> ## Content Index
> Fetch the complete content index at: https://policyops.ai/llms.txt
> Use this file to discover other available public pages before exploring further.

# Five AI questions for government agency leaders
- URL: https://policyops.ai/five-ai-questions-for-government-agencies/
- Published: 2025-03-19T01:25:14.000Z
- Updated: 2026-01-09T03:41:26.000Z
- Description: Government agency leaders are navigating a new age. Somewhere between all of the hype and the fear lies significant opportunity. What might they need to consider right now?
- Author: Stuart MacKinnon

Government rightly doesn't want to be left behind on AI - the opportunities are real. But adopting AI for its own sake could lead to problems too. As an agency leader, here are a few questions I'd be asking right now.

🔎 **How are we already using AI?**

From AI embedded in productivity apps, to larger investments, building a comprehensive picture of AI's current footprint in your organisation is always a good starting point.

🔎 **Do we really need AI for this?**

The [Centre for Data Ethics and Innovation](https://data.govt.nz/leadership/centre-for-data-ethics-and-innovation/guidance/artificial-intelligence-guidance) recommends asking: "*Is AI really the solution? Would another tested technology be just as good for your problem?*" This is solid advice. There's no point adding unnecessary complications.

🔎 **Do we need to be doing this at all?**

In all the excitement, proposals and initiatives can take on a life of their own. Are AI projects aligned with organisational strategy? If it's exploratory, or on the leading-edge, it needs to be tightly ring-fenced.

🔎 **How will we know AI made a real difference?**

Are we clear on the public value being delivered, or are we relying on ROI white papers and vendor 'research' to justify this? "Studies show..." is not a benefit realisation plan. How will we define and measure value, right here, on the ground?

🔎 **What does AI trust and transparency look like?**

Public interest is likely to be on a spectrum, with lower interest in how AI tools improve daily work, and far higher interest in AI that impacts the public personally. Check out the draft [Public Service AI Trust Model](https://policyops.ai/public-service-ai-trust-model-v1/) I posted up for discussion if this is something you are also thinking about.

---

As AI conversations shift from a sense of novelty to a sense of necessity, these questions may separate agencies that thrive in the new age from those that struggle. What would you add to the list?

---