Course outline
Module 1
Module 1: Foundations of AI-Assisted Writing
Foundations of AI-Assisted Writing
Knowledge check
Q1. What is the most accurate description of how a large language model generates text?
- A. It searches a database of pre-written sentences and copies the closest match
- B. It predicts the most likely next word (token) based on patterns learned from large amounts of text ✓
- C. It understands the meaning of a request and then looks up verified facts for each claim
- D. It follows a fixed set of hand-written grammar rules to assemble sentences
Why: LLMs generate text by predicting likely next tokens given the preceding context, using patterns learned during training. They do not search a database of sentences, nor do they follow hand-written grammar rules or verify facts on each claim.
Q2. Why is a clear, specific prompt important when using AI to draft a business email?
- A. Longer prompts always cost less money to process
- B. Specific details about audience, purpose, and tone help the AI produce output closer to what you need ✓
- C. AI tools only respond to prompts written in formal English
- D. Specific prompts prevent the AI from making any factual errors
Why: The model has no access to your context unless you provide it. Stating the audience, purpose, tone, and key facts gives the model what it needs to produce a usable draft. Specificity does not guarantee accuracy, and language formality is not a requirement.
Q3. Which of the following is the best practice before sending an AI-generated report to a client?
- A. Send it immediately, since AI output is generally accurate
- B. Remove any sections that sound too formal
- C. Review and verify facts, figures, and claims, and adjust it to fit your voice and context ✓
- D. Ask the AI to confirm that its own content is correct and then send it
Why: AI can produce plausible-sounding errors, including invented statistics or citations. A human should verify facts and tailor the content. Asking the AI to check itself is not reliable verification.
Q4. What does it mean when an AI model 'hallucinates' in writing tasks?
- A. It writes in a creative or poetic style when not asked to
- B. It produces information that sounds confident and plausible but is false or unsupported ✓
- C. It refuses to complete a task because the request is unclear
- D. It repeats the same paragraph multiple times in one response
Why: Hallucination refers to fluent, confident output that is factually wrong or fabricated, such as nonexistent sources or incorrect figures. It is a known limitation that makes verification essential.
Q5. Which prompting approach is most useful when you want AI to match your organization's writing style?
- A. Asking the AI to write as creatively as possible
- B. Providing a short sample of your past writing and describing the desired tone ✓
- C. Using only one-word prompts to keep the output simple
- D. Requesting the longest possible version of the text
Why: Giving examples of your existing writing and describing the tone helps the model imitate your style. Creativity requests, one-word prompts, and length requests do not target style matching.
Module 2
Module 2: Crafting Clear and Effective Emails
Crafting Clear and Effective Emails
Knowledge check
Q1. Which subject line is most effective for a clear, professional email requesting a meeting?
- A. Hey, quick thing
- B. Request: 30-minute meeting on Q3 budget review, Tuesday ✓
- C. Important!!!
- D. Read this now
Why: A specific subject line that names the purpose, length, and timing helps the recipient understand the email's content and priority at a glance, increasing the chance it gets opened and handled promptly.
Q2. What is the best opening sentence for a professional email to a client you have not contacted before?
- A. Yo, hope you're doing well lol.
- B. I hope this email finds you well. My name is Dana Lee, and I'm the account manager at Brightline Solutions. ✓
- C. Your company owes us money and we need to talk.
- D. Long time no see, as always!
Why: A courteous opening that introduces yourself and your role establishes credibility and context right away. The other options are either too casual for a first contact or inappropriate in tone.
Q3. When using AI to draft an email, what should you do before sending it?
- A. Send it immediately since AI is always accurate
- B. Review it for accuracy, tone, and facts, then edit it to match your voice and situation ✓
- C. Remove all paragraphs and send only the subject line
- D. Ask the AI to send it on your behalf without checking
Why: AI drafts can contain factual errors, mismatched tone, or generic phrasing. Reviewing and personalizing the draft ensures accuracy and that the message reflects your intent and relationship with the recipient.
Q4. Which email structure is generally most effective for busy readers?
- A. A long narrative beginning with background history before stating the request
- B. Key request or main point first, followed by supporting details and a clear next step ✓
- C. Multiple unrelated topics mixed together in one paragraph
- D. The main request placed in the final sentence after several paragraphs
Why: Leading with the main point, then supporting details and a clear call to action, respects the reader's time and makes it easy to understand what is needed and what to do next.
Q5. Which closing line best ends a professional email requesting a response?
- A. Whatever, do what you want.
- B. Let me know if this works for you by Friday, so I can confirm the schedule. Thanks for your time. ✓
- C. Bye.
- D. If you feel like it, maybe reply someday.
Why: A clear closing that specifies a deadline and the reason for it sets expectations and makes a response easier. Polite thanks at the end reinforces a professional tone, while the other options are dismissive or vague.
