AI-powered recovery in Malaysia

Malaysian banks are operating in an increasingly complex credit environment. Elevated household debt levels, post‑pandemic repayment stress, and rising non‑performing loans (NPL) have intensified pressure on recovery and collections teams. Meanwhile, Bank Negara Malaysia (BNM) continues to raise expectations around fair treatment of customers, transparency, and governance, particularly when automation and advanced analytics are used in decision‑making.

From reactive collections to intelligent, trusted recovery

While many banks have invested in AI, they’ve focused on front‑office use cases to enhance marketing automation and customer experience, where impact will be most keenly felt. Yet, one of the most immediate and measurable opportunities lies in AI‑enabled recovery and collections. Despite being one of the most outcome-critical functions in the bank, recovery and collections remain heavily manual and fragmented.

As a key customer touchpoint with the bank, intelligent and empathetic AI orchestration can be the difference between frustration and satisfaction for already distressed customers. By integrating AI directly into collections operations, Malaysian banks have an opportunity to converge financial outcomes, regulatory compliance, and build customer trust through customer‑centric recovery journeys.

Agentforce vs. traditional automation, and why it matters for BNM compliance

Transforming recovery and collections isn’t just about plugging in new tools. Many banks already use automation in collections, typically in the form of rules engines or robotic process automation (RPA). While useful, these approaches are limited as:

• Rules are static and require constant manual tuning
• RPA executes tasks but does not understand context 
• Decisions remain siloed and hard to explain end‑to‑end

The real value of your AI transformation lies in integrated platforms, clear governance, and humans working alongside AI as partners. Which is why Agentforce and Salesforce Financial Services Cloud (FSC) represent a step change.

While Salesforce Financial Services Cloud (FSC) unifies customer, product, interaction, and payment information into a single platform, Agentforce enables AI‑powered propensity‑to‑pay models that analyse behavioural, transactional, and engagement signals within FSC. 

So instead of simply automating tasks, Agentforce‑powered AI agents: 

• Interpret customer behaviour and context in real time

• Recommend next‑best actions rather than blindly triggering them

• Operate within policy, escalation, and human‑in‑the‑loop controls

• Produce auditable decision trails aligned to governance expectations 

What it does is allow banks to shift from chasing delinquency to anticipating behaviour.

Independent academic studies show that machine‑learning‑based risk and recovery models consistently outperform traditional rule‑based approaches. This allows for earlier identification of at‑risk accounts and more accurate recovery forecasting, all while remaining auditable for regulatory oversight.

For Malaysian banks managing large retail and SME portfolios, this proactive approach reduces both credit losses and operational strain. 

Furthermore, Agentforce reasons, prioritises, and collaborates with humans within defined governance guardrails. This distinction is critical in a Malaysian regulatory context. Banks will need to scale AI without compromising BNM’s expectations for transparency, accountability, and responsible use of AI. 

Five ways AI‑powered collections unlock value in Malaysia

1. More accurate risk segmentation using unified customer intelligence

Effective recovery depends on knowing who requires intervention, who needs flexibility, and who is likely to self‑cure.

FSC provides a unified customer view spanning repayment history, product exposure, service interactions, and behavioural signals. Agentforce uses this data to dynamically segment customers based on risk and recovery propensity.

Forecasting studies on NPL recovery demonstrate that behavioural signals during the recovery process (contact history, engagement patterns) significantly improve recovery rate prediction compared to rule‑based approaches.

For Malaysian banks, this supports more proportionate treatment strategies; a key BNM expectation.

2. Higher contact and promise‑to‑pay rates through personalised engagement

Malaysian customers increasingly prefer digital‑first engagement, which includes WhatsApp, email, mobile apps. But they want more than fast answers. They still expect empathy, consistency, and relevance even as they reach banks across multiple channels.

FSC orchestrates omnichannel outreach from a single platform, while Agentforce determines:

• When to engage

• Which channel to use

• What message or repayment option is most appropriate

According to Forrester, this personalised approach has driven 25–35% improvements in contact and promise‑to‑pay rates, compared to traditional collections methods.

More importantly, this reduces unnecessary repeat contact, which makes up a large proportion of customer complaints. From a governance perspective, centrally-managed engagement with your customers also simplifies monitoring, complaint handling, and regulatory reporting.

3. Faster recovery through connected sales, service, and collections

In many Malaysian banks, a customer may be interacting with a branch or relationship manager, a contact centre agent, and a collections officer, all without shared visibility.

FSC connects these teams on a single platform, while Agentforce ensures customer commitments, disputes, and service requests are visible in real time.

Salesforce users report that improved information flow across sales, service, and collections has resulted in an additional 10–15% reduction in losses, driven by faster, coordinated action. 

This directly improves recovery timelines while reducing customer friction.

4. Higher customer trust through fair, frictionless treatment

BNM places strong emphasis on the fair treatment of customers in financial difficulty. AI‑driven collections must therefore be consistent, explainable, and proportionate.

Agentforce enables banks to:

• Apply policy‑aligned treatment strategies consistently

• Escalate edge cases to human agents

• Adapt outreach based on customer response and vulnerability indicators

Salesforce finds that banks adopting frictionless, customer‑centric recovery journeys have achieved 20–30% higher customer satisfaction scores, while reducing customer attrition.

This strengthens trust with your customers at a time when they need it most.

5. Improved agent productivity without losing human oversight

Collections teams often spend a lot of time on administrative tasks, from case allocation to follow‑ups, and reporting.

Agentforce augments agents by:

• Automatically prioritising high‑impact cases

• Monitoring promise‑to‑pay commitments

• Recommending next‑best actions rather than enforcing them

Industry and academic studies consistently show that AI‑supported collections workflows reduce manual effort, shorten recovery cycles, and allow agents to focus on negotiation and complex cases.

While AI enhances productivity, it is your team who bring judgment and accountability for decisions augmented by AI. This move not only aligns with regulatory expectations, but allows banks to demonstrate a thoughtful and responsible approach to AI.

Scaling AI responsibly under BNM expectations

To sustain value and regulatory confidence, Malaysian banks will need to focus on:

• Unified, trusted data foundations via FSC

• Human‑in‑the‑loop governance for AI‑driven decisions 

• Transparent, auditable workflows aligned to BNM guidance

• AI agents that augment, not replace, human judgment

Agentforce is designed to operate within these guardrails, enabling scale without sacrificing control.

With the right implementation approach, banks can confidently demonstrate not only AI capability, but AI responsibility. A delivery partner with deep banking experience and who understands the technology and expectations shaping the collections practice in Malaysia can help deliver the impact you’re looking for. 

The path forward for Malaysian banks

As credit pressure persists and customer expectations continue to rise, recovery and collections will increasingly define a bank’s financial resilience and brand trust. AI, delivered through Salesforce Financial Services Cloud and Agentforce, can be an engine for growth. 

The PwC and Salesforce alliance combines strategic vision, risk and regulatory expertise, and industry-leading AI capabilities to empower and modernise banking operations for the future. Recovery and collections are no longer just about managing delinquency. They are about unlocking value through responsible, intelligent AI integration. So whether you’re exploring your options or want to fine-tune your agentic AI setup, we can help you implement and scale with confidence.