How to responsibly use AI-enabled procurement in government agencies

Government procurement teams can spend considerable time in the planning phase or a procurement, as well as drafting, revising and aligning documents. AI offers a practical way to reduce that workload without compromising quality or introducing probity risks.

AI can analyse large datasets relating to suppliers, pricing trends, contract performance and market conditions to better inform procurement teams planning a procurement. Using approved templates, policies and inputs, AI can also produce tracked, reviewable first drafts of key procurement documents, giving procurement teams a head start in tackling this important task.

AI can safely assist with drafting procurement documentation across the lifecycle, including planning, market testing and evaluation materials. For example, AI can prepare tracked first drafts of:

• a risk management plan;
• an industry engagement strategy; and
• probity plans.

Similarly, AI can assist with drafting market testing documentation, including

• the conditions of tender;
• RFT response requirements; and
• the draft contract and statement of work.

AI can help ensure consistency across documents and identify gaps that require further human input. 

Once inputs are provided, documents can be generated and updated quickly as projects progress, shifting the drafter’s role from manual rewriting to review, refinement and judgment.   

During the evaluation phase, AI can assist by summarising tender responses, mapping responses to evaluation criteria and detecting inconsistencies or missing information. This can relieve tender evaluation panel members from administrative tasks,  allowing them to focus on substantive decision making, where humans decide.

In each scenario, AI is not evaluating scoring or determining value for money; it simply produces accurate drafts for human approval. 

AI can also assist with contract review and negotiation, such as pointing out unusual risk allocations and clauses frequently contested by suppliers and contract management and performance monitoring, such as measuring delivery timelines and service levels against contractual metrics. 

Procurement documentation often reuses information across multiple documents. AI helps maintain consistent terminology, structure and referencing across the entire suite, reducing internal inconsistencies, gaps and late-stage rework.

This leads to a more coherent, defensible and auditable procurement process, supporting better outcomes for both agencies and suppliers.

Government procurement requires confidentiality and, in many cases, heightened security. That is why controlled AI environments, such as Harvey, are essential. These platforms operate within secure, organisation-approved boundaries and rely on approved templates and precedents only.

In other words, AI enhances capability without introducing unnecessary risk.

Keeping the human in charge

The principle is simple:

• AI drafts.
• Humans decide.

Evaluation, compliance, probity and value-for-money decisions remain firmly with human decision-makers. Used this way, AI becomes a productivity and quality tool, rather than a decision-maker. 

AI is already entering procurement, often informally. The challenge now is ensuring its use is deliberate, safe and value driven.

By focusing on low‑risk applications such as drafting, summarising and aligning documentation, agencies can achieve:

• significant time and cost savings

• faster procurement timeframes

• greater consistency and defensibility

• reduced administrative burden on procurement staff,

all without compromising the RFT process or reducing its defensibility. 

AI should be used only where it supports the team in making defensible value-for-money decisions and not where it replaces judgement.