Module 3
Module 3: Prompting AI for Professional Reports
Prompting AI for Professional Reports
Knowledge check
Q1. Which element most improves a prompt asking AI to draft a professional report?
- A. Asking the AI to 'make it good'
- B. Specifying the audience, purpose, length, and required sections ✓
- C. Using as few words as possible
- D. Leaving the format open so the AI can decide
Why: Clear context about the audience, purpose, length, and structure gives the AI the constraints it needs to produce a report that fits the real use case, rather than a generic draft.
Q2. When prompting AI for a professional report, what is the best way to handle data the AI does not have access to?
- A. Ask the AI to invent realistic figures so the report looks complete
- B. Provide the data directly in the prompt or ask the AI to leave clearly marked placeholders ✓
- C. Assume the AI will look up the latest statistics automatically
- D. Remove all numbers from the report entirely
Why: Supplying the actual data or using placeholders keeps the report accurate. Invented figures are a serious risk in professional writing because they can be mistaken for real facts.
Q3. Which prompt structure is most effective for a report summary?
- A. Summarize this report.
- B. Write a summary of this report.
- C. Summarize the attached quarterly sales report in three bullet points for a non-technical executive, focusing on revenue trends and key risks. ✓
- D. Make a summary.
Why: Option C specifies the source, format, length, audience, and focus areas. These details let the AI tailor the output rather than guessing what the user wants.
Q4. Why is it useful to ask AI to produce a report in sections, such as Executive Summary, Findings, and Recommendations?
- A. It makes the AI respond faster
- B. It mirrors professional report conventions and makes the output easier to review and edit ✓
- C. It prevents the AI from making any errors
- D. It removes the need for a human to read the output
Why: Requesting a defined structure matches what readers expect from professional reports and makes each part easier to check, revise, and reuse. It does not guarantee accuracy, so human review is still required.
Q5. After AI generates a professional report draft, what should the writer do next?
- A. Send it immediately since AI writing is always accurate
- B. Review it for factual accuracy, tone, and fit with the audience, then revise as needed ✓
- C. Ask the AI to rewrite it without reading it first
- D. Delete any section that contains numbers
Why: AI drafts can contain errors, unsupported claims, or mismatched tone. Verifying facts and adjusting the content for the intended audience is an essential step before sharing a professional report.
Module 4
Module 4: Structuring and Editing Long-Form Documents
Structuring and Editing Long-Form Documents
Knowledge check
Q1. When using AI to help structure a long-form report, what is the most useful first step?
- A. Ask the AI to write the entire report in one prompt without any guidance
- B. Provide the AI with your purpose, audience, and key points, then ask for an outline
- C. Copy and paste the final conclusion and ask the AI to fill in the rest
- D. Request a single paragraph summary before defining any goals
Q2. Which section of a long-form document typically helps the reader quickly understand the main findings before reading in detail?
- A. Appendix
- B. Bibliography
- C. Executive summary
- D. Footnotes
Q3. When editing an AI-generated draft of a long document, what should you check first?
- A. Whether the font choice matches the company brand
- B. Whether the content is accurate, relevant, and supported by reliable facts
- C. Whether the word count is exactly 2,000 words
- D. Whether the document uses the most complex vocabulary available
Q4. Why is it useful to ask an AI to create transitions between sections of a long document?
- A. Transitions increase the total word count for the report
- B. Transitions help the reader follow the logical flow from one idea to the next
- C. Transitions replace the need for headings and subheadings
- D. Transitions are required by all publishing platforms
Q5. What is a good way to ensure consistency in tone and terminology across a long AI-assisted document?
- A. Generate each section in a separate chat with no shared instructions
- B. Mix formal and casual language randomly to keep readers engaged
- C. Create a style guide or list of key terms and include it in your prompts
- D. Avoid reviewing the document until the final section is complete
Module 5
Module 5: Generating and Refining Web Content
Generating and Refining Web Content
Knowledge check
Q1. What is the most effective first step when using AI to generate web content for a new product page?
- A. Publish the AI output immediately to save time
- B. Provide the AI with a clear brief including audience, purpose, tone, and key details ✓
- C. Ask the AI to write the page without any additional instructions
- D. Copy content from competitor websites and ask the AI to rewrite it
Why: AI produces the most relevant web content when given a clear brief covering the target audience, the page's purpose, the desired tone, and essential product details. Vague prompts tend to yield generic output that needs heavy rework.
Q2. Why should you always review AI-generated web content before publishing it?
- A. AI tools cannot produce any readable text at all
- B. AI may produce inaccurate facts, miss brand voice, or include unverified claims ✓
- C. Search engines penalize any content that was written with the help of a computer
- D. Review is only needed for content longer than 5,000 words
Why: AI can generate plausible-sounding but incorrect information, overstate claims, or drift from brand voice. Human review ensures accuracy, compliance, and consistency before the content reaches the public.
