Cloudcase, an Australian-founded digital origination and onboarding platform, has successfully demonstrated an artificial intelligence assistant completing a full home loan application in less than ten minutes. In a live walkthrough, the AI acted on behalf of a customer to execute a comprehensive application for two borrowers, managing everything from document verification and identity checks to eligibility assessment and final submission.
This demonstration brings the “machine customer” concept—identified by Gartner as a Top 10 Strategic Technology Trend—out of the realm of theory and into practice. Gartner forecasts that by 2030, at least 25 per cent of all purchasing decisions will be substantially delegated to machines.
“This is no longer theoretical,” stated Rian Fergusson, chief technology officer at Cloudcase. “We’ve demonstrated a complete, real-world lending process executed end-to-end by an AI assistant. Financial institutions need to be ready for customers who don’t fill out forms, they delegate.”
From conversation to conditional approval
The newly demonstrated workflow fundamentally shifts the user experience. The customer simply provides a single conversational instruction and uploads their supporting documents. From there, the AI assistant completely takes over, performing the following tasks:
-
Identifies suitable loan products.
-
Creates and completes the application.
-
Extracts and validates data from the provided documents.
-
Pre-fills hundreds of required fields.
-
Submits the application and receives conditional approval.
Crucially, throughout this entire process, the human customer never interacts with traditional forms, portals, or branch processes.
Intelligent reasoning and real-time problem solving
During the walkthrough, the AI assistant showcased advanced reasoning and transparency. It actively cross-referenced multiple documents and detected an address discrepancy. Rather than making assumptions, the system asked the customer to clarify the issue and requested missing information in plain language.
The system also demonstrated real-time problem-solving capabilities when faced with a serviceability shortfall. When the requested loan amount exceeded allowable limits, the AI clearly explained the issue, calculated the necessary adjustment, and guided the customer to resolve it within the exact same conversation. Following this collaborative adjustment, the application was seamlessly reassessed and approved.
An open AI ecosystem
Cloudcase integrates with these AI assistants via the Model Context Protocol (MCP), an open standard enabling AI systems to interact with external platforms. Because the capability is not tied to a single provider, any MCP-enabled assistant—including popular tools such as ChatGPT or Claude—can effectively act as a machine customer using the Cloudcase platform.
Developed through the company’s internal R&D program, this new capability is slated to be officially released to customers next quarter.
“The question is no longer whether this will happen, but how quickly,” Fergusson warned. “Institutions that can serve machine customers will capture a new generation of buyers. Those that can’t risk becoming invisible.”
