On Completion

By the end of the course, you will be able to:

• Leverage Better Quality Output using prompt engineering
• Structure Prompts Effectively
• Guide AI Systems by incorporating examples, instructions & context
• Employ Techniques like specificity, plain language & affirmative voice.
• Utilise Advanced Skills like list prompts, command chains, clarifying questions.
• Troubleshoot Issues like avoiding AI bias, overpromising, and hallucination.
• Apply Prompt Engineering across use cases – content writing, coding, customer service, and more.

01: Intro to Prompt Engineering

• This lesson covers structuring prompts properly so AI systems understand your intent:

• What prompts are & example usage
• Why prompt engineering is a crucial skill
• Core concepts
• Best practices for writing clear, coherent prompts
• Common mistakes to avoid
• Ethical considerations
• Overview of the prompt engineering process

02: Prompt Structure Fundamentals

• This lesson covers structuring prompts properly so AI systems understand your intent:

• Organising prompts into logical sections
• Writing clear topics & formatting line breaks
• Using concise component with bullet points etc
• Establishing length & detail for different AI tools
• Balancing brevity with sufficient context
• Conversational tone vs. impartial commands
• Punctuation & grammar tips

03: Using Clear Language & Formatting

• Clarity in language and formatting prompts for scannability optimises AI comprehension:

• Formatting prompts for easy scanning with line breaks and bullet points
• Writing in simple, unambiguous language the AI will understand
• Avoiding abbreviations, acronyms, idioms and phrases that may confuse the AI
• Tips for enforcing tone and formatting through punctuation
• Capitalising proper nouns and titles
• Establishing the perspective you want the AI to write from

04: Adding Context & Instructions

• Providing adequate context and clear instructions are key to guiding the AI:

• Frontloading prompts with background details for the AI
• Explaining the desired objectives and intent
• Specifying any requirements, constraints or parameters
• Defining the target audience, output format and tone
• Giving clear instructions to the AI using imperative language
• Ensuring the context and instructions align
• Quantifying requirements like word count, number of examples needed, etc.

05: Writing Prompts for Content Creation

• Crafting effective prompts for content generation requires nuance.

• Choosing types of content like blog posts, social media captions, scripts
• Establishing topic, audience, tone, length, format and purpose
• Using examples of successful content and explaining why they work
• Explaining the goals and desired outcome for the content
• Supporting creativity while controlling quality and accuracy

06: Optimising for Conversational Chatbots

• Chatbots require conversation-optimised prompts.

• Formatting prompts as a hypothetical conversation
• Writing natural-sounding questions to elicit helpful responses
• Thinking of follow-up questions the user may have
• Predicting what clarifying information the user might request
• Guiding tone, personality and voice with conversational cues
• Controlling length & complexity appropriate for back-and-forth exchanges
• Enabling contextual recommendations based on user intent

07: Code Documentation Prompts

• Documenting code requires explaining complexity simply.

• Clarifying programming language, coding style, tooling used
• Providing code snippets and explaining desired documentation format
• Articulating target level of detail for coder vs. non-coder audiences
• Instructing the AI to focus on logic flow, dependencies, usage
• Protecting against incorrect assumptions about code functionality
• Writing concise, scannable code documentation
• Testing documentation prompts on sample functions

08: Customer Service Prompts

• Service prompts must handle diverse customer issues tactfully.

• Anticipating common customer problems and optimal responses
• Gathering context like customer industry, location, prior issues
• Outlining the ideal tone – friendly, understanding, reassuring
• Explaining constraints like availability, costs, eligibility
• Writing ideal sample conversations between customer and AI
• Teaching nuanced social tactics like empathy, humor, positive framing
• Avoiding overpromising capabilities when uncertain

09: Creative Content with Dall-E

• DALL-E requires imaginative, descriptive prompting.