Q3. Which prompt is most likely to produce useful refinement of an existing web paragraph?
- A. Make it better.
- B. Rewrite this paragraph to be shorter, use plain language for a non-technical audience, and keep the call to action at the end. ✓
- C. Write something new about this topic.
- D. Translate this into French.
Why: Specific refinement instructions define the length, reading level, and elements to keep. Vague requests like 'make it better' leave the AI guessing what to change, often resulting in unnecessary or unwanted edits.
Q4. What is a good practice when using AI to write SEO-focused web content?
- A. Repeat the target keyword as many times as possible in every sentence
- B. Ask the AI to naturally include the primary keyword while prioritizing readability for human visitors ✓
- C. Avoid mentioning the keyword at all so the content seems natural
- D. Insert hidden keywords in white text on the page
Why: Good SEO content balances keyword use with readability. Keyword stuffing harms user experience and can be penalized by search engines, while hidden text is a deceptive practice that violates search guidelines.
Q5. What information should you give an AI tool to help it match your organization's brand voice?
- A. Only the company's founding year
- B. Examples of existing approved content plus a description of tone, vocabulary preferences, and terms to avoid ✓
- C. The CEO's personal social media handles
- D. Nothing, since AI always uses the same neutral voice
Why: Providing sample approved content along with descriptions of tone, preferred vocabulary, and terms to avoid gives the AI concrete reference points. This results in output that aligns far more closely with the established brand voice.
Module 6
Module 6: Tone, Voice and Audience Adaptation
Tone, Voice and Audience Adaptation
Knowledge check
Q1. Which tone is most appropriate when writing a follow-up email to a client who has not responded to your last two messages?
- A. Angry and demanding, to pressure a quick reply
- B. Polite, concise, and professional, with a clear next step
- C. Overly apologetic, apologizing for bothering them
- D. Casual and joking, to lighten the mood
Q2. When you ask an AI tool to rewrite a report for a non-technical executive audience, what should you specify?
- A. Only the word count
- B. The audience, their background knowledge, and the purpose of the report
- C. The font and formatting style only
- D. Nothing, since AI adapts automatically
Q3. What does 'voice' refer to in writing, as distinct from 'tone'?
- A. The speed at which text is read aloud
- B. The consistent personality and perspective of the writer across pieces
- C. The emotional mood of a single message
- D. The language the text is written in
Q4. Which prompt is most likely to produce content with a consistent brand voice?
- A. Write something nice about our company.
- B. Write a blog post. Make it good.
- C. Write a blog post in a warm, confident, plain-language voice, avoiding jargon, for small business owners, using the examples in the attached style guide.
- D. Write whatever sounds best.
Q5. A message must deliver bad news to a customer whose order was delayed. What is the best approach?
- A. Hide the delay and mention it only if asked
- B. Open with a clear acknowledgment of the delay, explain the reason briefly, and offer a concrete next step
- C. Bury the news in the middle of a long message about company achievements
- D. Blame the customer for not ordering earlier
Module 7
Module 7: Ethics, Accuracy and Final Review
Ethics, Accuracy and Final Review
Knowledge check
Q1. Why should you verify facts and statistics that an AI tool includes in a business report?
- A. AI tools always produce perfectly accurate data, so verification is unnecessary
- B. AI tools can generate plausible-sounding but incorrect or fabricated information, so facts must be checked against reliable sources
- C. Verification is only required for academic writing, not business documents
- D. AI tools store all facts in a central database that updates in real time
Q2. Which of the following is the most ethical way to use AI when writing a work report that your manager will read?
- A. Submit the AI output without changes and do not mention AI was used
- B. Review, edit, and fact-check the AI draft, and disclose AI assistance if your organization requires it
- C. Use AI to write the report and claim full authorship since you entered the prompt
- D. Avoid using AI entirely because any use is considered cheating in all workplaces
Q3. What is a sensible practice when AI-generated content includes a quote attributed to a named expert?
- A. Publish the quote, since AI tools only quote real people
- B. Remove the attribution and keep the quote as general knowledge
- C. Locate the original source to confirm the quote exists and is worded accurately before using it
- D. Assume the quote is correct if it sounds professional and matches your argument
Q4. Which type of information should you be most careful about sharing in an AI prompt?
- A. General questions about grammar rules
- B. Publicly available industry trends
- C. Confidential customer data, passwords, or private company information
- D. Requests for email subject line ideas
Q5. Why is it important to check AI-written content for bias?
- A. AI outputs are always neutral because they are generated by software
- B. AI models learn from existing data, which may contain biases that can appear in the output and affect readers unfairly
- C. Bias only matters in creative writing, not in emails or reports
- D. Checking for bias is only needed when writing for government agencies