Promptlē is a lending intelligence platform for document and data verification in lending workflows. It combines AI with real-world evidence checks to produce trusted, audit-ready outcomes across categorisation, extraction, fraud detection, serviceability validation, and policy execution.

How to automate document categorisation in lending?

Automating document categorisation involves using probabilistic AI models to identify document types (e.g., payslips, bank statements) based on visual patterns and text markers, achieving higher speed and accuracy than manual sorting.

Does Promptlē replace our lending application?

No. Promptlē does not capture the initial application. Your frontend or broker portal continues to handle the customer experience. Promptlē receives applicant data and documents via API or direct portal uploads, runs the verification pipeline, and returns structured results.

Can Promptlē automate approvals and declines?

Yes. Promptlē features a Credit Policy Engine that ingests your specific lending rules to automate decisioning. Alternatively, it can just pass verified data and AI reasoning back to your existing decision engine.

Is Promptlē fully modular?

Yes. You can consume the full platform or pick specific AI modules via API (e.g., just document categorisation and extraction, or just the serviceability and credit policy engine without document processing).

What data formats and integration standards does Promptlē support?

Promptlē supports a wide range of interfaces, including XML and JSON, giving all inputs and outputs seamless integration points with the Australian lending ecosystem. Promptlē also provides clean REST and GraphQL APIs.

Does Promptlē store our documents long-term?

No. Promptlē is a processing engine, not a Document Management System (DMS). Documents are encrypted during processing and automatically purged once the workflow is complete or expires, ensuring customer PII remains secure.

What is the difference between declared and verified serviceability?

Declared serviceability calculates borrowing capacity using applicant-stated figures. Verified serviceability recalculates using AI-extracted data from source documents like bank statements and payslips. Promptlē highlights discrepancies between the two.

Technical Glossary

Defining the core concepts of intelligent lending infrastructure and AI lending services.

Intelligent Lending Infrastructure

The underlying software framework and AI orchestration layer that automates document verification, data extraction, and credit policy execution between application capture and loan fulfilment systems.

Mortgage AI Middleware

A specialised software layer that sits between fragmented front-end application sources (brokers, portals) and legacy back-end banking systems, providing automated intelligence and data normalisation.

Standardised Data Interoperability

The use of common, structured data formats to exchange information between systems. Promptlē supports a wide range of interfaces, including XML and JSON, for seamless integration across the Australian mortgage and lending industry.

Stateless Processing

An architectural approach where sensitive data and documents are processed in-memory and purged immediately after the output payload is successfully dispatched, ensuring zero long-term storage of PII.

Verified Serviceability

The calculation of an applicant's borrowing capacity based on AI-verified data extracted from source documents (payslips, bank statements), as opposed to declared figures provided in the initial application.

Context-Aware Financial Extraction (CAFE)

An advanced AI technique used by Promptlē to understand the relationship between different financial data points on a document, leading to higher accuracy than traditional OCR.

Probabilistic Document Categorisation

The use of machine learning models to identify and classify hundreds of document types (e.g., payslips, tax returns, identity docs) based on visual and textual patterns with high confidence scores.

Deterministic Credit Policy

The execution of fixed lending rules and knockout criteria against verified data points to ensure 100% auditable and consistent decisioning outcomes without human bias.

Verified Output Payload

A standardised data packet produced by Promptlē that merges original application data with AI-verified truth, fraud indicators, policy outcomes, and a detailed reasoning audit trail. Delivered via a wide range of interfaces, including XML and JSON.

Zero-Touch Conditions Management

The automated identification and request for missing documentation or information required to satisfy a lender's credit policy, managed via real-time customer and broker portals.

Data Sovereignty (AU)

The principle that sensitive financial data remains within the legal jurisdiction of Australia. Promptlē ensures this by hosting all processing infrastructure in Sydney-region AWS.

Forensic Metadata Analysis

The examination of hidden digital markers within uploaded documents (e.g., software used to create/edit, timestamps, layer history) to detect sophisticated digital tampering or fraud.

Industry Data Schemas

Modern, JSON- and XML-friendly data structures used for granular data exchange in Australian mortgage origination. Promptlē's outputs are designed to integrate seamlessly with these evolving industry schemas.

Automated Serviceability Checks

The process of programmatically calculating borrowing capacity by comparing real-time AI-extracted income/expense data against lender-specific credit policy buffers and HEM benchmarks.

APRA CPS 234 Alignment

A prudential standard aimed at ensuring APRA-regulated entities (banks, insurers) maintain information security capabilities to withstand cyber threats. Promptlē's infrastructure is architected to align with these requirements.

Want to see these concepts in action?

Explore Solutions Request a Demo