Southern Cross Credit Union Deploys Salesforce Agentforce Voice to Slash Wait Times

Southern Cross Credit Union (SCCU) has deployed "Nova," an AI-powered voice agent built on Salesforce’s Agentforce Voice, to manage routine balance inquiries. By automating these high-volume calls, the credit union aims to reduce wait times for urgent issues like fraud reporting, marking one of the first major implementations of this technology in the Australia and New Zealand region.

What was announced

SCCU, a customer-owned financial services provider, launched Nova to address a surge in wait times, which had climbed from under two minutes to nearly eight. Nova is integrated with SCCU’s Vonage contact centre solution and currently handles 7% of all incoming customer calls. For those specifically selecting the balance enquiry option, Nova successfully provides account details to 95% of verified customers.

The system uses Agentforce Voice to understand natural language, including varied accents, phonetic variations, and customer-specific nicknames for accounts rather than requiring formal account numbers. Security is handled through a verification process that SCCU claims is more secure than traditional one-time passcodes. This allows the credit union to treat its service centre as its largest branch, catering to regular customers who call several times a week for routine updates on recurring payments like pensions.

The deployment was completed in six weeks in collaboration with Salesforce Forward Deployed Engineers (FDEs). The development process included workshops at SCCU’s head office to define Nova’s personality, tone, and levels of empathy, followed by testing in an Agentforce Testing Centre to simulate diverse dialects. Nova also features sentiment monitoring; if a caller appears frustrated or distressed, the agent routes the call to a human staff member, passing along the full conversation context to prevent the customer from having to repeat themselves. Safeguards are also in place to automatically escalate calls based on specific vulnerability criteria.

The companies involved

Southern Cross Credit Union is an Australian customer-owned financial services provider. Operating as a credit union, the organisation prioritises its member-owners and treats its service centre as its largest branch rather than a traditional contact centre. This focus on service led the institution to seek automation solutions that could handle routine inquiries without compromising the customer experience for more complex or urgent needs, such as lost cards or scam reports.

Salesforce is a major global provider of customer relationship management (CRM) and cloud-based software solutions. The company has a significant presence in the financial services industry, offering tools designed to streamline customer interactions and data management. Its recent development of the Agentforce platform marks a strategic shift toward autonomous AI agents that can be integrated into existing workflows, such as contact centres. Salesforce’s Forward Deployed Engineers (FDEs) work directly with clients to implement these new technologies, as seen in the rapid six-week deployment of the Nova voice agent. The company has been a frequent subject of industry reporting, reflecting its role as a primary infrastructure provider for digital transformation in banking.

What this means

The deployment of Nova signals a maturation of voice AI within the financial sector, moving beyond simple IVR menus to agents capable of understanding intent and sentiment. For the broader fintech market, this implementation serves as a proof of concept for "agentic" AI—systems that do not just follow a script but interact dynamically with core banking data. The challenge for the industry now lies in maintaining security and trust; while SCCU claims their verification is more secure than passcodes, the sector will be watching closely to see how these autonomous systems handle the nuances of identity theft. This move puts pressure on larger incumbents to move beyond basic chatbots toward more sophisticated, voice-first automation.