AI-Augmented Leadership Learning
RockMouse has designed a suite of AI learning activities that map directly to the DeakinCo Leadership Program journey — June through November. Each activity is grounded in the program's content and values, using AI as a coach, connector and insight engine — not a replacement for facilitation. Three activities are specifically designed to support the Evaluation Framework, generating evidence of impact at the learner, cohort and program level.
10
AI Learning
Activities
Activities
6
Program
Months
Months
3
Evaluation
Activities
Activities
3
Content
Tools
Tools
Design Principles
P1
AI as Coach, Not Assessor
Every activity supports reflection, extends thinking and personalises feedback. AI never scores or ranks participants. Psychological safety is protected by design.
P2
Context Before Content
Activities are anchored to the participant's real experience — their organisation, their current challenge, their workshop touchpoint. Generic prompts are replaced by program-specific framing.
P3
Learning Loops, Not Events
Activities build on each other. A pre-program reflection informs graduation. A workshop insight shapes the nudge recommendation. AI provides connective tissue across 6 months.
P4
Four Engagement Modes
Solo self-paced · Cohort and peer-based · Participant and manager together · Facilitator-led with AI supporting. Each activity specifies its primary mode.
P5
Data for Learning Only
Where AI generates data — reflection themes, decision patterns — that data belongs to the learner's journey. Aggregate insights inform program design; individual data stays private.
P6
Progressive Complexity
Early activities are low-stakes and confidence-building. By graduation, participants are fluent in AI-assisted reflection. The tool grows with the learner.
Agent Types Used in This Program
🔮 Insights Agent — clusters, analyses & surfaces patterns from qualitative input
🎯 Scenario Coach — delivers branching dilemmas & personalised decision coaching
🗺 Mapper / Connector — maps tasks, skills & nudges to goals & program stages
🤝 Cybernetic Teammate — live collaborative partner for coaching & group work
Program at a Glance
👩
Meet May — Our Learner Persona
Frontline Team Leader · Healthcare Sector · Mid-career · First leadership program
"I know my team, but I'm not sure I know myself as a leader yet. I want to understand how I come across and how I can get the best out of my people — especially when things get hard."
May's AI-Augmented Journey — June to November 2026
10 AI Learning Activities
Each activity is fully embeddable and runs via the Anthropic API. Click any activity to expand and try the live demo. Set your API key in Settings first.
4
Agent
Types
Types
3
Eval
Mapped
Mapped
AI Content Creation Tools
Three purpose-built tools to accelerate and improve the quality of learning content development — from currency checking to expert knowledge capture to storyboarding.
Content Currency
Keep your learning content current and accurate
Paste existing learning content and AI will audit it for currency — flagging outdated statistics, superseded regulations, obsolete tools or language, and suggesting specific updates with rationale.
Identifies outdated facts, figures and references
Flags legislative or regulatory changes
Suggests specific, sourced replacement content
Highlights terminology that has evolved
Produces a structured currency report
▶ Try Content Currency
SME Interview Tool
Extract expert knowledge efficiently
Tell AI the topic and your subject matter expert's background. It will generate a structured interview guide with smart follow-up questions — then analyse the transcript to extract structured learning insights.
Generates tailored interview question sets
Smart follow-up and probing questions
Analyses transcripts for key themes
Structures knowledge into learning-ready content
Identifies gaps and contradictions
▶ Try SME Interview
Storyboard Tool
From objectives to learning narrative
Describe your learning objective, target audience and context. AI will generate a full scene-by-scene storyboard script — complete with narration, on-screen text, interaction cues and instructional notes.
Scene-by-scene script with narration
On-screen text and visual direction notes
Interaction and branching cues
Instructional design rationale per scene
Formatted for developer handover
▶ Try Storyboard
Evaluation Framework — AI-Supported Measurement
Three of the 10 AI activities are specifically designed to generate evaluation evidence across the Kirkpatrick Four Levels. Together they provide a continuous thread of program impact data — from participant reaction through to workplace behaviour change.
Activities ① ⑩
Kirkpatrick Levels 1 & 2
Reaction & Learning — Welcome Pulse + Graduate Story
Activity ① (Welcome Pulse, June) captures baseline expectations, confidence and anxieties before the program begins. Activity ⑩ (Graduate Story, November) repeats comparable questions at program end — and AI surfaces the delta: what shifted, what grew, what surprised the participant. Together they provide Kirkpatrick Level 1 (Reaction) and Level 2 (Learning) evidence.
Level 1 — Reaction
Pre and post sentiment · Satisfaction indicators · Perceived relevance scores
Level 2 — Learning
Confidence shift in key competencies · Language complexity growth · Self-assessed capability gains
AI Output
Individual narrative arc · Cohort-level aggregated themes · Program facilitator summary
▶ Demo: Graduate Story — AI Comparison
Activity ⑨
Kirkpatrick Level 3
Behaviour — Cohort Insights Synthesiser
After each workshop, facilitators submit structured session observations and participant reflection summaries. The Cohort Insights Synthesiser clusters themes across the cohort — surfacing patterns in how participants are applying learning back at work, where behaviour change is visible, and where it is not yet translating.
Level 3 — Behaviour
Transfer to workplace · Manager-observed changes · Peer-reported shifts
Cohort Patterns
Common themes · Outlier insights · Facilitator priority flags
Facilitator Value
Evidence-informed debrief prep · Targeted follow-up nudges · Program design feedback
▶ Demo: Cohort Insights Synthesiser
Activities ① ⑨ ⑩
Kirkpatrick Level 4
Results — Program Impact Report
At program end, AI aggregates data from all three evaluation activities — Welcome Pulse baselines, Cohort Insights themes, and Graduate Story narratives — to generate a Program Impact Report. This report connects participant growth to organisational outcomes: retention, team performance indicators, and leadership capability uplift.
Level 4 — Results
Program ROI narrative · Cohort capability uplift · Organisational impact indicators
Report Sections
Executive summary · Learner journey highlights · Recommendations for next cohort
Audience
L&D leads · Senior stakeholders · Program sponsors · Next cohort design
How it works: Paste a summary of your cohort's journey — Welcome Pulse themes, workshop insights and graduation reflections. AI generates a structured Program Impact Report formatted for senior stakeholders.
▶ Generate Program Impact Report
API Configuration
Live API requests are routed through a secure Cloudflare Worker proxy. The Anthropic API key is held server-side as an encrypted secret and never exposed to the browser.
Proxy Endpoint
How it works:
1. The browser POSTs requests to
2. A Cloudflare Worker receives the request, attaches the API key, and forwards it to
3. The model response streams back through the Worker to the page.
To rotate the key: Cloudflare dashboard → Workers & Pages →
Demo Mode (default) works without any backend — use the toggle in the header to switch into Live API mode.
1. The browser POSTs requests to
/api/claude on this domain.2. A Cloudflare Worker receives the request, attaches the API key, and forwards it to
api.anthropic.com.3. The model response streams back through the Worker to the page.
To rotate the key: Cloudflare dashboard → Workers & Pages →
deakinco-claude-proxy → Settings → Variables and Secrets → update ANTHROPIC_API_KEY.Demo Mode (default) works without any backend — use the toggle in the header to switch into Live API mode.
Model Information
Model
claude-sonnet-4-6
Max Tokens
1,000 per demo request
Temperature
0.7 (creative + structured)