AI startup case study

Background

Hundreds of pages of transcripts (Hansard) are published each Parliament sitting day.

For those whose jobs depend on knowing what happened in Parliament it can be time consuming to keep up, especially when you only care about specific topics (eg: a government affairs manager at a leading telecommunications company might only want to know what was said about telecomms).

The work

We ingested historical Hansard data and developed a process to ingest new transcripts each sitting day.

Users can login to a web application and can access the information they need in one of 2 ways:

• Hansard Chat: using AI, users can ask questions about what has happened in Parliament. For example, you might want to know what a particular member's stance is on onshore manufacturing, or what Labor has said about tariffs in the past 18 months.
• Hansard Podcast: users can enter the topics they care about (Economy, or AI, or telecommunications etc), choose a date range and get an AI generated podcast about what was said in regards to their chosen topics on the sitting days they have chosen. The user can dictate the depth that the podcast should go into and set a podcast length. The AI generates a podcast based on your chosen settings.

Services

Since the launch of Claude Opus 4.6 in February 2026 we have focussed on using AI more and more to write code. We really leaned in to this approach on this project. The Auspol project was completed at least 50% faster than pre-AI coding.

Web design Web app development AI Agentic developers

View project