​​Artifical intelligence

 

Gulanga brings together sovereign AI, an agentic delivery factory, and a suite of platforms under one Indigenous-led brand. It is built on a simple conviction that has guided Gulanga from the start: the oldest continuous culture and the newest contemporary technology belong together.

 

The portfolio spans three layers, a sovereign AI engine and its conversational interface (Dukes and Yarning Circle), the factory that builds and ships software (Pathfinder), and the products that factory delivers.

​

Dukes - Sovereign Indigenous AI engine

 

The sovereign large language model and agentic engine at the heart of Gulanga.AI.

​

Dukes is named in honour of Craig Dukes, a proud Mara and Jingili man who founded Gulanga in 2015. A trailblazer across Indigenous health and technology, Craig moved with strength and wisdom between Indigenous and non-Indigenous worlds and opened pathways for First Nations people in ICT through mentorship, scholarship and opportunity. He passed away unexpectedly in June 2022. Naming Gulanga's AI engine after him carries that legacy forward, a technology that walks confidently in both worlds and makes sure no one walks their journey alone.

​

Dukes is delivered through an OEM partnership with an Australian sovereign AI platform provider. That gives Dukes a foundation that is fully Australian-owned and operated, hosted in Australia, Defence-grade data centres, aligned to the Information Security Manual and engineered so customer data never leaves Australian soil.

​

On top of that sovereign foundation, Gulanga adds what makes Dukes distinct - data sovereignty and governance. Data ownership, access and use are governed in line with Data Sovereignty principles, so agencies and communities retain authority over their own knowledge rather than surrendering it to an offshore model.

​

In practice, Dukes provides:

• Agentic AI - autonomous and semi-autonomous agents that can reason, retrieve and act within defined guardrails

• Conversational access to an organisation's entire corporate knowledge base

• On-shore hosting and encryption suitable for security-sensitive environments

• A governance layer built around data sovereignty, and transparent, auditable use

​

 

Yarning Circle - the conversational interface for Dukes

​

Where people yarn with their organisation's knowledge.

​

Yarning Circle is the chat interface that sits on top of Dukes. It turns the engine into an everyday tool: a secure, plain-language space where staff can ask questions, draft and analyse work, and put agents to task, all grounded in their own trusted, onshore knowledge.

​

The name draws on the yarning circle, a long-standing First Nations practice of sharing knowledge, listening and making decisions together as equals. That is the experience the product is designed to create, not a black box that hands down answers, but a place where people and AI reason together, with the source of every insight visible and open to challenge.

​

Yarning Circle offers:

• A clean conversational experience for individuals and teams

• Retrieval-grounded answers with transparent sourcing and evidence

• Access to Dukes agents with appropriate permissions and approval workflows

• The same sovereign hosting, security and data-governance as Dukes itself

​

​

Pathfinder - the AI software factory

 

The delivery operating system that turns ideas into shipped products every time.

​

Pathfinder is Gulanga’s software factory, a structured delivery operating system that integrates process, workspace and AI agents into one native experience built for how product teams’ work.

​

At its core is the HASTE process - Hypothesise, Assess, Specify, Throughput, Evaluate a repeatable, five-stage path from a belief to a shipped product, with clear stage gates and accountability at every step. Around that process sits a connected workspace where concepts, decisions, risks, features and brand all live together, and a team of AI agents that pick-up work alongside people rather than in a separate tool.

​

Pathfinder brings together:

• The HASTE process - five stages with conditional stage gates, concept briefs and PRDs

• A unified project workspace connecting documents, decisions, risks and delivery stages

• AI agents with four autonomy levels, approval workflows, budget limits and a complete audit trail

• Feature Mapper (PVE scoring and job-to-be-done traceability), a sprint Work Board shared by people and agents, and a Brand System that turns briefs into design tokens

​