This Register details KPMG Australia's AI-enabled solutions used for client services and external stakeholders. It includes descriptions of each solution, the type of AI employed, and an overview of applied controls.
Updated periodically, this Register underscores our commitment to AI transparency.
At KPMG Australia we develop, procure, use and sell a range of AI tools. These tools help us to manage our business with greater effectiveness and efficiency and enhance our services and products.
In all we do, we are guided by our Trusted AI Framework and continually review and update our approach to align with current best practices and regulatory standards. KPMG Australia has been certified as complying with ISO 42001:2023 (AI Management System).
Overview
KPMG Clara AI – chat is KPMG’s proprietary AI – chat solution, specifically for Audit & Assurance. It is hosted on the KPMG’s Microsoft Azure Cloud and leverages the standard language model from Azure OpenAI. AI – chat is designed to improve productivity, research and efficiency in our audit and assurance engagements, as well as to be used for other administrative tasks.
AI Type
Generative AI
Controls
KPMG is committed to the responsible design, development and deployment of AI solutions. Consistent with the firm’s Trusted AI Framework.
KPMG's use of Clara AI meets leading standards of responsible AI, including with regard to human oversight, data governance, privacy, and security.
Key contact
Overview
KymChat is an internal AI agent developed to enhance the way KPMG interact with KPMG data and knowledge resources. Built on Azure OpenAI models, KymChat allows access and leverage internal KPMG information.
AI Type
Generative AI – Chat GPT 4 Turbo
Controls
KPMG is committed to the responsible design, development and deployment of AI solutions consistent with the firm’s Trusted AI Framework.
KPMG KymChat meets leading standards of responsible AI, including with regard to human oversight, data governance, privacy, and security. KPMG KymChat has undergone a Trusted AI Impact Assessment and is a certified ISO42001:2023 AI solution.
Key contact
Overview
KymTax, a custom-built generative AI tool designed to put the breadth and depth of KPMG’s Tax proprietary knowledge at the fingertips of the firm’s tax specialists.
AI Type
Generative AI – Chat GPT 4o
Controls
KPMG is committed to the responsible design, development and deployment of AI solutions consistent with the firm’s Trusted AI Framework.
KPMG KymTax meets leading standards of responsible AI, including with regard to human oversight, data governance, privacy, and security. KPMG KymTax has undergone a Trusted AI Impact Assessment and is a certified ISO42001:2023 AI solution.
Key contact
Overview
GitHub Copilot, an AI generative tool to help users improve their code quality and coding efficiency both on internal activities and on Salesforce related client engagements. The tool integrates out of the box with a number of Integrated Development Environment (IDE) solutions.
AI Type
Generative AI
Controls
KPMG is committed to the responsible design, development and deployment of AI solutions consistent with the firm’s Trusted AI Framework.
KPMG's use of Github CoPilot meets leading standards of responsible AI, including with regard to human oversight, data governance, privacy, and security.
Overview
KPMG Regulation Sync uses Generative AI capabilities in Kym Chat to enable our KPMG Climate Change and Sustainability subject matter experts perform a compliance gap analysis in their client’s sustainability report disclosures based on AASB S2 standards. The results of KPMG Regulation Sync get categorised into three categories of fully compliant, partially compliant, and non-compliant. Regulation Sync also provides the subject matter expert with reasoning on why it has come up with its assessment by providing citation to the relevant chunks of sustainability reports it has used in its assessment.
AI Type
LLM, GPT-4o, and Semantic Processing
Controls
KPMG is committed to the responsible design, development and deployment of AI solutions. Consistent with the firm’s Trusted AI Framework.
KPMG Regulation Sync meets leading standards of responsible AI, including with regard to human oversight, data governance, privacy, and security. KPMG Regulation Sync has undergone a Trusted AI Impact Assessment and is a certified ISO42001:2023 AI solution.
Key contact
Overview
The Tax Classifier enabled by Kym uses AI, including generative AI to bridge gaps, thereby accelerating and enhancing classification accuracy for users internally. This tool provides users across Tax and Legal at KPMG with self-serve access to a web-based classification tool, which augments existing working methods and increases efficiency.
AI Type
Machine Learning, Natural Language Process, Generative Pre-trained Transformer, Generative AI, Large Language Model
Controls
KPMG is committed to the responsible design, development and deployment of AI solutions. Consistent with the firm’s Trusted AI Framework.
TaxClassifier meets leading standards of responsible AI, including with regard to human oversight, data governance, privacy, and security. TaxClassifier has undergone a Trusted AI Impact Assessment and is a certified ISO42001:2023 AI solution.
Key contact
Overview
Microsoft 365 Copilot is an AI-powered productivity assistant that integrates with Microsoft Office 365 applications, including Word, Excel, Outlook, and Teams. It enables natural language prompting of large language models combined with data in an individual's Microsoft Graph to generate more tailored draft content and analyses.
AI Type
Generative AI
Controls
KPMG is committed to the responsible design, development and deployment of AI solutions consistent with the firm’s Trusted AI Framework.
KPMG's use of Microsoft 365 CoPilot meets leading standards of responsible AI, including with regard to human oversight, data governance, privacy, and security.
Overview
Synthesia is a generative AI tool that creates 'talking head' videos based on scripts written by the user.
The platform allows the user to create and edit videos using synthetic voices and images.
AI Type
Generative AI
Controls
KPMG is committed to the responsible design, development and deployment of AI solutions consistent with the firm’s Trusted AI Framework.
KPMG's use of Synthesia meets leading standards of responsible AI, including with regard to human oversight, data governance, privacy, and security.
Overview
KymR&D is an intelligent GenAI workflow that supports the structuring and referencing of R&D information into a project description, ensuring quality, consistency, and alignment to the eligibility criteria.
AI Type
Generative AI – Chat GPT 4o
Controls
KPMG is committed to the responsible design, development and deployment of AI solutions. Consistent with the firm’s Trusted AI Framework.
KymR&D meets leading standards of responsible AI, including with regard to human oversight, data governance, privacy, and security. KymR&D has undergone a Trusted AI Impact Assessment and is a certified ISO42001:2023 AI solution.
Key contact
Overview
KPMG Learning Design Studio enabled by Kym, is a custom-built application that accelerates KPMG's learning offering in delivering transformational learning and development outcomes for organisations and their people.
AI Type
Generative AI – Chat GPT 4o
Controls
KPMG is committed to the responsible design, development and deployment of AI solutions consistent with the firm’s Trusted AI Framework.
KPMG Learning Design Studio enabled by Kym meets leading standards of responsible AI, including with regard to human oversight, data governance, privacy, and security. KPMG Learning Design Studio enabled by Kym has undergone a Trusted AI Impact Assessment and is a certified ISO42001:2023 AI solution.
Key contact
Overview
BreachAI leverages the power of AI for KPMG Auditors, supporting rules-based assessments required to process and identify regulatory breaches of the audited entity.
AI Type
LLM, GPT-4o, AI Persona, Prompt Chain (Flowise)
Controls
KPMG is committed to the responsible design, development and deployment of AI solutions. Consistent with the firm’s Trusted AI Framework.
BreachAI meets leading standards of responsible AI, including with regard to human oversight, data governance, privacy, and security. BreachAI has undergone a Trusted AI Impact Assessment and is a certified ISO42001:2023 AI solution.
Key contact
