The landscape
AI products by autonomy level.
Where each tracked product sits on the 5-level ladder, from Chat to Executive, across software dev, personal/life, business ops, and frameworks.
The full framework with the PM-trust lens, jumps that matter, and decision table lives in the agentic AI landscape essay.
Quick summary
Level
Role
What it is
L1
Chat
AI = Responder
Human = Instructor
Answers when asked. No tools, no action. Reactive: you prompt, it replies, it stops.
L2
Pilot
AI = Assistant
Human = Editor
Same shape as L1, but reaches for tools mid-answer. Scope stays narrow. One question in, one answer out.
L3
Agent
AI = Teammate
Human = Reviewer
Plans multi-step work. Executes in a loop. Self-corrects. You review.
L4
Coworker
AI = Deputy
Human = Manager
Persistent. Runs 24/7 on its own hardware. Triggers its own work. Reports back.
L5
Executive
AI = Business Unit
Human = Owner / Board Member
Runs a mission. Spawns its own sub-agents. Reports by exception only.
Where each product sits
Levels × categories. Reference only, no clicks.
Software Dev
Personal / Life
Business Ops
Frameworks
L1
Chat
- ChatGPT
- Claude Chat
- Grok
- Siri
- Gemini
- Notion AI
- Jasper
none
L2
Pilot
- Copilot
- Bolt
- v0
- Amazon Q
- Perplexity
- HubSpot AI
- Salesforce Einstein
- LangChain
L3
Agent
- Claude Code
- Cursor
- Windsurf
- Lovable
- Replit Agent
- Copilot Workspace
- Claude Cowork
- OpenAI Operator
- Jules / Mariner
- Lindy
- Gumloop
- Agentforce
- CrewAI
- AG2
- MS Agent Framework
L4
Coworker
- Devin
- OpenHands
- OpenClaw
- Hermes Agent
- Zapier
- Glean
- Artisan
- LangGraph
L5
Executive
not yet shipped in 2026.04
Want the framework, not just the map?
Read the full 5-level framework essay.
Each level with the capability lens, the PM-trust lens, jumps that matter, decision table, and a field guide for picking the right level for the task.
Read the essay on janeyoubradley.com →