How $2.5bn PEXA secured and scaled their AI strategy with Redactive
Redactive's AI Data Security platform enabled PEXA to launch a permissions-aware, enterprise-grade generative AI assistant to 1,000+ employees in just 8 weeks.
About PEXA
PEXA is a world-leading digital property exchange platform and property insights solutions business.
Since 2014, PEXA has facilitated more than 15 million property settlements through the PEXA Exchange in Australia. Today, more than 90% of all property transactions in Australia are completed via PEXA.
PEXA facilitates property settlements on behalf of more than 160 financial institutions and 10,000 practitioner firms.
Industry
Financial Technology
Partners
Thoughtworks
AWS
Data Sources
SharePoint
Confluence
Key Highlights
A pioneering, permissions-aware AI assistant
Live data retrieval, with no data stored at rest
Improved overall data security posture
The Vision
A pioneering, permissions-aware AI assistant
Operating in a highly regulated environment, PEXA sought to launch a permission-aware AI assistant designed to empower employees and drive productivity improvements whilst ensuring secure use of enterprise data.
To achieve this, PEXA's AI assistant needed to understand each employee’s access to enterprise data in real time.
CHALLENGES
Existing data security solutions cannot sufficiently prepare and protect data for AI access
PEXA faced a common set of enterprise challenge when considering their AI strategy. While they wanted to harness generative AI to boost employee productivity and innovation, they couldn't risk their AI assistant exposing property and financial data.
Existing data security solutions do not adequately secure data for AI
PEXA faced a common set of enterprise challenge when considering their AI strategy.
While they wanted to harness generative AI to boost employee productivity and innovation, they couldn't risk their AI assistant exposing property and financial data.
AI tools typically require data to be copied from its source of truth
Because AI tools typically cannot access data at its source of truth, a copy of the data is required, which the AI tool then accesses.
This strips the data of its fine-grained permissions, exposing the potential for data leaks and creating a critical security risk.
Data security concerns were blocking AI progress
Like many other businesses that operate in regulated industries, PEXA could not move forward with their AI strategy without a solution to their data security challenges.
A Guide to the Emerging AI Security Risks for Enterprises
Learn more about the data security risks posed by Generative AI in enterprise environments.
SOLUTION
Redactive enabled PEXA to securely scale AI while maintaining high data security standards
PEXA partnered with Redactive, AWS, and Thoughtworks to implement a security-first generative AI assistant that revolutionized how their employees safely interact with enterprise data.
Unlike existing models that often rely on outdated architectures and lack fine-grained access control, PEXA’s solution leverages Redactive to ensure robust data security at both the model and data levels, retrieving live information from within the organisation.
The solution integrated Redactive's technology with Amazon Bedrock, creating a permissioned Retrieval Augmented Generation (RAG) system that ensures AI only accesses data that each employee is authorized to see, and controls what data is passed to the AI in small, secure segments.
Single source of truth for data
By accessing information directly from the source of truth without storing or copying data, all data maintains its fine-grained permissions when accessed by PEXA's AI assistant.
Redactive's approach minimizes risks of unauthorized data access while meeting compliance standards.
Real-time permission checks
Redactive enables real-time permission checks to ensure PEXA's AI assistant can only access data that each user is authorized to see, and controls what data is passed to the AI in small, secure segments.
Live data retrieval
Because Redactive enables the retrieval of live information from within the organization, PEXA's AI assistant is always able to leverage the most up-to-date information.
As a result, PEXA employees are able to trust that the information the AI assistant is returning is accurate and reliable.
Redactive's permissioned RAG enabled us to responsibly scale generative AI across our workforce without compromising our high data security standards.
Eglantine Etiemble
Group CTO
OUTCOMES
A significant step forward in PEXA's journey toward an AI-enabled future
Redactive enabled PEXA to build an AI assistant that doesn't just understand the company's information, but also understands who should have access to what information. This ensures that the AI assistant only returns information appropriate to the data access permissions of each individual user.
PEXA was able to successfully deploy enterprise-wide generative AI without compromising their data security posture.
Fine-grained control over data
With Redactive, PEXA was able to safely scale AI access across the workforce while maintaining fine-grained control over unstructured data.
Every AI interaction is secured
Redactive enables PEXA to ensure every employee interaction with AI remains secure and contextually appropriate based on the individual's data access permissions.
Improved overall data security posture
With Redactive, PEXA was able to improve their overall source data permissions to ensure continuously maintain appropriate data access, and future-proof their AI strategy by securing the foundations for future AI agent adoption.
Is data access debt blocking your AI rollout?
Learn how Redactive enables security and engineering teams to gain precise, contextual control over what data employees, agents, and AI applications can access—ensuring unauthorized data exposure is prevented.