Call 1300 950 251     Follow us :

DTA AI Procurement Guidance Explained

DTA AI Procurement Guidance Explained

DTA AI Procurement Guidance Explained

Thursday, 1 October 2026

Key takeaways

  • The Digital Transformation Agency (DTA) released Guidance on AI Procurement in Government in December 2025 alongside the updated Policy for Responsible Use of AI v2.0 and the AI Impact Assessment Tool
  • The guidance maps AI considerations across three phases of the Digital Sourcing Lifecycle: Plan, Source and Manage
  • AI Model Clauses v2.0 (published March 2025) provide contract terms covering transparency, audit rights, data governance, intellectual property and ethical use
  • Agencies must implement mandatory AI impact assessments for all in scope use cases by 15 December 2026
  • Procurement teams are the front line for implementing this guidance because they write the specifications, evaluate the bids and manage the contracts through which AI enters government

The Digital Transformation Agency (DTA) released Guidance on AI Procurement in Government in December 2025 as part of a broader AI governance package that included the updated Policy for Responsible Use of AI in Government (v2.0) and a new AI Impact Assessment Tool. The DTA’s announcement described the procurement guidance as equipping agencies with the practical tools required to buy responsibly and deliver outcomes Australians can trust. For procurement teams, this is not supplementary reading. It is the operational guidance that determines how AI products and services are specified, evaluated, contracted and managed across the Australian Public Service (APS).

What does the guidance cover?

The procurement guidance maps AI considerations across three phases of the Digital Sourcing Lifecycle. In the Plan phase, agencies must define the problem the AI solution is intended to address, assess whether AI is the appropriate technology, identify the data requirements and conduct a preliminary risk assessment. The Lexology analysis of the DTA AI Model Clauses notes that the Model Clauses align with Australia’s technology regulations in privacy, cybersecurity and intellectual property, covering transparency, audit rights, data handling, model updates and exit provisions.

In the Source phase, the guidance recommends assembling multidisciplinary procurement teams that include technical, legal, privacy and domain expertise alongside procurement professionals. Tender documents should specify how AI will be used, what data it will process, what transparency and explainability requirements apply and what ongoing obligations the supplier will carry. In the Manage phase, contract management must include monitoring the AI system’s performance, managing updates and model changes, maintaining audit trails and ensuring ongoing compliance with the Policy requirements.

What are the AI Model Clauses?

The AI Risk Aware analysis of the DTA Model Clauses describes them as setting contractual norms that will influence private sector AI procurement practice across Australia. The Model Clauses v2.0 (published 17 March 2025) create obligations governing the relationship between the buyer and the seller of AI products. They cover AI system transparency (how the system works and what data it uses), audit rights (the buyer’s ability to inspect the system), data governance (ownership, handling and retention of data processed by the AI), intellectual property (who owns outputs generated by the system), ethical use (alignment with Australia’s AI Ethics Principles) and exit provisions (what happens when the contract ends, including data return and model decommissioning).

The Australian Government Solicitor (AGS) consulted on the development of the clauses and recommends that buyers consider the broader legislative context before drafting: privacy obligations, cybersecurity requirements, PSPF Directions (including the prohibition on DeepSeek and Kaspersky products) and existing whole of government arrangements.

What must agencies do by December 2026?

The Policy for Responsible Use of AI v2.0 took effect on 15 December 2025, with a structured implementation timeline. The Trusenta analysis of the DTA AI Policy confirms that agencies must implement mandatory AI impact assessments for all in scope use cases by 15 December 2026. This means every AI system that processes personal information, makes decisions affecting individuals or operates in a context that carries significant risk must be assessed using the DTA’s AI Impact Assessment Tool and the results documented in the entity’s governance framework.

For procurement teams, the practical implication is that AI procurements initiated between now and December 2026 must be designed to satisfy both the procurement guidance and the policy requirements. Specifications must include impact assessment requirements. Evaluation criteria must assess the supplier’s ability to meet transparency, data governance and ethical use obligations. Contracts must include the Model Clauses or equivalent protections.

How can procurement professionals prepare?

AcademyGlobal (AG), the Chartered Institute of Procurement & Supply (CIPS) Centre of Excellence for Australia, New Zealand and Southeast Asia, delivers programs that build the capability procurement teams need to implement AI procurement guidance effectively. AG’s Digital as Procurement Enabler program addresses how digital technologies, including AI, are changing the procurement function and how procurement professionals can evaluate, procure and govern technology solutions responsibly. For procurement leaders responsible for embedding AI governance into their team’s practice, AG’s Strategic Procurement Leadership program builds the strategic capability to lead procurement through regulatory change and technological disruption.

The DTA guidance does not ask procurement teams to become AI experts. It asks them to apply procurement discipline to AI: define the requirement clearly, evaluate suppliers rigorously, contract with appropriate protections and manage the arrangement through its lifecycle. These are procurement skills applied to a new category.

Frequently asked questions

What is the DTA AI Procurement Guidance?
Guidance released in December 2025 that maps AI considerations across the Plan, Source and Manage phases of the Digital Sourcing Lifecycle. It equips agencies with practical tools for buying AI responsibly.
What are the AI Model Clauses?
Contract terms published by the DTA in March 2025 (v2.0) covering transparency, audit rights, data governance, intellectual property, ethical use and exit provisions for AI procurements.
When must agencies complete AI impact assessments?
By 15 December 2026 for all in scope use cases. The Policy for Responsible Use of AI v2.0 took effect on 15 December 2025 with a one year implementation window.
Do agencies need multidisciplinary teams for AI procurement?
Yes. The guidance recommends including technical, legal, privacy and domain expertise alongside procurement professionals. AI procurement involves risks that procurement teams cannot assess alone.
Where can I build AI procurement capability?
AcademyGlobal’s Digital as Procurement Enabler program addresses how to evaluate, procure and govern technology solutions including AI in government settings.

References

  1. DTA, “AI Policy Overhauled with New Impact Assessment Tool and Procurement Guidance,” December 2025
  2. Lexology / Hogan Lovells, “Australia’s Model Clauses Provide Framework for AI Procurement,” June 2025
  3. AI Risk Aware, “Australian AI Procurement: The DTA Model Clauses and What They Mean,” June 2026
  4. AGS, “Legal Update No. 327: AI Model Clauses,” April 2025
  5. Trusenta, “DTA AI Policy: Mandatory Register and Supplier Obligations,” July 2026