Choose a scenario
Illustrative scenarios, modeled on real government uses. Each one runs through the same five levels; the resident 311 request is the default.
One resident, three requests at once
Before we climb the spectrum, here is where the work starts.
City service request form
Tap any node to see what it does, in plain language and in n8n. The gold marker is one request making the trip.
Which level fits your organization?
There is no single right level. It depends on your task, your data, and your tolerance for risk. A few questions to take back to your office:
Michael's rule from the first half, “treat AI as a structured drafting engine, not a knowledge source,” was about a chatbot. Does it still hold at Level 4?
- For a given task, what is the lowest level that would do the job? What would justify climbing higher?
- When a vendor's material says “agentic” or “AI-powered,” what does the tool actually do on its own? Which level would you place it at?
- For each use, where should the human stay? Is the AI referencing your own data, and do you know which sources?
- Where do your current and planned AI uses sit on this spectrum?
- What are the benefits and the limitations at the level you are weighing, and who carries the risk if it is wrong?
The level frame has a long research history: in 2000, Parasuraman, Sheridan, and Wickens described automation as a continuum of levels, from fully manual to fully automatic, and asked which functions should be automated, and to what extent.
The choice of level is also a central question in AI ethics research: a decide-to-delegate model, in which people retain the power to decide which decisions to take, and any delegation stays overridable. Across 84 guidelines, the field converged on five shared principles; researchers also caution that principles alone cannot guarantee ethical AI.
In a July 2026 review, we ran four public policies, as published in 2024 and 2025, through the policy self-check: Arizona's statewide generative AI policy (since revised, February 2026), San Francisco's generative AI guidelines, Seattle's AI policy, and Los Angeles's AI ethics code and safety checklist. All four cover the chatbot-era basics, data rules and disclosure among them. None clearly addresses any of the five agent-specific items, a gap of timing: the editions reviewed predate the agentic wave. Arizona’s February 2026 revision, checked after this review, likewise contains no agent-specific provisions. In March 2026 the National Association of State Chief Information Officers (NASCIO) published a report on agentic AI in state government recommending that state AI policies be “updated to include language around the risks and rewards of agentic AI” and noting that eight states already report agentic tools in production.
Where is the human?
Human oversight is its own spectrum. For decisions affecting rights or safety, the Office of Management and Budget (OMB) requires human oversight, intervention, and accountability: a person in or on the loop.
The system stops and waits for a person to approve before it acts. An approval gate.
Example: an AI flags a possible tumor; a radiologist makes the final call.
Fits when the decision is high-impact or high-stakes. Levels 1 to 4 live here.
The AI runs on its own; a person monitors and can override or hit the kill switch.
Example: supervised driving, or an escalation queue that flags edge cases for review.
Fits when it is routine work at scale, with anomalies escalated. Level 5 and scaled automation.
Full autonomy, no human. The AI senses, decides, and acts alone.
Example: high-frequency trading, too fast for a person to intervene.
For government: not appropriate for anything affecting rights or safety.