• Choosing compelling image concepts that interest users
• Writing detailed descriptions of objects, scenes, styles, colors
• Citing examples of desired aesthetics, composition, and mood
• Balancing creative freedom vs. guardrails to prevent incoherence
• Iterating on prompts based on initial image outputs
• Curating image datasets to train AI on desired themes
• Ethical considerations of AI art generation

10: Advanced Prompt Components

• Take your prompts to the next level using advanced tactics like:

• Example prompts showing ideal responses for the AI to model
• Follow-up questions to gather missing info from the user
• Multiple choice options to narrow down user intent
• Clarification of ambiguous terms that might confuse the AI
• “Avoid” instructions telling the AI what not to do
• Back and forth dialogue for complex conversations

11: Chaining Commands in Prompts

• Prompt chaining directs the AI through multi-step tasks:

• Breaking down requests into numbered/ordered action steps
• Identifying dependencies needed for later steps
• Defining variables for the AI to set and reference
• Leveraging output from previous steps in subsequent prompts
• Tracking context throughout conversations with users
• Re-prompting when output is inadequate or incorrect
• Maintaining consistent narrative voice across prompts

12: Troubleshooting Bad Responses

• Debugging prompts based on undesired AI output:

• Identifying if poor responses stem from flawed prompts or AI limitations
• Testing prompts interactively to isolate issues
• Simplifying prompts by removing ambiguous or contradictory instructions
• Explaining domain-specific terminology the AI may not grasp
• Adding examples for context if AI seems ignorant of basics
• Rewriting confusing language, vague instructions, overly complex prompts
• Adjusting tone if AI output seems too informal or disjointed

13: Mitigating AI Bias in Prompts

• Crafting inclusive prompts by:

• Removing biased wording, stereotypes, or assumptions
• Expanding context to avoid narrow interpretations by the AI
• Seeking diverse perspectives when providing examples to the AI
• Reinforcing values of equity and inclusion
• Avoiding charged political topics that lack impartial facts
• Testing prompts from the perspective of underrepresented groups
• Monitoring AI behavior continuously for emerging biases

14: Controlling Tone, Voice & Personality

• Tuning AI tone and voice through:

• Articulating target communication style – professional, casual, funny
• Using examples to model desired tone and personality
• Explaining preferences for word choice, sentence structure, figures of speech
• Outlining etiquette guidelines like avoiding profanity, rudeness
• Testing to ensure consistency across prompt interactions
• Reining in undesired tone shifts or eccentricities
• Balancing creativity vs. predictability in voice

15: Writing in Affirmative Voice

• Affirmative voice prompts increase AI decisiveness:

• Framing prompts as direct, unambiguous statements
• Avoiding equivocal phrasing like “maybe”, “perhaps”, “possibly”
• Cutting redundant phrases that obscure meaning
• Using decisive language like “Generate”, “Create”, “Summarise”
• Quantifying requirements rather than leaving them open-ended
• Reducing broad or subjective terminology that leads to confusion
• Re-prompting in affirmative voice if AI responds tentatively

16: Using Clarifying Questions

• Improving prompts through follow-up questions:

• Identifying missing context needed to answer original prompt
• Composing potential follow-up questions the user might ask
• Grouping related questions together in sections
• Quantifying questions whenever possible for decisive AI responses
• Prioritising questions to streamline pathways to resolution
• Phrasing questions conversationally while maintaining clarity
• Enumerating complex questions to simplify AI processing

17: Prompt Engineering for Emails & Letters

• Email and letter prompts require proper tone, format, and voice:

• Defining recipient and establishing appropriate tone
• Structuring prompts like a professional email draft
• Following expected email/letter formatting conventions
• Writing clear subject lines succinctly summarising intent
• Composing realistic greetings, introductions and sign-offs
• Quantifying expected length to prevent overly long responses
• Testing email prompts by sending drafts to yourself

18: Prompts for Long-Form Content

• Optimising prompts for long-form content like:

• Providing overview of topic, objective, and outline
• Breaking down into sections and ordering logically
• Defining required components like headings, paragraphs, visuals
• Setting length expectations section-by-section
• Maintaining consistent voice and tone throughout
• Grouping related instructions to prevent disjointedness
• Allowing flexibility while guiding structure

19: Optimising for Readability & Flow

• This lesson covers enhancing prompt responses through readability best practices:

• Establishing target reading level using grade level or age
• Limiting sentence/paragraph length for better comprehension
• Varying sentence structure to avoid repetition
• Improving flow with transitions between paragraphs
• Breaking up dense text with section headings
• Balancing brevity with adequate explanations
• Reviewing responses to identify areas needing clarification

