Skip to content
all frameworks
06 · Original model

Dream Teamlead AI-native

Individual practice → team system → org culture
L1Individual practice
local speed ≠ delivery
Accelerates individual artifacts
Frames task and intent
Verifies AI output
Maintains deep skills
L2Team system
human-agent contract
Designs the human-agent loop
Sets autonomy by risk
Verification in Definition of Done
Context: ADR · runbooks · docs
L3Org operating model
platform + policy
Golden paths & platform
Shared policy & observability
Metrics & training
Trust & risk governance
AI-native Dream Teamlead impact levelsL1Individual practicelocal speed ≠ deliveryAccelerates individual artifactsFrames task and intentVerifies AI outputMaintains deep skillsL2Team systemhuman-agent contractDesigns the human-agent loopSets autonomy by riskVerification in Definition of DoneContext: ADR · runbooks · docsL3Org operating modelplatform + policyGolden paths & platformShared policy & observabilityMetrics & trainingTrust & risk governance

An AI-native teamlead expands impact from individual practice to a team system and then to organizational conditions. The individual layer accelerates artifacts but does not yet guarantee value delivery.

At the team layer, the lead designs the human-agent loop: risk-based autonomy, verification in the Definition of Done, and maintained context through ADRs, runbooks, and documentation.

The organizational layer supplies platforms and golden paths, policy, observability, metrics, and training. The central thesis is that a teamlead designs a system of human-agent work rather than merely spreading a toolkit.

How to use this model
01

Separate individual acceleration from team delivery: local speed alone does not make a system AI-native.

02

Design the human-agent loop with risk-based autonomy, verification in the Definition of Done, and context through ADRs, runbooks, and docs.

03

Support the team system at the organizational level with platforms, policy, observability, metrics, and training.

Sources and related materials
Author materials