Framework

HELM

Human-first Execution and Leadership Model

Practical guidance for working with AI agents. Choose what to delegate, check the results, and agree which decisions people own.

HELM 1.0.1 · Updated 2026-09-17

Practitioner Guide Leadership Guide

Principles

Six principles for working with agents

Read all

Simplicity First

Start with the simplest solution possible. Only increase complexity when it demonstrably improves outcomes.

Redesign, Don't Automate

Rethink the work before adding AI. Simplify steps and focus on the result you need.

Agents Execute, Humans Are Accountable

Agents can help with tasks. People remain responsible for decisions and results.

Guardrails Are Non-Negotiable

Set limits and checks before giving agents more work or freedom.

Structure Over Tooling

Agree who owns the work, who checks it, and who decides. Tools cannot make those choices for you.

Team-Wide Adoption Over Individual Mastery

Help the whole team learn how to use and review agent work, rather than relying on a few experts.

Read all principles

For Practitioners

How we work with agents

Choose which tasks to give agents
Set clear limits before work starts
Check results before they are shared
Keep people responsible for important decisions
Learn from each piece of work

Read the guide

For Leaders

How we lead and support the team

Agree who owns goals, decisions, and review
Introduce new ways of working in small steps
Support people as their responsibilities change
Review quality, outcomes, and effort together
Spot common problems before they spread

Read the guide

Roles & Skills

How responsibilities and skills change

Explore what AI-assisted work means for each role, what people remain responsible for, and which skills help the team work well.

Software EngineerStaff / Principal EngineerEngineering ManagerSRE / DevOps EngineerQA Engineer / SDETPlatform / Infrastructure EngineerProduct ManagerProduct Designer8 roles

Explore roles

Latest updates

View all updates

2026-09-17HELM 1.0.1Published

Keeping HELM simple

HELM 1.0.1 is easier to read and use, with simpler principles, practical guides, shared skills, and a focused roadmap.

See what changed

2026-09-05HELM 1.0.0Published

A shared starting point for human–agent teamwork

The first numbered release brings HELM’s principles, practical guides, and role descriptions together, so your team can refer to the same version when discussing how to work with AI agents.

See what changed