20: Planning Structure for Complex Requests

• Approaching multifaceted prompts by:

• Outlining all required components before writing prompt
• Prioritising components based on dependencies
• Grouping related instructions to build on previous context
• Quantifying expectations like word count for each section
• Maintaining consistent voice, tone and terminology
• Indicating sequence for components if order matters
• Allowing flexibility where order doesn’t affect output

21: Thinking through possible AI failures

• Mitigating risks of incorrect or inadequate responses by:

• Considering areas where AI might lack necessary context
• Predicting edge cases that prompt didn’t account for
• Identifying terms or concepts AI may misunderstand
• Drafting potential follow-up clarifying questions
• Adding examples illustrating desired response if prompts are ambiguous
• Explaining safety considerations if AI response could cause harm

22: Balancing Specificity vs Creativity

• This lesson examines the tradeoffs between guidance and originality:

• Evaluating whether to constrain AI creatively or not
• Adding guardrails to prevent completely off-topic responses
• Quantifying expectations like number of ideas to balance flexibility
• Using a few examples for inspiration while allowing new perspectives
• When to let AI riff creatively vs. controlling output more tightly

23: Testing & Iterating prompts

• Methods for refining prompts through experimentation:

• Identifying key variables in the prompt to A/B test
• Trying slight variations in wording, tone, or structure
• Testing on smaller examples to pinpoint issue areas
• Evaluating which components add value vs. create confusion
• Documenting learnings after each iteration to optimise future prompts

24: Documenting Prompts for Future Use

• This lesson explores best practices for documenting quality prompts:

• Cataloging prompts in a centralised, searchable database
• Tagging prompts by use case, industry, content type, etc.
• Recording optimal parameters like length, specificity, examples
• Noting lessons learned from iterations and AI errors
• Sharing prompts across teams to avoid duplication of efforts
• Building libraries of prompts over time for consistency

25: Identifying New Applications for Prompting

• Finding new use cases by:

• Observing pain points in daily workflows to prompt solutions for
• Looking for repetitive or rules-based tasks that AI could assist with
• Testing tools against existing processes to identify improvement areas
• Encouraging teams to ideate promptable concepts during brainstorms
• Starting with simple content prompts then expanding scope over time

26: Building a Library of Quality Responses

• This lesson covers compiling AI responses to benchmark against:

• Cataloging well-written passages covering various topics
• Curating examples of creative output like images, videos, code
• Gathering conversational dialogues showing humanlike exchanges
• Saving responses addressing edge cases for future reference
• Building corpus covering different formats like emails, reports
• Using library to set standards of quality when testing prompts

27: Streaming Workflows with Prompt Templates

• Prompt templates accelerate creating prompts:

• Developing a menu of templates categorised by use case
• Pre-populating templates with common instructions
• Creating blanks for users to quickly fill in context
• Testing templates with target users and refining based on feedback
• Versioning templates to improve while maintaining consistency

28: Collaborating with Others on Prompts

• Cross-team prompt collaboration enables sharing best practices:

• Identifying team members with complementary prompt engineering skills
• Co-editing prompts using cloud-based shared docs
• Assigning team members specific prompt sections based on strengths
• Reviewing prompts together to improve coherence
• Comparing prompts side-by-side to identify optimal phrasing

29: Prompt Engineering Ethics & Pitfalls

• This lesson explores ethical considerations:

• Mitigating bias through expanded context and testing
• Ensuring transparency by documenting prompt process
• Explaining AI limitations to set proper expectations
• Reporting harmful content prompts may generate
• Establishing approval processes for high-risk prompts
• Seeking diverse perspectives when creating prompts

30: The Future of Prompt Engineering

• The course concludes by looking ahead at the evolution of prompt engineering:

• How advanced AI capabilities like memory will impact prompts
• New applications of prompting as technology develops
• Building scalable template libraries as more teams adopt prompting
• Streamlining collaboration as tooling improves
• The importance of transparency and ethics as reliance on AI increases
• Preparing for paradigm shifts in how humans interact with AI