A useful, bounded starting point
Identify work where AI can help, with a defined scope, agreed sources and an understandable measure of progress.

Public value, practical delivery
Make AI useful within the environment you are responsible for.
Start with public value, the work people need to do and the policy, security and operational constraints that shape delivery.
Discuss your delivery contextWith Nicolai · GovernmentPublic-sector leaders need a clear purpose for change, appropriate information handling and a workable route through the organisation. We begin with the actual environment: what is permitted, who is accountable and what would make the service or team more effective.
Identify work where AI can help, with a defined scope, agreed sources and an understandable measure of progress.
Bring approved information into briefs that retain links to evidence and make gaps or uncertainties clear.
Build understanding alongside practical capability, involving the teams responsible for policy, security and continued operation.
My current engineering work includes auditable workflows in UKOPRA’s live platform, research into AI priorities in Third Layer and knowledge systems with source references and human confirmation. My public-sector experience comes from Inland Revenue New Zealand, where I worked on transformation strategy and implementation for the government’s Telecommunications as a Service programme. My work included partnerships, programme teams, commercial risk and an Innovation Lab. During that time, I also worked with the Ministry of Business, Innovation and Employment, the Department of Internal Affairs and the Ministry of Social Development. At Spark, I developed enterprise networking propositions for government and financial services, with earlier mobility work including NZ Police, Housing New Zealand and New Zealand’s security services.
About NicolaiAn illustrative application: use approved, appropriately classified documents to prepare an internal briefing. Keep source references visible, flag missing information and require a responsible person to review the result. Access, hosting and information-handling requirements are agreed before implementation.
A possible starting point, not a completed-client case study. Scope is agreed together.Approved sources and clear boundaries
Referenced findings and visible gaps
Human review before use
Start with the leader’s operating environment and a bounded, useful scope. Over six months, develop and review the work with the relevant teams, followed by a clear briefing and handover. Engagement and delivery arrangements must fit your organisation’s processes.
Explore Transform · US$11,500 + travelTell me a little about your situation in the short brief below. It takes about three minutes and only I read it. If it looks like a fit, you’ll get a link to book a free half-hour with me. If it doesn’t, I’ll say so plainly and point you somewhere more useful.
Share your brief first so I have the context before we speak. You’ll then be able to choose a time on Cal.com. A free half-hour. No card details, no obligation.
Information access, classification, approved tools and hosting requirements must be agreed with your organisation before any data is used. No assumption is made that a public AI service is appropriate.
No certification is included. The work is shaped around your organisation’s requirements and responsible teams; approvals remain with the relevant authorities.
Yes. The Diagnostic is half a day on an agreed example using appropriate information, and it stands on its own. Transform and Implement are there when the organisation is ready for a broader engagement.
I work with a handful of leaders at a time. I don’t name clients unless they ask me to. I’ll sign your NDA before I see anything, or you can use mine, and we agree where your information lives before we start.
If you’d rather your team didn’t know you were exploring this yet, that’s how most conversations begin.