Second Edition · Updated for 2026

The New Learning Stack.

Using AI to Transform Instructional Design and Digital Learning

Charles Hinds, M.Ed.

Core insight

Training that stops when a course ends is only half the job. The other half is supporting people while they work.

AI is no longer a horizon you are moving toward. It is the environment learning design now happens inside. The designers who will define the next decade are not the ones who know the most tools — they are the ones who ask the deepest questions about what learners actually need.

60–75%

Reduction in information-search time in early Role IQ testing

Chapters covering every phase of the AI-enabled design pipeline

The core problem

Why learning stops working at the worst moment.

Learning and development has a gap problem. Learners finish courses, receive certificates — then return to their desks and face real situations the training never anticipated.

Learners are overwhelmed and self-reliant

People search Google, ask colleagues, or guess — because the training finished and the system went silent.

Training doesn't reflect the real world

Courses cover the clean version of a process. Real work is messy. The gap between them is where performance fails.

Feedback arrives too late to matter

If a learner makes the wrong call on Monday, they won't hear about it until the quarterly review — if at all.

Personalization rarely delivers

Most systems promise adaptive learning but deliver the same experience to everyone. Real personalization requires better infrastructure.

"AI cannot just live in a private back-and-forth between a learning designer and a tool. Real learning design has always been built through conversations. AI should join that dialogue — not replace it."

— The New Learning Stack, Introduction

Central framework

The Learning to Performance Bridge.

The Role Intelligence Layer sits between what learning designers build and the reality employees face the moment the course window closes. It closes the gap that training alone never could.

The Learning to Performance Bridge

Learning Environment

Courses & eLearning Modules

Training Programs

Microlearning

Knowledge Articles

Job Aids

Role Intelligence Layer

Trusted answers during real work
Role-specific guidance on demand
Grounded in approved knowledge

Work Environment

Real Tasks & Decisions

Operational Procedures

Customer Interactions

Unexpected Situations

Time Pressure

Figure I.1 — The Role Intelligence Layer connects formal learning to real work demands, providing trusted, role-specific answers at the moment of need.

What's inside

A playbook across the full design pipeline.

Each chapter gives learning designers practical, AI-enabled tools for a specific phase of their work — from initial research through to real-time performance support.

Chapter 1

AI-Powered Needs Analysis

Use AI to surface learner needs faster — moving from vague stakeholder requests to sharp, evidence-based design decisions.

Chapter 2

The Discovery Phase

Structure your discovery conversations with AI so no critical context falls through the cracks before design begins.

Chapter 3

Structuring Learning

Build modular, searchable content architecture that evolves — instead of monolithic courses that become impossible to update.

Chapter 4

AI as a Design Partner

Understand how agentic AI handles sequences of design tasks autonomously — and where human judgment is irreplaceable.

Chapter 5

Designing for the Learner

Empathy is the foundation. Use AI to generate personas, simulate perspectives, and pressure-test assumptions before launch.

Chapter 6

Writing & Narration

Produce clear, engaging scripts faster — while maintaining the human voice that makes learning feel real, not generated.

Chapter 7

Review & Feedback Loops

Redesign your review process so AI handles the routine, stakeholders focus on the meaningful, and cycles run in days.

Chapter 8

Visuals & Media

Generate on-brand visuals, diagrams, and multimedia assets that support comprehension — without requiring a design team.

Chapter 9

AI-Powered Narration

2026 tools interpret emotional context, not just execute voice parameters. This chapter shows how to use that shift.

Chapter 10

Learner Experience & UX

Apply UX principles to learning design. Remove friction, reduce cognitive load, and build experiences people complete.

Chapter 11

Assessments That Matter

Build assessments that reveal real understanding, not just recall. AI can personalize challenge levels and evaluate responses.

Chapter 12

Data & Learning Analytics

Turn learner behavior data into design decisions. Know where people drop off and what actually drives performance improvement.

Chapter 13

Future-Proofing Your Stack

Build modular, indexed, AI-ready content. Your content library becomes the infrastructure — organized content powers smarter systems.

Chapter 14

Adaptive Learning Systems

Design LMS and LXP experiences that respond to actual learner behavior — not static completion — for outcomes that scale with the individual.

