Leveraging-AI-for-Enhanced-Productivity.pdf

A short introduction to the use of AI for daily tasks

1. Leveraging AI for Enhanced
Productivity
Applications of Chatgpt and Other Platforms in Healthcare
by Ferwahn Fairis

3. A Brief History of AI
Evolution
1
1950s-60s
Early AI algorithms and neural networks emerge.
2
1990s-2000s
Machine learning grows; early iterations include Deep
Blue (chess).
AI9s theoretical backbone, harnessed hardware gains
to pivot toward scalable, practical applications.
3
2022-2025
Large language models (GPT-4, Med-Gemini) and
multimodal AI transform capabilities.
Leaps: Faster processing, real-time analysis, higher
accuracy (MedQA hits 91.1% in 2024).
Impact: AI interprets scans, predicts risks,
automates workflows.

4. Key AI Technology

5. Objective
1
Demonstrate AI's impact
Show how AI tools will enhance healthcare productivity.
2
Practical guidance
Provide examplesto use AI tools.

6. Why Healthcare Workers Must Embrace AI
The Imperative for AI in Healthcare
We are facing immense pressure from aging
populations, increasing chronic diseases, and staff
shortages, necessitating innovative solutions.
Healthcare Demands: An aging population and a
shrinking workforce are stretching resources thin.
AI9s Role: Automates repetitive tasks, giving
healthcare professionals more time to focus on
patient care.
Safety: Anticipates risks like patient decline or
workflow bottlenecks, improving outcomes for both
staff and patients.
Precision: Sharpens diagnostic accuracy and
optimizes resource use, from staffing to supplies.
Key Benefits
By adopting AI, healthcare facilities can enhance
patient care quality, streamline operations, and support
a more resilient workforce4vital for sustaining a high-
performing healthcare system.
Call to Action
Start integrating AI tools into your everyday routines to
unlock immediate gains in efficiency, safety, and
patient care delivery

7. Where do we begin?
FIRSTLY, WE USUALLY WRITE A PROMPT
A prompt is any input4such as text, an image, or a voice command
4that a user provides to an AI system to generate a response.
It serves as an instruction or query that guides the AI model in
producing relevant outputs, whether in the form of text, images,
code, or other data.

8. Introduction to
Prompt Engineering
Writing a good prompt matters because it directly
determines the quality and relevance of the language
model's output, enabling you to effectively harness its
capabilities for your specific needs.
Engineering Powerful Prompts
Role
Define the persona the language model should
adopt.
Objective
Clearly state the desired outcome of the
prompt.
Audience
Specify who the response is intended for.
Context
Provide relevant background information or
constraints.
Tone
Indicate the desired style of the response.
Output
Specify the format of the expected result.

9. Tips on good
prompting
Here are some pointers on refining your approach:
Role
Use role-playing to guide the model's
perspective. (e.g., "As a historian, explain...",
"Imagine you are an MSQH Surveyor...") Objective
Break down complex tasks into smaller,
manageable objectives. (e.g., "First, analyze the
MSQH Standard 2. Then, list the key points for
its audit.")
Audience
"Use analogies for a non-technical audience,"
"Provide detailed explanations for experts."
Context
Use specific examples and scenarios to provide
richer context. (e.g., "Based on the district
hospital setting of...", "In a situation where...")Tone
Employ nuanced tone adjustments to achieve
specific effects. (e.g., "Use a persuasive tone,"
"Write in a humorous style.") Output
Specify output formats with examples for
clarity. (e.g., "Provide a table with columns
for...", "Write a script with dialogue and stage
directions.")

10. Sample of a Bad vs Good Prompt
"Tell me about Chemical Safety""
Answer:
Chemical safety is about handling chemicals
carefully to avoid harm. It includes storing them
properly, using protection, and knowing what to do if
something goes wrong. It9s important in places like
labs or factories to keep people safe.
"As an OSH expert, list 3 essential tips for
new workers in handling chemicals in
hospital lab setting. Use a clear,
authoritative tone and focus on practical,
actionable advice".
Answer:
Wear Proper PPE: Always use appropriate personal protective
equipment4gloves, goggles, and respirators as required4to
shield yourself from exposure.
Store Chemicals Safely: Keep chemicals in labeled, secure
containers, away from heat or incompatible substances, to
prevent spills or reactions.
Know Emergency Procedures : Familiarize yourself with spill
response plans and first-aid steps4quick action can minimize
harm if an accident occurs.
Why is it bad?
Too vague: No specific goal or focus.
No role: Who9s answering4an expert, a worker?
No audience: Who9s this for4managers, employees?
No context: What kind of workplace4factory, office?
No tone or format: How should it sound, and how should it be
structured?
Why is it good?
Objective: What you want (e.g., safety tips).
Role: Who9s answering (e.g., OSH expert).
Audience: Who it9s for (e.g., new workers).
Context: Where it applies (e.g., hospital lab).
Tone: How it9s said (e.g., authoritative).
Format: How it9s presented (e.g., a list).

