Clinical Efficiency & Ambient Intelligence

The Diagnostic Prompt IMI — The Framework at the Heart of This Series

The session that anchors everything. The Diagnostic Prompt IMI is the structured practice that runs through all three programmes — a way of working with AI that starts before either of you produces anything. The framework operates across three dimensions: Question surfaces the assumptions you didn't know you were carrying. Clarity replaces the assumed problem with the real one. Mirror is where your thinking and AI's thinking sharpen each other in the exchange. Live demonstration across Claude, ChatGPT/GPT-4o, and Gemini — so you see the same problem handled across all three tools. Perplexity AI introduced for real-time cited answers on shift.

Applying the Diagnostic Prompt IMI to a Clinical Task

You select one genuine problem from your clinical practice — a documentation challenge, a recurring communication task, or an administrative burden. Before turning to AI for a solution, you work through the Diagnostic Prompt IMI: surface your assumptions, clarify the real problem, and engage AI as a thinking partner rather than an answer generator. You record what emerged — what you assumed the problem was at the start, and what it actually turned out to be.

AI-Assisted Clinical Documentation, Patient Communication & Australian Compliance

Structured online teaching session covering AI ambient scribing tools, a three-step review protocol for AI-generated clinical documents, AI-assisted patient communication drafting (discharge summaries, letters, SMS recall), quality verification framework, Australian Privacy Principles, My Health Record obligations, and data residency considerations.

Reflection on AI-Assisted Documentation & Patient Communication Drafting

Participant uses an AI ambient scribing tool for real clinical documentation tasks, applying the three-step review protocol. Participant also drafts one piece of patient communication using AI and applies the quality verification framework before using or sending it.

No-Code Workflow Automation — Connecting AI to Clinical Processes

Structured online teaching session covering agentic automation using Claude Cowork — connectors, skills, and scheduled tasks — taught platform-agnostically, since the same discipline applies whether the underlying tool is Claude, ChatGPT, or Gemini. The Custodian Prompt™ is introduced as the interrogation step run before any automation is built — surfacing what patient- or organisationally-identifiable information a task would touch, before it touches it. Live demonstration coordinating email and calendar as a single automation and scheduling it to run without manual repetition, specified using a reusable formula (job, tool, categories, output, boundary). Two reusable skills are built live: The Diagnostic Prompt™ packaged as a callable skill, and a second skill that turns a source document into a slide outline, always asking purpose and audience first. Human-in-loop checkpoint design taught as a staged trust model (visible, efficient, automated, delegated) — no automation in a clinical setting reaches full autonomy without deliberate, earned trust. Five high-value clinical automation patterns applicable across specialties.

Reflection on Automation Deployment in Practice

Participant builds and deploys one clinical workflow automation relevant to their practice. They run it with real or realistic data and document what it produced, what they corrected, and what impact it had on the clinical process.

Ambient Scribe Templates and Your First AI Skill

Working live session. Participants arrive with an ambient scribe template they have already tested and a first Claude or ChatGPT skill. Session time is used to refine both, troubleshoot failures, and make the outputs reliable enough to keep using.

Reflection — Scribe Template, Skill and What You Learned

Participant writes up what they learned from building and testing an ambient scribe template and a first Claude or ChatGPT skill — including what they corrected and what they will keep using.