The central concept

The Role Intelligence Layer.

Most training ends the moment a learner clicks Submit on their final assessment. But real work begins right after. The questions that matter most arrive days or weeks later — in the middle of an actual task, under time pressure, with no instructor in sight.

The Role Intelligence Layer is an AI-powered support system that lives between training content and real work. It connects learning materials, policies, job aids, and operational knowledge — and delivers trusted, cited answers exactly when employees need them.

It uses retrieval-augmented generation (RAG): rather than guessing, it pulls answers directly from the organization's own approved content. That's what makes it trustworthy enough to act on.

See the full Role Intelligence Loop

60–75% reduction

in information-search time in early Role IQ testing, with a significant drop in routine HR and support escalations.

How Role IQ works in Microsoft Teams

Employee completes training

Formal course content builds foundational knowledge.

Real work begins

They encounter a situation the course didn't fully cover.

Question asked in natural language

Instead of interrupting a colleague, they ask Role IQ directly inside Teams — where the work is already happening.

Grounded, cited answer arrives instantly

Pulled from approved internal content. Auto-updates when source changes. Includes citations so the employee can verify.

Performance improves — and feeds back in

Interaction data reveals gaps in training content, driving continuous improvement of the learning system.

"The content you structure, tag, and maintain is the retrieval layer. The designer's craft becomes the infrastructure."

— The New Learning Stack, Chapter 13

The continuous model

The Role Intelligence Loop.

Learning doesn't stop at course completion, and neither should the system supporting it. The Role Intelligence Loop shows how training, real work, AI support, and continuous improvement connect into a cycle that gets smarter over time.

Figure 13.1 — The Role Intelligence Loop

Training & Learning Content

Structured courses and materials build foundational knowledge and role-specific skills.

Simulation & Practice

AI-driven scenarios let learners rehearse decisions before consequences are real.

Real Work Tasks

Learners apply what they know. This is where the gap between training and performance becomes visible.

Questions & Uncertainty

Edge cases arise. This is where most systems go silent.

Role Intelligence Layer — The Bridge

Trusted, grounded, role-specific answers — pulled from approved content, delivered in the tool where work happens.

Trusted Guidance & Answers

The employee gets an answer they can act on. Cited. Current. Specific to their role.

Improved Performance

Better decisions. Fewer escalations. Faster task completion. Performance that compounds over time.

New Insights Feed Back In

What people ask reveals what training missed. Those gaps improve the next cycle of content.

Key insights

What this book changes about how you think.

The course is not the finish line

Training that ends at completion is only half the design. The other half is the system that supports performance after the course window closes.

Your content library is infrastructure

Modular, tagged, searchable content isn't just easier to update — it's the retrieval layer that powers AI-driven performance support.

Build judgment, not tool loyalty

AI tools release on near-monthly cycles. Designers who build principles outlast every platform change.

AI joins the conversation, not replaces it

Good learning design is built from real voices: learners, SMEs, stakeholders. AI amplifies that dialogue.

Empathy is still the foundation

Every AI capability in this book serves one goal: designing for the real, specific, time-pressured person on the other side of the screen.

The infrastructure already exists

Role Intelligence systems can be built inside Microsoft Teams and Slack today. What's missing is designers who know how to use them.

About the author

Written for people doing real work under real pressure.

Charles Hinds

M.Ed., Western University · Learning Design Leader · Author

Charles Hinds began his career as a technical writer, focused on taking complexity out of technology so real people could use it without frustration. That instinct — clarity, empathy, and design for the actual person on the other side of the screen — became the foundation of his approach to learning design.

He has led learning and performance initiatives for organizations including Walmart Canada, Costco Canada, Canada Post, and Manulife Financial. During his M.Ed. at Western University, he researched AI-powered learning interventions and built Role IQ — a production-viable AI support agent for enterprise environments.

The New Learning Stack is the distillation of that research, practice, and the conviction that learning design's most important work is still ahead of it.

Instructional Designers

L&D Directors

Learning Developers

eLearning Specialists

L&D Managers

HR Leaders

Training Coordinators

"Every field eventually finds its divide. In learning design, that divide is happening now. The ones who adapt will find AI does not replace their skill — it amplifies it."