Redesigning Online Learning with AI

Lessons learned from letting go of the blank page

Introduction

In online course design, the “blank page” rarely means a literal empty screen. It’s the moment of uncertainty that comes at the very start — when the learning designer is waiting for input from the academic lead, and the academic (or subject matter expert) is unsure how to turn years of teaching and disciplinary expertise into a fully online learning experience.

At Learning Design Solutions, we always begin with collaboration. That initial conversation between SME and learning designer is where a course really takes shape. But over the past year, we’ve been exploring how generative AI can act as a third voice in that early-stage design: not replacing the human exchange, supporting it.

This post shares what we learned when we deliberately designed a demo course — Principles of Responsible Management — using a GPT-based AI assistant trained in our own pedagogical approach. Our aim was to test whether AI could support a structured, theoretically grounded design process suitable for UK higher education at postgraduate level. And what we found was that it didn’t just help us start — it helped us start well and keep going.

Why the Blank Page Is a Shared Problem

For subject matter experts, the blank page is about translation: how do I take everything I know about teaching this subject and structure it as an online learning experience?

For learning designers, the blank page is often about waiting: we know how to scaffold a course, but without early input, we’re left in a holding pattern.

This is where we’ve found AI most useful — not for finishing tasks, but for starting conversations. Using a GPT-powered tool that we’ve trained in key HE pedagogical frameworks, we’re able to generate structured, justified design drafts that both parties can respond to. It gives us a shape to work with — and that makes a huge difference.

The Principles of Responsible Management Pilot

To explore the potential of AI in course design, we created a demo module: Principles of Responsible Management. It was a short (5-credit) postgraduate course, co-developed by an experienced academic and one of our senior learning designers, with AI used to support:

• Drafting intended learning outcomes at master’s level

• Designing aligned assessments

• Proposing weekly topics and learning activities

• Suggesting case studies, micro-lecture content, and interactive formats

Importantly, nothing was handed over to AI entirely. Every output was reviewed, revised and aligned in conversation with the academic partner — who had full ownership of the final content.

This process gave us a glimpse of how AI might fit into a scalable, human-led course design methodology.

How Pedagogical Theory Was Embedded

The AI wasn’t working from scratch. We trained it in the core frameworks we use across our learning design practice:

• Backward Design (Wiggins & McTighe): Starting from intended outcomes and assessment, then planning content and delivery methods.

• Constructive Alignment (Biggs): Ensuring that all content and activity types clearly support the module’s outcomes and assessment tasks.

• Bloom’s Taxonomy: Matching verbs and tasks to the appropriate level of cognitive demand for postgraduate learners.

• Laurillard’s Six Learning Types: Using theory to select and justify learning activities based on how students learn through practice, inquiry, production or discussion.

• Gagné’s Nine Events of Instruction: Ensuring that each module not only contains the right elements but sequences them for learner engagement and retention.

Because of this, the drafts we received were more than just filler — they were pedagogically purposeful. They provided the basis for meaningful design conversations, not generic content.

Letting AI Help Us Continue the Design

The biggest impact wasn’t just in getting started — it was in keeping momentum. Once a draft outline was available, the AI continued to support development by providing just-in-time content and structure suggestions at each stage of the storyboard.

With a structured starting point already on the table, the SME was able to:

• Reflect on whether the proposed activities matched their goals

• Suggest changes in format or tone with more confidence

• Feel that they were collaborating in a process that respected their academic judgement and understanding of the subject

Meanwhile, learning designers could focus on alignment, accessibility, and scaffolding — rather than waiting for the next SME draft to arrive. AI became part of the iteration, not just the initial draft.

What We Learned

Starting well changes the tone. Having a structured, theory-informed draft allowed us to begin conversations with focus and clarity. The SME wasn’t asked to ‘fill in’ empty templates or produce large volumes of text from scratch. Instead, they engaged in evaluating and shaping content in relation to clear pedagogical principles — a far more effective use of their time and expertise.

AI works best when roles are clear. We learned that AI doesn't flatten expertise — it helps each role do its job better. The SME retained ownership of content and disciplinary direction. Learning designers brought their lens of pedagogical alignment, accessibility and learner experience. And AI provided scale and speed, without replacing insight.

Theory increases trust. When suggestions from the AI were grounded in familiar pedagogical models it helped build confidence in the process. Academic staff were more comfortable engaging with proposals that were theoretically justified, rather than “automated guesses.”

Momentum matters. AI didn’t just help us start. It helped us keep going. By supporting iterative development — such as proposing alternative ways to present concepts, suggesting case study prompts, or restructuring activities for variety — it helped us avoid stalling at typical bottlenecks. For the learning designer, this freed up time to focus on more nuanced design decisions.

Next Steps

This pilot became the foundation for a wider implementation. We’re now applying the same approach across three full degree programmes — MBA, MSc Psychology, and MSc Computer Science — for a London-based higher education provider. In these projects, AI is used from the very beginning, always in collaboration with a module author and a learning designer.

What’s emerging is not automation, but a new kind of teamwork, where AI scaffolds early decisions, theory anchors the design, and human judgement leads the process.

We’ll share more on that in future posts — including what we’re learning about assessment, multimedia, and the pedagogy of prompting.

Explore Further

You can view the full Principles of Responsible Management demo course online. Simply go to:

https://learningdesignsolutions.moodlecloud.com/course/view.php?id=9

Select “Access as a guest” when prompted.

Book a Free Consultation

If you’d like to explore how AI-supported design could work in your institution, we’d be happy to talk.