Lendi lifts loan lodgement with AI agents & Twilio

Lendi Group has built two AI agents on Twilio's communications platform for its home loan process, with the system increasing the rate of customers moving from initial enquiry to loan lodgement by 2.5 times, according to the companies.

The work centres on two custom agents, Guardian and Linda, used across the group's Lendi and Aussie brands. The tools combine voice, SMS and email into a single customer view, allowing brokers and AI systems to respond to the same customer activity during a loan application.

The change is intended to reduce the breaks that often appear as customers move between brokers, lenders and other parties during a home purchase. In mortgage applications, those handoffs can slow progress and force borrowers to repeat information across channels.

In Lendi Group's updated app-based journey, Guardian, the lead intake agent, asks qualifying questions when a customer responds by text or voice. It can then match those responses to live product options in real time.

Linda, the outbound agent, follows up on missing information and selects the channel and message for further contact while observing customer communication preferences. The system has also been used to re-engage borrowers who had stopped responding during the process.

Over a six-month period, around 20% of customers completed fact-finding and received product options, compared with less than 3% through the earlier app-based process, according to Lendi Group. It also said about 90% of customers who engage with Guardian continue the conversation, while up to 16% of previously disengaged high-intent customers return to the process after personalised AI outreach.

Some tasks that previously took weeks can now be completed in under an hour, the company said, allowing many applicants to get through much of the process in a single sitting. That speed matters in a market where borrowers often compare products, gather documents and wait for multiple parties to respond before an application can proceed.

Closing gaps

Devesh Maheshwari, Chief Technology Officer, Lendi Group, said the company's approach is based on acting on live customer signals rather than simply logging events after they occur.

"The customer does not experience your organisation chart or your systems architecture. They experience the gaps between them. A traditional organisation uses software to record activity after it happens. As an AI-native organisation, we use data, signals and agents to understand what is happening now and what should happen next," said Devesh Maheshwari, Chief Technology Officer, Lendi Group.

Twilio provides the communications layer linking the different channels customers use during the application journey. Lendi Group said that layer is integrated with its Aurora and Vectorium systems, which hold customer context, agent logic and workflow connections for lending tasks.

The companies are also testing Twilio Conversation Relay for real-time AI voice interactions. The pilot is intended to explore whether voice can extend the same shared customer context already used in digital channels.

Conversation Relay handles functions including speech recognition, text-to-speech and conversational interruption handling, while Twilio Voice provides the live voice connection. This setup allows voice interactions to be attached to the same customer record and agent flow as messaging and email exchanges, according to Lendi Group.

Maheshwari said the voice element could improve the quality of information reaching brokers.

"Conversation Relay is interesting because it helps us think beyond one-way communication. It creates an opportunity for more interactive engagement, where the customer can respond, the system can capture intent, and the broker can receive better context," Maheshwari said.

Lendi Group operates a national network of about 1,350 brokers and 215 retail stores, with staff across Australia and the Philippines. Its wider business spans property search, buyer advocacy, mortgage broking, conveyancing and home ownership tools.

For Twilio, the project offers an example of how financial services groups are combining communications software with AI systems in customer-facing processes that require high levels of trust and record-keeping. Mortgage applications are among the more complex retail financial journeys because they involve sensitive personal data, changing customer circumstances and input from several external parties.

Robert Woolfrey, Vice President, Asia Pacific & Japan, Twilio, said shared customer context across channels was central to the work.

"Lendi Group demonstrates what is possible when communications are treated as a foundation for intelligence rather than disconnected tools. Unifying voice and messaging with AI preserves critical customer context and enables both AI agents and brokers to work from a shared understanding," said Robert Woolfrey, Vice President, Asia Pacific & Japan, Twilio.

"Borrowers navigate their journey without repeating themselves, while brokers gain better visibility into what customers need. We expect this signal-driven approach to define customer engagement across Australia's highly regulated financial services market," Woolfrey said.