Most AI roadmaps teach capability without control. They show you how to prompt, fine-tune, and deploy, but not how to stay in charge of what you build. This roadmap teaches both. You will gain AI skills and the habits that keep those skills under your authority.

The difference matters. Organizations adopting AI face a choice: let capability outpace governance, or build governance into capability from day one. This roadmap assumes you choose the second path.

Detailed Course Catalog Here

Before You Begin: The Governance Foundation

Before selecting a track, you will establish the operating system that runs beneath every phase. This takes approximately two hours. The investment pays dividends throughout your learning and career.

Three Core Ideas Structure Your Work:

1. Role-Based Collaboration (HAIA-RECCLIN)

HAIA-RECCLIN is a multi-AI governance framework that assigns explicit roles to AI systems under human constitutional authority. The acronym stands for Human Artificial Intelligence Assistant, with seven specialized roles: Researcher, Editor, Coder, Calculator, Liaison, Ideator, and Navigator. Each role carries defined responsibilities:

• Researcher: Finds sources, verifies facts, gathers evidence, flags contradictions
• Editor: Refines structure, enforces consistency, adapts to audience, catches errors
• Coder: Writes, reviews, and debugs code with documented assumptions
• Calculator: Performs mathematical analysis, quantitative modeling, data processing
• Liaison: Coordinates perspectives, manages stakeholder communication, bridges disciplines
• Ideator: Generates creative options, brainstorms approaches, surfaces novel possibilities
• Navigator: Documents disagreements, presents trade-offs, resists picking sides prematurely

The framework emerged from 16 years of systematic development, beginning with the original 2012 Factics methodology. Its core principle: AI works for your judgment, not the reverse.

Full Framework: HAIA-RECCLIN: The Multi-AI Governance Framework

Quick Start Version: HAIA-RECCLIN Lite: 30-Day Implementation Guide

2. Checkpoint-Based Governance (CBG)

Checkpoint-Based Governance establishes three phases of human oversight that apply to every AI-assisted task:

Before: Define purpose, constraints, and success criteria before AI begins work. Write one paragraph with three specific requirements. This prevents scope drift and establishes what “done” looks like.

During: Pause mid-task to ask “What is the counter-argument?” Review AI work at natural breakpoints. Adjust intensity based on stakes: single review for low risk, multiple checkpoints with independent reviewers for high-consequence decisions.

After: Answer three questions before accepting output. Can I explain this decision to a colleague? Do I know the sources? Would I bet my reputation on this? Record your acceptance or rejection with rationale.

Checkpoint records must be immutable. Oversight evidence cannot be rewritten after the fact. This creates audit trails that prove governance happened, not merely that governance was intended.

Full Framework: Checkpoint-Based Governance

3. Personalization Prompts

Personalization prompts force AI systems to surface doubt, alternatives, and limitations. They transform passive consumption into active verification. Examples:

• “Review as critical editor. List errors and missing perspectives.”
• “Provide an alternative conclusion with supporting evidence.”
• “Identify the parts of this response most likely to be wrong or outdated.”
• “If I were to take the opposite view, what evidence would support me?”
• “What assumptions are you making that I should verify independently?”

These prompts work with any AI system. They establish the habit of reading AI output as draft material requiring human judgment, not finished product requiring human acceptance.

Your First Deliverable

Before entering either track, complete this exercise:

1. Create your personalization prompt set (3-5 prompts tailored to your work context)
2. Use these prompts on one real task this week
3. Document what the prompts revealed that you would have missed without them

This takes under one hour. It establishes the governance habit before technical learning begins. Keep your prompt set and refine it throughout your journey.

Two Tracks, One Foundation

Both tracks share the governance foundation. They diverge in application. Choose based on your role and goals.

TraTrack 1: Leadership Path

You are building the capability to direct AI, evaluate AI outputs, and set AI policy, not just use AI tools.

Time Commitment: 1-2 hours per week | Horizon: 12-18 months | Total Cost: $600-900

This track builds credibility through recognized credentials while developing the judgment to lead AI initiatives. You will understand enough technical foundation to ask the right questions without becoming an engineer.

Phase 1: Language and Credibility (Months 1-4)

Goal: Establish AI literacy and earn credentials that signal competence to stakeholders.

Begin with free courses to build foundational knowledge, then pursue certifications that demonstrate competence to employers and colleagues.

Governance Integration: Practice CBG checkpoints on each course module. Use your personalization prompts to summarize each module and compare to your notes.

Phase 1 Outcome: Two to three credentials. Established checkpoint thinking habit. Evidence of critical reading through documented prompt exercises.

