practices and the increase in demand for personalized banking are pushing banks to turn to artificial intelligence (AI) for security.
While others adapt, Maya is building its own system, customizing services, and scaling digital finance.
At the recent Amazon Web Services (AWS) AI Day, Maya Group chief technology officer Alfred Lo unveiled homegrown AI breakthroughs, including a fraud detection technology that can reduce losses by up to 98 percent, a chatbot that handles most customer inquiries, and a recommendation system that boosts user engagement.
“AI is reshaping banking to ensure every Filipino benefits from the digital economy,” Lo said.
“At Maya, we’re competing with AI itself — to stay ahead of fraudsters, shifting customer behaviors, and a digital economy that never stops moving.”
The Bangko Sentral ng Pilipinas (BSP) has flagged an increase in scams, phishing attacks and account takeovers, particularly among new bank customers.
However, most fraud detection systems still rely on rule-based methods, AI with simple feature engineering, or graph-based techniques.
Maya’s innovative fraud detection approach, Transaction Sequence Embeddings, analyzes and uncovers subtle patterns between transactions, flagging those that resemble fraudulent behavior or deviate from a user’s normal patterns.
The company said it is finalizing a technical paper showcasing proprietary innovation for publication in a premier journal.
“Our model can reduce fraud losses by up to 98 percent, significantly enhancing the effectiveness of anti-fraud systems,” Lo said.
The use of AI can also assess credit risks, predicting repayment likelihood based on spending habits. However, security is just one way Maya is driving innovation.
Most banks still segment customers by broad categories like age or income, missing real behavioral insights.
Maya has built an AI-powered Action Card feature on the Maya app to guide users at the right time, suggesting products that are highly relevant. Maya’s personalized approach has helped increase average revenue per user by 45 percent.
The next frontier would be an AI-powered rewards system that learns from past offers to predict which incentives, like free insurance or higher interest to deliver the most value to millions of users while remaining cost-effective. This ensures that every peso spent on rewards maximizes impact.
“Now, we enter the era of multi-objective optimization. AI orchestrates and balances cost, conversion, and customer delight. At Maya, we’re at the forefront of this evolution. We turn data into decisions, and decisions into delight.” Lo added.
Yet, even the best banking services mean little if customers struggle to get support. AI is addressing that, too.
Most banking chatbots handle only basic inquiries, with limited responses. Maya’s AI chatbot, however, now resolves up to 95 percent of customer issues instantly, reducing the need for human support tenfold while improving customer satisfaction.
Lo sees the future of AI beyond customer service — turning chatbots into full-fledged financial assistants.