Get more out of Geb AI
Battle-tested prompts you can copy, edit, and ship. Each template explains why it works so you build the skill, not just the prompt.
Three techniques recur in clinical prompting: explicit patient context, ranking instead of listing, and asking for the mechanism. Geb gives stronger answers when you show your clinical reasoning in the prompt itself.
Template 1 · Drug interaction screening with patient context
Analyze this medication profile for clinically significant interactions. Patient: {{age}}yo {{sex}}, key conditions: {{conditions}}. Medications: {{med_list}}. RANK interactions Major/Moderate/Minor with mechanism for each. Flag cumulative risks (additive QTc, bleeding, sedation). Suggest specific monitoring parameters with timing.
{{age}}{{sex}}{{conditions}}{{med_list}}Three techniques at work: (1) explicit patient context narrows the analysis, (2) RANK forces prioritization instead of a flat list, (3) asking for mechanism gets you the WHY, not just the WHAT. The cumulative-risk ask catches what pair-wise checkers miss.
Template 2 · ADR differential diagnosis
Patient started {{drug}} {{dose}} {{duration}} ago. New symptoms: {{symptoms}}. PMH: {{pmh}}. Analyze likelihood each symptom is an ADR vs other causes. Provide mechanism. Suggest alternatives in same class and different class with one-line tradeoffs.
{{drug}}{{dose}}{{duration}}{{symptoms}}{{pmh}}Asking for likelihood + mechanism + alternatives gives you a complete clinical picture. The 'one-line tradeoffs' constraint keeps the response actionable, not encyclopedic.
Template 3 · Patient counseling guide in any language
Create a step-by-step counseling guide for {{device_or_med}} in {{language}}. Reading level: 6th grade. Include: numbered actions, 3 teach-back questions to confirm understanding, warning signs with what to do. Format as a 1-page handout.
{{device_or_med}}{{language}}Reading-level + teach-back questions + warning-sign list are pharmacy gold standards baked into the prompt. The format constraint ('1-page handout') controls length and structure together.
Template 4 · Prior authorization appeal letter
Draft a prior-auth appeal letter. Patient: {{age}}yo with {{diagnosis}}. Insurance denied {{drug}}, citing {{denial_reason}}. Patient has tried and failed: {{failed_treatments}}. Cite the latest {{specialty}} guidelines (e.g., {{guideline_name}}). Anticipate common reviewer objections and address them in the body.
{{age}}{{diagnosis}}{{drug}}{{denial_reason}}{{failed_treatments}}{{specialty}}{{guideline_name}}The 'anticipate reviewer objections' instruction is the killer move — it produces letters that pre-empt the typical second denial. Citing specific guidelines makes the letter defensible.
Template 5 · Tapering schedule with counseling
Generate a conservative tapering schedule for {{drug}} {{dose}} taken for {{duration}}. Goal: minimize discontinuation syndrome. Provide: calendar-based schedule with specific dose reductions, expected symptoms with patient-friendly explanation, when to slow down vs hold, when to call the prescriber.
{{drug}}{{dose}}{{duration}}Calendar-based output gives the patient and prescriber something to reference. Distinguishing 'slow down' from 'call prescriber' creates a clear escalation path.
Drop one of these into Geb AI and see what comes back.
Open Geb AI→