AI governance for New Zealand government agencies, from ministries to councils.
We work with central government departments, Crown entities, state-owned enterprises, and local councils. Every engagement is scoped against the Public Service AI Framework, the Algorithm Charter, Government Procurement Rules, and Te Tiriti o Waitangi, and closes with an evidence pack an Auditor-General review can work through.
Built for
What your agency walks away with.
Full engagement methodologyFramework implementation plan
A phased roadmap mapped to your agency's AI maturity, with gap analysis against the Public Service AI Framework and the Algorithm Charter.
AI procurement evaluation criteria
Supplier scorecards, due-diligence checklists, AI-specific contract clauses, and exit-strategy templates aligned with Government Procurement Rules.
Treaty-aligned governance
Iwi and hapū engagement frameworks, cultural impact assessment templates grounded in te ao Māori, and kaitiakitanga embedded in policy.
Audit-ready evidence pack
Documentation Te Kawa Mataaho Public Service Commission, the Office of the Privacy Commissioner, or an Office of the Auditor-General review can work through.
Where government AI programmes stall.
Four pressure points recur across the agencies we work with. The obligations behind them are set out in full on our Public Service AI Framework explainer. This page is about the engagement that resolves them.
- 01Procurement gap
AI procurement without AI-specific criteria.
Government Procurement Rules require assessment of supplier reputation, pricing, supply chain risks, and privacy and security posture. The Responsible AI Guidance for GenAI adds expectations around data traceability and exit strategies. The Algorithm Charter commits agencies to transparency. Without AI-specific evaluation criteria aligned to these commitments, procurement teams are guessing.
Build procurement evaluation criteria - 02Risk gap
Risk registers not designed for algorithmic harm.
The Public Service AI Framework calls for proportionate risk management. Most agency risk registers were not designed for algorithmic bias, model drift, or third-party AI dependencies. AI-specific risk categories need to be layered onto existing processes, aligned with OECD AI Principles on robustness, security, and accountability.
Build an AI risk management framework - 03Constitutional obligation
Te Tiriti without AI-specific guidance.
Te Tiriti o Waitangi requires government agencies to protect Māori interests. When AI systems process Māori data, affect Māori communities, or inform decisions with Treaty implications, generic frameworks fall short. The principle of kaitiakitanga demands that Māori data sovereignty be built in from the start, not bolted on.
Integrate Treaty-compliant governance - 04Scrutiny gap
Public scrutiny arrives before the governance does.
An Official Information Act request, a select committee question, or an Office of the Auditor-General review can land months before an agency has an algorithm register, a public-facing explanation, or a documented human-oversight pathway. The agencies that answer well are the ones that built the evidence pack first.
See the obligations the evidence pack has to answer
The PolyGovern tracks that apply to government.
Three practice tracks built around the Public Service AI Framework, the Algorithm Charter, and Te Tiriti obligations. Each engagement closes with a documented evidence pack.
Track A
Governance and strategy
Public Service AI Framework operating models, Algorithm Charter implementation, and AI policy tied to Cabinet expectations and Te Kawa Mataaho Public Service Commission scrutiny.
Track B
Assessment and assurance
Risk assessment, AI impact analysis, and supplier due diligence aligned with Government Procurement Rules and the Responsible AI Guidance for GenAI.
Track C
Compliance and advisory
Privacy Act 2020 alignment, Māori data governance practice grounded in kaitiakitanga, AI literacy across the workforce, and standing advisory support for cross-agency programmes.
Across every level of NZ government.
Each level of government carries different governance pressure. Central departments answer to Cabinet and Te Kawa Mataaho. Crown entities operate inside their own statutes. Councils answer to ratepayers and mana whenua.
Central government departments
Ministries and departments with direct Cabinet accountability. Public Service AI Framework, Algorithm Charter commitments, and all-of-government procurement obligations all apply.
Local government and councils
Auckland Council, Wellington City Council, and regional councils deploying AI for resource consenting, transport planning, and citizen services, with accountability to ratepayers and mana whenua.
Crown entities and SOEs
Te Whatu Ora (Health New Zealand), education institutions, and state-owned enterprises building AI governance inside their sector-specific statutes, from the Health Information Privacy Code to education standards.
Common questions from government agencies.
Which parts of government do you work with?
Central government departments and ministries, Crown agents and autonomous Crown entities, state-owned enterprises, and territorial and regional councils. The instruments differ by tier: departments answer to Cabinet and Te Kawa Mataaho, Crown entities to their own statutes, councils to ratepayers and mana whenua. The engagement tracks on this page are scoped to the tier you sit in.
How is a government engagement different from a private-sector one?
Three things change. Procurement runs under the Government Procurement Rules, so evaluation criteria and exit terms have to be defensible before contract signing. Records fall under the Public Records Act and the Official Information Act, so every assessment is written on the assumption it will be released. And Te Tiriti obligations are statutory for public service agencies under the Public Service Act 2020, which puts iwi and hapū engagement inside the governance design. The full list of obligations is on the Public Service AI Framework explainer.
What are our Te Tiriti o Waitangi obligations for AI?
When AI processes data about Māori, informs decisions affecting Māori communities, or is deployed in contexts with Treaty implications, agencies need to ensure Māori data sovereignty and kaitiakitanga are respected through meaningful partnership with iwi and hapū, not consultation as a checkbox.
How do we coordinate AI governance across multiple agencies?
Cross-agency AI projects, shared platforms, and all-of-government procurement arrangements need shared risk-assessment standards, data-sharing protocols, vendor-management responsibilities, and clear accountability for outcomes. Without explicit coordination, each agency builds its own approach.
What should an AI exit strategy look like?
The Public Service AI Framework and Government Procurement Rules both emphasise exit planning. You need to be able to retrieve your data, switch vendors without service disruption, and retain ownership of fine-tuned models or training data. Exit strategy is defined before contract signing.
Further reading
AI policy template
The New Zealand edition carries Algorithm Charter commitments, OIA clauses and a Te Tiriti and Māori data section.
Get the templateAI governance maturity model
Five levels an agency can place itself on, with the Public Service AI Framework pillars each one satisfies.
Read the guideThe PolyGovern AI Governance Framework
The eleven elements, mapped to the Framework’s principles and pillars and to the Algorithm Charter.
Read the guideScope the engagement your agency needs.
A scoping call maps your AI inventory, procurement pipeline, and Te Tiriti obligations to one of the three tracks above, and ends with a written proposal your chief executive can take to the leadership team.