AI adoption is accelerating across Malaysia's financial sector. According to Bank Negara Malaysia (BNM), 71% of banks have already adopted some form of AI — though most applications centre on fraud detection, customer analytics, and digital identity verification (confirming identity via a smartphone or an online platform), rather than on loan assessment itself.

A few players are ahead of the curve. Funding Societies, one of Malaysia's leading P2P lending platforms, has been developing machine-learning-based credit-scoring models. Among the five BNM-licensed digital banks, two — GXBank and Ryt Bank — have incorporated AI into their consumer loan assessment processes. Beyond these pioneers, however, AI-driven loan screening remains uncommon across the broader lending industry.

What Is AI Screening? How Bee Informatica Is Approaching It

AI screening refers to the use of artificial intelligence to automatically analyse financial data and interview content, generating an assessment score that supports and accelerates human decision-making. It is gaining significant attention across the financial sector.

Bee Informatica has been actively incorporating the latest developments in the field into its own operations. Our AI-powered assessment process is structured around two key stages.

Stage 1: Quantitative Assessment — Automated Financial Data Analysis

Applicants upload financial documents such as bank statements to our system. The AI then automatically reads and analyses the data, calculating an assessment score based on metrics such as the average monthly balance over the past six months. Tasks that were previously performed manually — data extraction, aggregation, and analysis — are now handled by AI, resulting in significantly faster processing and greater objectivity.

Stage 2: Qualitative Assessment — Evaluating the Business Owner

The second stage assesses the business owner's understanding of their operations, their command of financial figures, and their overall financial literacy. Previously, this was done through online video interviews (camera on), with assessors manually recording scores.

We are currently developing the next step: a unique AI system that transcribes interview recordings and audio, then generates an automatic five-level score across three dimensions:

• Emotional stability, as read from voice tone.
• Consistency of responses across the interview
• Accuracy and reliability of the information provided

With this system in place, what currently takes an assessor about 30 minutes is expected to take 5 minutes. Final lending decisions continue to be made by our credit committee, who review and finalises scores based on the AI-generated outputs.

Why Now? The Case for AI Assessment

There are two main reasons behind our decision to introduce AI-powered screening.

The first is delivering better outcomes for our customers. Whereas traditional banks typically take two to three months, the approval process now takes just two to three working days. By utilising AI, we aim to complete the transfer in as little as one working day. This reduction in processing time offers a significant advantage to customers who require funds urgently.

The second is handling rapid growth in assessment volume. The number of applications we process has grown severalfold in recent years, and further increases are expected. To maintain quality while operating with a lean team, improving efficiency through AI is not just an option — it is a necessity.

An Unexpected Discovery: What AI Reveals That Humans Miss

As we developed our AI screening system, we gained insight we had not fully anticipated: the richness of voice data.

In traditional interviews, assessors naturally focus on what is being said — the content and meaning of the words. As a result, non-verbal signals such as tone of voice, confidence, and emotional steadiness often go unnoticed. AI, however, can systematically analyse these acoustic qualities and convert them into measurable data points.

This means that AI screening does not simply make the process faster — it surfaces information that human assessors are unlikely to catch on their own. For Bee Informatica, the deeper value of AI lies not just in speed, but in new layers of insight.

Looking Ahead: Expanding AI Across the Business

Bee Informatica is committed to continuously improving the accuracy and scope of our AI-assisted assessment. Beyond the screening process itself, we are extending AI capabilities into other areas of the business:

• Customer communication: An AI-powered communication tool has already been deployed.
• Early-stage collections: AI-assisted outreach is being used to support the initial stages of our collections process.

By harnessing technology to improve the quality of service for our customers and the efficiency of our internal operations, Bee Informatica is working to set a new standard in Malaysia's lending industry.

Reference: Fintech News Malaysia – List of Digital Banks in Malaysia

https://fintechnews.my/48009/digital-banking-news-malaysia/list-of-digital-banks-in-malaysia/

Reference: The Edge Malaysia

https://theedgemalaysia.com/node/784279

‍