Case study · AI-enabled course design · Live & public, no confidentiality

Leading Through the Unknown: An AI Leadership Story

A narrative-driven leadership course with a live, generative AI coaching simulator built in—designed, built, and deployed end to end by a single instructional designer.

Live AI coachingreal-time, personalized, generative—not scripted branching
Solorole: sole designer & developer
Articulate-freehand-built HTML/CSS/JS
Netlifydeployment
Actual product — live, functioning AI coaching. Real screens from a real interaction; the coaching text below was genuinely generated, not scripted. Find it yourself at Module 2, Lesson 3 ("The Priya 1:1").

The learner writes what they’d actually say in a difficult 1:1 conversation, in their own words—then gets personalized coaching back, generated live by an LLM grounded in the scenario, not a bank of pre-written branches.

The Priya 1:1 live AI coaching screen, showing the scenario setup and an open text box for the learner's response The same screen after submission, showing a real generated coaching response from the P3 Leadership Coach
Actual product — course landing & narrative. Also real screens, not recreations.
Leading Through the Unknown course landing screen, showing the module outline sidebar and course introduction Module 1 screen, 'The Mandate', showing the pre-teach video and reflection prompt
A — Analyze

The problem

Mid-level managers were being told to "lead through AI disruption" with no concrete model for what that meant day to day—just anxiety about job security and vague mandates to "adopt AI." The gap wasn’t technical literacy; it was a leadership framework for making decisions under genuine uncertainty, with AI as the specific catalyst.

Audience: people managers, not individual contributors—so the course needed to teach judgment and change leadership, not tool tutorials.

Constraint: built and shipped by one person, with no production team, no LMS budget, and a compressed timeline—which shaped almost every downstream decision in Design and Develop.
D — Design

Narrative and structure

Voice: second person, StoryBrand structure—the learner is the hero, P3 is the guide, not the hero of its own course. Content is sequenced to Bloom’s levels 2–5 (Understand through Evaluate) and Depth of Knowledge 1–3, so the course moves from comprehension into applied judgment rather than stalling at recall.

Citations are woven conversationally into the prose rather than dropped as a reference list at the end—keeping the tone consistent with the StoryBrand voice instead of breaking into an academic register.

D — Develop

Build

Built as a hand-coded HTML/CSS/JS course library (not an Articulate or Rise export), giving full control over the interaction design and page performance. The centerpiece is a live coaching simulator: a serverless function calls the Anthropic API with the scenario context and the learner’s own written response, and returns personalized, generative coaching—a genuinely different mechanism from a pre-scripted branching scenario, where every possible learner input already has an author-written reply waiting for it.

I — Implement

Deployment

Deployed via Netlify Drop to a public course library, with a per-module structure and the coaching simulator running live in production via a Netlify serverless function—confirmed working, not a demo stub.

In active polish: the custom learn.p3elearning.com subdomain isn’t connected yet (currently on the Netlify default domain), and a few scene/character image sets are still being finished. Neither blocks trying it—the coaching feature itself is live and functioning.
E — Evaluate

What this demonstrates

This case study exists to show range beyond a template-based authoring tool: narrative and instructional design, front-end build, generative-AI integration, and deployment—the full stack of what "solo delivery at team scale" actually looks like in practice. It's also the one artifact here you don't have to take my word for: click through and coach Renee yourself.

Tools used

HTML / CSS / JavaScript Netlify Anthropic API Claude

See the enterprise-scale version of this discipline