Workflow: Preparing a Teaching Slide Deck

Preparing a teaching session takes longer than giving one. Most of that time goes into structuring the content and finding the right words. An AI tool can accelerate both. This workflow covers building a teaching slide deck from first outline to finished speaker notes.

1. Define your session in one sentence

   Before you open any tool, write one sentence that describes the session: who the audience is, what they should know by the end, and how long the session runs. For example: "A 20-minute teaching session for Foundation Year 2 doctors on the initial assessment of abnormal uterine bleeding in the outpatient setting." This sentence becomes the anchor for every subsequent step.

2. Ask the tool to generate a session outline

   Open your AI tool and paste your one-sentence description. Then ask: "Suggest an outline for this teaching session with six to eight slides. For each slide, give a title and two or three bullet points summarising the key teaching points." Review the outline critically. Add anything missing, remove anything that is outside the session scope, and reorder if the flow is wrong.

3. Refine the outline with one follow-up instruction

   It is rare that the first outline is exactly right. Pick the one or two slides that need the most adjustment and ask for a revision. For example: "Expand slide 4 to include a brief explanation of the role of outpatient hysteroscopy." Or: "Slide 2 is too basic for this audience. Replace it with content appropriate for doctors with six months of clinical experience." The tool will revise only what you ask.

4. Draft speaker notes for each slide

   Once the outline is correct, ask the tool to draft speaker notes. Do this one slide at a time, not all at once. For each slide, paste the title and bullet points and ask: "Write speaker notes for this slide as if I am talking to Foundation Year 2 doctors for two minutes. Use plain, direct language. No jargon without explanation." Review each set of notes before moving to the next slide.

5. Check every clinical fact

   Read the speaker notes as a clinician, not as a writer. The tool can make plausible-sounding errors — particularly with clinical thresholds, investigation sequences, or current guideline recommendations. Any factual claim that you did not personally supply to the tool should be verified against your own knowledge or a current guideline document before you stand up and teach it.

6. Add case-based examples yourself

   AI tools are less reliable at generating realistic clinical vignettes than at structuring content. Write your own case examples. A patient scenario you have seen in your own practice is more credible to trainees than a fictional one generated by a tool, and it is certainly more clinically accurate.

7. Build the slides in your presentation software

   Copy your outline and speaker notes into your preferred presentation software. Add your own images, diagrams, or clinical images following your institution's guidance on patient confidentiality and image consent. The tool has done the writing. The design and layout are yours.

8. Ask the tool to generate audience questions

   Once your slides are ready, ask the tool: "Generate five questions an engaged trainee might ask after a 20-minute teaching session on abnormal uterine bleeding in the outpatient setting." This is a useful way to stress-test your own preparation and anticipate where trainees will push back or seek more detail.

What this means for you

The quality of a teaching session depends on clinical experience and educational skill — not on how long the preparation took. This workflow compresses the structural work so you spend your preparation time on the parts only you can contribute.

When not to use this workflow

Do not use this approach for a session on a topic where your own knowledge is uncertain. The tool will fill gaps with plausible content that may be incorrect. If you would not be confident answering trainee questions without notes, prepare the session the traditional way — or work alongside a colleague who has clinical expertise in that area.

Remember: AI is a helpful assistant, not a clinician. You make the call.