Skip to main content

Prompt Engineering 101

K
Written by Kay Chen

Welcome to Prompt Engineering 101! This session is designed to help you become confident and effective in using CoAuthor AI for document analysis, summarization, and content generation. Whether you're new to AI tools or looking to refine your prompting skills, this training will give you practical strategies to get the most out of CoAuthor.

Prompt Engineering 101: Expanded Agenda

Welcome & Introduction

  • Meet your host: Dan Schindler, Technical Account Manager

  • Overview of CoAuthor AI and its role in your workflow

  • Addressing common concerns about AI tools


1. Understanding CoAuthor AI

  • What CoAuthor is and how it differs from tools like ChatGPT

  • Data privacy and source control in CoAuthor

  • How CoAuthor uses attached sources for focused responses


2. Everyday AI Examples

  • Spell checker in Word using machine learning

  • Sentence completion in Outlook and smartphones

  • Drawing parallels to CoAuthor’s backend technology


3. Getting Started with Prompts

  • Introduction to summarization as a first step

  • Pitfalls of general summarization

  • Live Example: Summarizing a full document vs. targeted sections


4. Enhancing Prompt Quality

  • Importance of context: document sections vs. whole documents

  • Structuring prompts:
    Context → Task → Formatting

  • Live Example:
    Prompt with context as a medical writer → focused summary with methodology and patient count


5. Formatting Techniques

  • Using phrases like “be concise,” “answer with a paragraph,” “convert decimals to percentages”

  • Leveraging repetition to reinforce instructions

  • Live Example:
    Comparing outputs with and without formatting constraints (e.g., 100-word paragraph)


6. Avoiding Hallucinations

  • What hallucinations are and why they happen

  • Strategies to reduce them:

    • Attach only relevant sections

    • Use strict formatting instructions

    • Repeat constraints clearly

  • Live Example:
    Prompt causing hallucination vs. corrected version with section filtering


7. Reusing Prompts Across Projects

  • Generalizing prompts for repeat use

  • Saving prompts by sponsor or project

  • Ensuring consistency across similar documents


8. Best Practices

  • Make small changes to refine prompts

  • Always provide context and specify exclusions

  • Use formatting and repetition to guide responses

  • Live Example:
    Adjusting prompt to include missing data (e.g., number of patients)


9. Support & Resources

  • Using the support sandbox for safe experimentation

  • Contacting AI Support for troubleshooting

  • Link to CoAuthor FAQ and documentation


10. Q&A and Next Steps

  • Open floor for questions

  • Preview of future sessions (e.g., Prompt Engineering 201)

  • Encouragement to explore and test prompts

Did this answer your question?