Lendi Group Overhauls Data Architecture With MongoDB

Australian fintech Lendi Group has rebuilt its data infrastructure on MongoDB Atlas as part of an ambitious plan to become a fully AI-native business by June 2026, claiming the shift has already cut time to market for AI features by 40 per cent.

The company, which has $107 billion of home loans under management, operates an integrated ecosystem spanning property search, buyer advocacy, mortgage broking, conveyancing and ownership tools. Its platform is increasingly powered by AI agents that assist both brokers and customers through the lending process.

But getting to that point required Lendi Group to rip out a fragmented legacy data environment and replace it with something built to support AI workloads from the ground up.

Untangling A Post-Merger Data Problem

The complexity of Lendi Group’s previous infrastructure traces back to its merger with Aussie Home Loans. Over time, the combined organisation’s technology stack had grown to more than 500 deployable components, built across a mix of relational databases like PostgreSQL and various non-relational systems.

That architecture had become expensive and time-consuming to maintain. More critically, it lacked the consistency and flexibility the company needed to build and deliver AI-driven services at pace.

For Lendi Group’s AI agents to function effectively, they required a complete, real-time view of each customer. That meant pulling together complex and diverse data sets including property data such as real-time valuations, suburb trends and geospatial information, alongside finance data from credit reports and Open Banking feeds, and behavioural data covering customer goals, interactions and platform usage patterns.

Mortgage broking is also a tightly regulated industry, which added another layer of complexity. Lendi Group needed an AI platform that was compliant from the outset and could adapt as regulations evolved.

Why MongoDB Atlas

The company settled on MongoDB Atlas within the first week of its operational data layer project.

Will Hargan, Senior AI Systems Engineer at Lendi Group, explained the reasoning was clear-cut.

“There simply wasn’t another option that offered the flexibility of the document model and the power of MongoDB’s integrated, AI-ready data platform,” Hargan noted.

Four requirements drove the decision. First, the organisation needed the ability to manage complex data structures. Lendi Group adopted what it describes as a “document first” approach, creating a unified schema strategy that standardises data contracts across domains.

Second, MongoDB’s built-in AI capabilities, including its vector search functionality, allowed the team to prototype and iterate on AI applications without introducing a separate vector database.

Third, scalability was a consideration. MongoDB’s native horizontal sharding gave Lendi Group a path to handle the data growth anticipated from its expanding AI capabilities without creating an operational burden.

Fourth, security and compliance features built into the platform enabled Lendi Group to maintain a continuous audit trail at the database layer, supporting governance controls, traceability and regulatory compliance.

Building The Operational Data Layer

The operational data layer, or ODL, serves as the foundation for what Lendi Group describes as a shift from a “human-motion” to an “agentic-motion” business.

In practical terms, that means using AI agents to automate the time-consuming, lower-value parts of the mortgage process – document checks, follow-ups and rate monitoring – while freeing human brokers to focus on complex structuring and customer guidance.

The first product to emerge from the new architecture was Lendi Guardian, which was built and delivered in a 12-week development cycle.

Devesh Maheshwari, Chief Technology Officer for Lendi Group, pointed to the speed of that delivery as evidence of the platform’s impact.

“MongoDB has given us operational simplicity and incredible developer velocity for AI features,” Maheshwari remarked. “The successful launch of Lendi Guardian demonstrates the speed and quality of what we’re able to do now.”

Faster Feature Delivery

The company reports that its developers are now bringing new features to production 40 per cent faster than they could under the previous data architecture.

That improvement in velocity is central to Lendi Group’s broader ambitions. The company is positioning itself not just as a user of AI tools but as a business that is structured around AI operations at a fundamental level.

Looking ahead, Lendi Group intends to use its MongoDB-powered ODL to build what it calls an “elastic workforce” – a model where AI agents handle routine processes while human brokers concentrate on areas that require empathy, complex decision-making and trust.

The company has set June 2026 as its target date for reaching fully AI-native status, a timeline that will test whether its rebuilt data architecture can deliver on the pace of innovation Lendi Group is banking on.

Last Updated on March 21, 2026 by Nick Ross