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.
L1
Chat
L2
Pilot
L3
Agent
L4
Coworker
Software Dev
- ChatGPT
- Claude Chat
- Grok
- GitHub Copilot
- Windsurf← L3
- v0
- Claude Code
- Cursor
- Codex
- Lovable
- Replit Agent
- Bolt← L2
- Copilot Coding Agent
- Devin
- OpenHands
- Jules
Personal / Life
- Apple Intelligence
- Gemini app
- Perplexity
- Arc Search
- Dia browser
- Claude Cowork
- ChatGPT Agent Mode
- Gemini Agent
- Manus
- Genspark AI Employee
- OpenClaw
- Hermes Agent
Business Ops
—
- HubSpot Breeze← L1
- Salesforce Einstein
- Notion AI← L1
- Jasper← L1
- Lindy
- Gumloop
- Agentforce
- HubSpot Breeze Agents
- Clay
- Artisan← L4
- Zapier
- Glean
- 11x.ai
Frameworks
—
- LangChain
- Vercel AI SDK
- OpenAI Agents SDK
- LangGraph← L4
- CrewAI
- Microsoft Agent Framework
- Pydantic AI
- Mastra
- Google ADK
- LangGraph Platform
Voice
- OpenAI Realtime
- Gemini Live
- Hume EVI 3
- ElevenLabs
- Sesame CSM
- Suno Studio
- Udio v4
- ElevenLabs Agents
- Vapi
- Retell
- Bland
—
Robotics
- Boston Dynamics (teleop)
- NVIDIA Isaac GR00T
- Gemini Robotics (VLA)
- Physical Intelligence
- Skild Brain
- 1X Neo
- Apptronik Apollo
- Figure 03 + Helix-02
L5
Executive
not yet shipped, this row is the horizon
← L2 marks a tool that graduated from a lower level since the last refresh
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 →