Phase 2: Applied Literacy and Governed Practice (Months 5-10)

Goal: Develop technical intuition and practice multi-AI collaboration under governance.

These courses build programming literacy and machine learning intuition. You do not need to become an expert engineer, but you need enough depth to evaluate AI systems and ask informed questions.

Governance Integration: Role-based exercise with two AI systems (Researcher and Editor roles). CBG documentation for one course project. Personalization prompts on all assignments.

Phase 2 Outcome: Programming literacy. ML intuition. One documented example of governed multi-AI practice.

Phase 3: Enterprise Authority (Months 11-18)

Goal: Earn enterprise certification and produce governance artifacts demonstrating leadership capability.

This phase culminates in credentials that prove enterprise readiness and a capstone project that demonstrates your ability to govern AI implementations.

Capstone: Governed AI Implementation Plan with six components:

1. Use Case Definition
2. Role Assignment (HAIA-RECCLIN)
3. Checkpoint Map (CBG)
4. Personalization Protocol
5. Dissent Preservation
6. AI Evaluation

Phase 3 Outcome: Enterprise certification. Governance artifact proving leadership capability.y.

Track 2: Engineering Path

You are building the capability to architect governance into AI systems, not bolt it on afterward.

Time Commitment: 4-6 hours per week | Horizon: 18-24 months | Total Cost: $900-1,100

This track builds technical depth while embedding governance thinking into every layer of development. Your GitHub will not just show code. It will show code with accountability built in.

Phase 1: Foundations with Governance Thinking (Months 1-6)Phase 1: Foundations with Governance Thinking (Months 1-6)

Goal: Establish programming fluency and mathematical foundation while developing the habit of seeing governance requirements in technical work.

Phase 1 Outcome: Fluent Python. Applied mathematical foundation. Habit of seeing governance requirements in technical work.

Phase 2: Machine Learning with Built-In Accountability (Months 7-12)

Goal: Build working ML systems with governance documentation that makes them auditable.

Phase 2 Outcome: Working ML projects with governance documentation. Models that are auditable, not just accurate.

Phase 3: Advanced AI with Multi-System Dissent (Months 13-18)

Goal: Master advanced AI architectures while implementing multi-AI collaboration and dissent preservation.

Phase 3 Outcome: Advanced AI projects demonstrating not just capability but accountability.

Phase 4: Governed Deployment (Months 19-24)

Goal: Earn enterprise certification and deploy a governed system demonstrating production-ready governance architecture.

Capstone: Governed System with five components:

1. HAIA-RECCLIN Architecture Diagram
2. CBG Checkpoint Map
3. Dissent Logging Infrastructure
4. Audit Trail
5. Failure Mode Documentation

Phase 4 Outcome: Enterprise certification. Deployed system with governance architecture. Portfolio demonstrating auditable system design.

Enterprise Cloud Certifications

Choose ONE cloud platform for your enterprise credential. Each provider offers free preparation resources before the paid certification exam.

Migration Between Tracks

The tracks share foundation courses. Migration does not mean starting over.

• Track 1 Phase 1 to Track 2 Phase 1: No additional time. Python for Everybody appears in both tracks.
• Track 1 Phase 2 to Track 2 Phase 2: Add approximately 2-3 months for Mathematics for ML and MLOps.
• Track 2 to Track 1: Subtract time. Engineering track includes all Leadership content.

Courses Are Examples. Structure Is Constant.

The specific courses listed here will evolve. What remains constant: governance foundation before technical depth, role assignment clarifying AI contribution scope, checkpoints preserving human decision authority, personalization prompts ensuring critical reading, dissent preservation surfacing alternatives, and capstones producing auditable artifacts.

Detailed Course Catalog Here

Optional Deeper Reading

For those who want deeper grounding in why governance matters at this moment in AI development:

Governing AI: When Capability Exceeds Control

This book expands the “why” behind the frameworks used in this roadmap. The book is not required. Everything you need to practice governed AI is here. The book provides context for those who want to understand the deeper principles.

Summary

This roadmap teaches what most AI education omits: how to stay in charge of what you build or direct. Two tracks serve different roles. Both share a governance foundation: role-based collaboration through HAIA-RECCLIN, checkpoint-based human oversight, and personalization prompts that force critical reading.

The goal is not AI skill alone. The goal is AI skill paired with the judgment and habits that keep those skills under human authority.

Governance without evidence is belief. Governance with checkpoints is proof.

This catalog supplements the curated AI Learning Roadmaps. Track 1 (Leadership) and Track 2 (Engineering) remain the recommended structured learning sequences.
Last Updated: December 2025

Detailed Course Catalog Here