How AI Changes Storyboarding for L&D

This is Part 1 of The L&D Workflow, a series about how the important steps in creating training change when AI works directly on the course.

Storyboarding sits between thinking and building. An instructional designer turns source material into a screen-by-screen plan, then someone recreates that plan in PowerPoint, Storyline, Rise, Captivate, HTML or another authoring tool.

Across projects I have worked on, storyboarding often absorbs roughly 20–30% of the hands-on creation effort. Treat that as a planning range, not an industry benchmark. The widely cited Chapman Alliance development-time study (PDF) shows how much total effort goes into instructional design, slide development and stakeholder review, but it does not isolate storyboarding as a separate phase.

AI changes the output of this work. The first storyboard can now be a working version of the training.

The Old Storyboard Was a Handoff

A traditional storyboard gave the team a cheaper artifact to review than a finished course. It also introduced a translation gap.

The designer described the screen. Reviewers commented on that description. A developer interpreted both. The team often discovered later that the copy was too long, the interaction felt awkward or the example failed in context.

	Traditional storyboard	AI-assisted storyboard
Output	Description of a screen	A working screen or editable slide
Review	Reviewers imagine the experience	Reviewers use the experience
Revision	Rebuild a section in the authoring tool	Change one screen and preserve the rest
Design system	Recreated by hand	Reused from an approved course
Handoff	A developer interprets the document	The draft becomes the starting artifact

A Worked Example: A Pricing-Objection Lesson

Consider a six-screen lesson that teaches a new sales rep to handle a pricing objection.

The instructional designer starts with an approved talk track, a fixed compliance statement and an existing course whose visual system has already passed review. The first request defines the audience, performance objective, scenario, practice and feedback. The result is a working HTML lesson or editable slide deck that the team can click through.

During review, screen five feels too dense. The designer gives the agent a narrow instruction:

On screen 5, tighten the interaction note to one short sentence. Keep the approved compliance language. Leave screens 1–4 and 6 unchanged.

The agent updates that screen and reports the scope of the change. The designer then checks the revised interaction against the objective, sources, accessibility requirements and surrounding flow.

A bounded course-chat request changes one screen while the rest of the storyboard remains untouched.

A Practical AI Storyboarding Workflow

1. Define the performance

Write what the learner should be able to do and in which situation. Do this before deciding how many screens the course needs.

2. Separate approved sources from reference material

Mark the current policy, product facts and required language. Tell the agent what may be rewritten and what must remain exact.

3. Generate something reviewers can experience

Request HTML, a PPTX or both. Include the audience, objective, learning flow, practice activity and feedback.

4. Revise narrowly, then validate

Name the screen, the allowed change and the content that must stay fixed. The instructional designer still checks accuracy, cognitive load, accessibility, practice quality and alignment with the objective.

Where AmpUp Fits

Most training teams already have courses they trust. Existing SCORM packages, including courses built with Articulate Storyline and Rise, contain approved decisions about typography, color, spacing, navigation, interactions and quiz states.

AmpUp Courses imports material from different SCORM authoring tools and H5P, identifies those repeatable decisions and organizes them as reusable brand tokens. The team refines the new course through chat and a visual editor, then exports the approved result as SCORM for its LMS. Those tokens become a specification for future courses rather than a screenshot someone has to imitate.

Where This Still Breaks

Faster production does not protect a team from weak inputs or weak judgment.

• Outdated source material produces a polished but outdated lesson.
• Generic prompts produce generic practice and feedback.
• Brand reuse does not prove that an interaction is accessible.
• A valid SCORM package may still behave differently across LMSs.
• Proprietary material needs an approved data-handling policy before it enters any model.

Human review gates remain essential. This is an example of augmentation before automation: the agent handles production work while the instructional designer owns the learning decision.

The Takeaway

Storyboarding is becoming the beginning of the course itself. Teams can experience the draft, revise a bounded part and carry approved design decisions forward without recreating every screen.

That changes the next phase too. Once the storyboard is working, review comments can point to the exact content that needs attention. Part 2 follows that collaboration loop.

Frequently Asked Questions

Can AI create the complete training course?

No. It can move a team from approved sources to a working draft much, much faster. People still own the learning design, accuracy, accessibility, testing and final approval.

Can an existing SCORM course become a design reference?

Yes. Its recurring visual and interaction decisions can be extracted into a reusable token set, then reviewed before use across new courses.

Where does AmpUp fit?

AmpUp connects an existing course library to the courses a team needs to build next. It supports SCORM and H5P import, brand-token reuse, chat-based and visual editing, and SCORM export for the LMS.