Workato partners with Indigenous supplier to offer 'sovereign AI' to Australian businesses and government

Sydney, October 14, 2026 , Enterprise automation vendor Workato today unveiled a commercial agreement with Gulanga Group, a Supply Nation certified Indigenous ICT provider, and Trellis Data to deliver so called sovereign AI to Australian government agencies and private sector customers. The companies said early testing is already under way with Gulanga’s Dukes model, powered by Trellis Data’s Agentaus large language model, and that integration work will allow customers to run Australian trained models inside Workato workflows while keeping data and model processing onshore.

What the deal does

Workato will assess and integrate Australian model providers recommended by Gulanga, then expose those models as selectable options inside the Workato platform. That means an organisation using Workato to automate processes, generate documents, or support casework will be able to route prompts to a locally trained model such as Dukes rather than automatically sending them to an offshore frontier model. The partners emphasised that Workato will remain model neutral, so sovereign and non sovereign models can sit side by side depending on customer needs.

Under the arrangement, Gulanga will perform market assessment and help with procurement and governance fit, while Trellis Data supplies the underlying Agentaus technology used by Dukes. Workato said the integration focuses on governed execution, so any model chosen operates inside the platform’s existing controls for access, logging and approval.

Why sovereign AI matters for Australia

Australian federal and state agencies increasingly face procurement requirements and compliance obligations that favour local hosting, local data handling, or demonstrable governance over algorithmic decisions. For organisations with particularly sensitive customer or patient data, the ability to run model inference onshore and to rely on an Australian data set for fine tuning reduces legal and reputational risk.

The partnership seeks to answer that demand with a commercially oriented approach. Workato executives framed sovereign AI not as a constraint but as a commercial option that can deliver cost, latency and relevance benefits for domain specific tasks such as policy drafting, regulatory compliance checks or specialised customer service workflows.

Commercial and social angles

The deal also carries an Indigenous economic intent. Gulanga is Supply Nation certified and said revenue from its Dukes service includes token economics designed to fund Indigenous social outcomes. That structure ties model usage to community returns and positions Gulanga as an intermediary that can both vouch for local capability and channel economic benefit into Indigenous programs.

For Workato, the commercial upside includes expanding addressable market for its automation platform in the public sector and regulated industries, where procurement officers and chief information officers have been cautious about using offshore models. For Trellis Data, close product integration with a major automation provider offers a route to production scale usage and steady inference demand.

Market reaction and policy context

Industry commentators and early adopter customers quoted by the partners welcomed the move. One public interest technology buyer highlighted the appeal of a model trained on Australian law and policy for government drafting tasks, which the partners say improves relevance and reduces the need for extensive human correction. Vendors and customers often cite such domain specialisation as a primary advantage of locally trained models.

The announcement comes as Australia debates regulatory frameworks for AI, including governance expectations for public sector procurement and data residency. Recent parliamentary committee hearings have underscored the need for platforms and vendors to make local model options available while preserving interoperability with international providers. Workato said it spoke to those debates at a parliamentary committee session last week when it advocated for neutral platforms that enable competition between local and frontier models.

What to watch next

Key near term milestones to monitor include completion of testing and the first commercial deployments inside Australian agencies or regulated enterprises, timing for which the partners have not specified. Procurement cycles in government can be lengthy, so early pilots or proof of concept projects will be the first indicators of traction.

Investors and corporate buyers will be watching whether the technical integration lives up to claims around latency and cost, and whether the token economics model yields meaningful social outcomes without creating new governance complexity. For other model providers in Australia, Workato’s platform integrations could create demand for specialised models, but they also set a higher bar for formal accreditation and proof of governance.

Why this matters for Australian business and finance

From a business and finance perspective, the deal signals a maturing market in which Australian suppliers can capture higher value activity beyond advisory and pilot work. If onshore models become more widely available inside enterprise execution layers, that could shift some AI spend from hyperscaler hosted services to locally provided inference and managed hosting. Over time that shift could create recurring revenue streams for Australian model operators, systems integrators and hosting providers, while also shaping how capital is allocated across domestic tech firms.

For the public sector, sovereign AI options offer a pragmatic compromise between adopting powerful AI capabilities and meeting statutory duties around data protection and procurement. The commercialisation of those options, combined with explicit Indigenous participation, also reframes part of the sovereign debate around opportunity rather than only risk.

Workato, Gulanga and Trellis Data said they will announce pilot customers and more technical details in coming months. For now the partnership provides a commercially oriented test case of how sovereign AI could be delivered at enterprise scale in Australia.