Human-AI Operations FrameworkCore Framework v0.1 · 1 November 2026

A vendor-neutral operating framework

Better work.
Not more AI.

HAIOF helps organisations decide what work should exist, what humans should own, where AI should assist or execute, and when ordinary automation is the better answer.

Author: Mxolisi SilabelaPublished: 1 November 2026Working draft v0.1

The operating question

Organisations are adopting AI quickly, but many still lack a practical model for sharing work, authority, knowledge and evidence across people, AI and automation.

01

What work should exist?

Redesign before automating. Remove work that no longer serves a useful outcome.

02

Who should perform it?

Choose human-led, AI-assisted, AI-executed or deterministic automation deliberately.

03

How does control remain clear?

Separate recommendation, execution and approval. Define evidence, boundaries and stop conditions.

Trustworthy progress

Useful work moves forward while material risk remains controlled, human authority stays clear and organisational knowledge remains durable.

Five work classes

Before deciding who or what performs the work, first ask whether the work should exist at all.

01Eliminate
02Human-led
03AI-assisted
04AI-executed
05Deterministic automation

From principle to practice

The Core Framework defines the operating model. The Practitioner Guide turns it into assessments, workshops, records, pilots and transformation steps.

CORE

Understand the model

Vision, principles, maturity, work classification, responsibility, knowledge, governance, metrics and transformation.

GUIDE

Apply it to real work

Assess reality, redesign workflows, define boundaries, evaluate outcomes and scale deliberately.

TOOLS

Leave durable evidence

Use practical records for intent, decisions, knowledge, governance, evaluation, incidents and change.