11. Multi-Modal Input Options to
assist Prompts

12. Enhanced Productivity with AI
Data Analysis
Presentation Design
Workflow Automation

13. How have i utilize AI Platforms in my daily
work?
' Information Retrieval & Summarization
 Summarized AI research on facial plethysmography
 Collected insights for Casemix documentation improvement
 Compiled cost-benefit analysis for virtual clinics
' Content Generation
 Drafted formal letters & proposals (e.g., Casemix website, virtual clinic)
 Created presentation scripts & reports

14. How have i utilize AI Platforms in my daily
work?
 Data Analysis & Research
 Assisted in analyzing facial scan dataset for oxygen saturation validation
 Helped outline variables for virtual clinic cost-benefit study
¿ Automation & Coding Assistance
 Provided guidance on ICD-11 coding automation
 Assisted with Python-based data analysis for research

15. How have i utilize AI Platforms in my daily
work?
ô Language Translation
 Translated documents (English µ Malay)
 Assisted in clearer documentation for medical officers
w Question Answering
 Explained ICD-11 coding automation for SMRP
 Answered AI-related queries for research & hospital use

16. Overview of AI Tools
1
ChatGPT/ Google Gemini / Grok/ Deepseek
Automate tasks like document creation and proposal
writing.
2
Perplexity
Efficient information retrieval, enhancing literature
review & research
3
Gamma & Prezi
Easy to create and visual engaging presentations
4
Canva
Powerful tool for content creation
5
Leonardo AI
Generate visuals for reports and presentations.
6
Google Studio AI
Powerful tool for knowledge management, workflow
optimization, and decision support.
7
Napkin
Amazingly quick generator for infograms i.e.
flowcharts and process

17. Live Demo Session
ChatGPT
Generate content and
simplify tasks.
eg. Drafting an official
letter.
Napkin
Create quick visual aids.
eg. Infogram
Gamma
Generate presentations
Leonardo
Image Generator based
on text and uploaded
image
Perplexity
Research a topic and summarize the points for a presentation.
Google AI Studio
Build prediction models for healthcare data.

18. Pitfalls of AI : Be careful!
Hallucinations & Errors
AI might generate inaccurate information or
misinterpret processes.
Ex: Providing incorrect instructions for a task.
Bias Amplification
Reflects biases in training data, skewing results or
decisions.
Ex: Favoring certain demographics in
recommendations.
Context Blindness
Overlooks specific details or situational factors.
Ex: Offering generic suggestions that ignore unique user
needs.
Overreliance Risk
Users might depend too heavily on AI, bypassing critical
thinking.
Ex: Accepting automated predictions without double-
checking.
Lack of Transparency
Unclear how AI arrives at its conclusions or
suggestions.
Ex: Not explaining why a particular option was
prioritized.
Data Staleness/Overload
Relies on outdated information or struggles with
excessive input.
Ex: Failing to incorporate the latest industry standards.

19. Ensuring AI Reliability
Cross-Check Sources
Verify AI outputs with clinical
guidelines (e.g., UpToDate,
PubMed) to ensure evidence-
based care.
Ask Specific Questions
Query AI with patient details
(e.g., <Antibiotics for 60yo with
COPD?=) to reduce vague or
irrelevant output.
Test Consistency
Rephrase medical queries for
reliability. Example: <Best
asthma treatment?= vs. <Asthma
management options?=
Demand Evidence
Ask AI for supporting studies or reasoning (e.g., <Cite
trial data=) to validate clinical relevance.
Use Multiple Tools
Compare AI outputs from different platforms (e.g.,
ChatGPT vs. medical-specific AI) to spot potential
errors.

20. Summary & Takeaways
1
Key Benefits
AI enhances productivity.
2
Practical Applications
Apply AI to your daily tasks.
3
Pitfalls of AI
Be aware of its limitations. Always cross-check
especially facts.

21. Q&A Session & Closing

22. AI Adoption in Healthcare
40%
Increase in efficiency.
25%
Reduction in errors.
30%
Cost savings.