Frameworks

Original frameworks and models for designing, building, and governing enterprise AI agents

Agentic Academy’s original frameworks provide a shared vocabulary for designing, building, and governing AI agents in enterprise environments. Each framework is grounded in production experience and designed to be immediately applicable.

These frameworks are free to use, cite, and build upon. If they help clarify your thinking or improve your team’s conversations, we’ve done our job.

Core Framework

Agentic Primitives

The building blocks of AI agent systems

A taxonomy of 17 primitives across four domains—Mind, Hands, Voice, and Wiring—that provides a shared vocabulary for designing, building, and governing AI agents regardless of framework or platform.

Mind

Actors & Instructions

Hands

Tools & Retrieval

Voice

Interactions & Patterns

Wiring

Connections & Orchestration
Governance Framework

Four Dimensions of Agent Autonomy

A multi-dimensional model for understanding and governing agent independence

Autonomy isn't a single dial you turn. This framework decomposes agent autonomy into four independent dimensions—Tool, Task, Plan, and Collaboration—enabling precise governance conversations and incremental trust-building.

Tool Autonomy

What capabilities an agent can use

Task Autonomy

What work an agent can identify

Plan Autonomy

How an agent sequences actions

Collaboration

How agents coordinate with peers
Operational Framework

Applied Autonomy Framework

Bridging design intent and operational reality

The gap between what you design and what you can operate is where most enterprise AI initiatives stumble. This framework makes the gap between aspiration and operational reality explicit, measurable, and actionable—with maturity gates, trust scores, and risk assessments.

Design Profile

Intended autonomy levels

Operating Profile

Actual deployed autonomy

Maturity Gates

Conditions for expanding autonomy

Trust Scores

Earned confidence over time
Architecture Model

Tasks, Skills & Tools

The holy trinity of AI agent capability architecture

A practical framework for understanding the three essential components that make AI agents work in enterprise environments: the work they do (Tasks), the capabilities they have (Skills), and the systems they interact with (Tools).

Tasks

The work agents perform

Skills

Capabilities agents possess

Tools

Systems agents interact with

Using these frameworks

These frameworks are published under Agentic Academy and are free to reference in your presentations, internal documentation, architecture reviews, and published work. When citing, please attribute to the original authors and link back to the source article.