Opening Case: The Statistic That Wasn’t

Noa Villaseñor-Katz read the headline twice: Bridger Valley Provisions: Mountain West’s Fastest-Growing Organic Jerky Brand. The trade journal’s website loaded the rest of the article, and there it was in the second paragraph — “with 340% year-over-year growth across the region.” She had never seen that number in any of the client’s quarterly reports because it didn’t exist.

Her phone sat at the edge of her desk, still warm. Two minutes ago, Margaux Prentiss, CEO of Bridger Valley Provisions, had called with the measured calm of someone choosing her words carefully. “I’m sitting across from my CFO right now, Noa. He’s asking me where this growth figure came from. I’d like to give him an answer that doesn’t make both of us look foolish.”

Noa had promised to call back within the hour. She hadn’t promised an explanation, because she didn’t have one yet — though she had a suspicion.

Meridian & Lark Creative occupied the top two floors of a renovated grain elevator on East Main Street in Bozeman, Montana. Noa had founded the agency a decade ago after leaving a corporate communications firm in Denver, drawn to the idea of building something smaller and more intentional. The elevator’s original wooden beams still ran across the ceilings, and on clear days the fourth-floor conference room framed the Bridger Range through a wall of windows that Noa’s architect had called “aggressively optimistic.” For ten years, the view had matched the agency’s trajectory — steady growth built on meticulous copy, personal client relationships, and a reputation for getting the details right.

That reputation was the reason the phone call stung. Meridian & Lark didn’t make factual errors. Or they hadn’t, until six weeks ago, when Noa made the decision that now seemed to be unraveling on her screen.

The catalyst had been losing the Gallatin Outfitters account. Gallatin was their second-largest client — a Bozeman outdoor gear company that had been with them since year three. When the contract came up for renewal, a larger agency out of Missoula had pitched against them and won. The Missoula team had produced twice the campaign concepts in half the time, complete with polished social media calendars, SEO-optimized blog drafts, and A/B-tested ad copy. Their secret was no secret at all: they had integrated AI writing tools into every stage of their workflow.

Noa had watched the industry shift for over a year. Clients weren’t just accepting AI-assisted work — they were expecting it. Speed had become a proxy for competence, and agencies that couldn’t match the pace were losing bids. So she had done what she thought a responsible leader would do: she called an all-hands meeting and announced that everyone at Meridian & Lark would integrate AI writing applications into their workflow within sixty days. She framed it as evolution, not replacement. “The writing is still yours,” she told the room. “AI just gets you to the starting line faster.”

The reactions had split along predictable lines.

Reeve Thalberg, her senior copywriter, had crossed his arms and said nothing during the meeting — which, for a man who’d spent twenty years in newspaper journalism before joining the agency, was louder than any objection. At fifty-one, Reeve had built his career on the conviction that good writing came from reporting: interviewing sources, verifying claims, choosing each word with the precision of someone who’d once had editors red-pen his copy at the Billings Gazette. After the meeting, he stopped by Noa’s office and said, “You’re asking me to co-author everything I write with a machine that makes things up for a living.” He had been using AI only for the most basic tasks since — reformatting bullet points, generating meeting agenda templates — and kept it away from anything that carried a client’s name.

Brin Okafor had been the opposite. Hired six months ago as a junior account coordinator, Brin was twenty-four, a recent graduate from the University of Montana’s business program, and fluent in every AI tool Noa had heard of and several she hadn’t. He embraced the mandate with enthusiasm that Noa initially found encouraging. He was producing drafts faster than anyone on the team, handling three accounts with the output of someone managing one, and his early work had been solid enough that Noa hadn’t looked closely at his process.

She was looking now.

Suki Amundsen, the agency’s media strategist, had taken what she called the “Minnesota approach” — cautious, practical, and unlikely to generate a headline. At thirty-seven, Suki had been at the agency five years and handled media placements, press strategy, and client communications planning. She had started using AI quietly for internal work — brainstorming pitch angles, summarizing media coverage reports, drafting talking points that she then rewrote by hand. Nothing she produced with AI assistance went to a client without her rewriting it first, and she had been vocal about wanting written guidelines before the sixty-day mandate took full effect.

Tomás Delacroix, the graphic designer and social media manager, had his own version of the problem. At twenty-nine, Tomás had come to Bozeman from Lafayette, Louisiana, bringing a designer’s eye and a sharp sense of humor that showed up in the agency’s social campaigns. He had started using AI image generators for mood boards and initial concept work, and the results had been impressive — until a client asked, during a campaign review, whether the hero images in a social media concept were “real photographs or that AI stuff.” Tomás hadn’t known what to say. The images were AI-generated, but the client’s contract said nothing about it, and the agency had no policy on disclosure.

Now, Noa pulled up the press release that had started the fire. Brin had written it — or rather, he had prompted ChatGPT to write it. She could see the prompt in the shared workspace: Write a press release announcing Bridger Valley Provisions’ growth in the organic jerky market in the Mountain West region. Make it sound professional and newsworthy. The AI had delivered exactly what was asked for: a professional-sounding press release with a fabricated statistic that made the client sound like a market leader. Brin had reviewed it for tone and grammar, added the client’s contact information, and sent it to the media list. He hadn’t checked the growth figure because it looked right. It had the confident precision of a real data point — 340%, not a round number, not an obvious guess. That was what made hallucinations dangerous: they didn’t announce themselves.

Noa closed the laptop and looked out at the Bridger Range. She had four people on her team, four different relationships with AI, and a client waiting for a call she wasn’t ready to make. She had two weeks before the quarterly client review, and she needed something she should have created before the sixty-day mandate: a policy. Not just rules about what tools to use, but a framework for thinking about when AI helped and when it hurt — when it made their work faster and when it made their work wrong.

She picked up her phone and typed a message to the team Slack channel: Monday, 9 AM, fourth floor. One agenda item. Bring your questions, because I’ve got more of them than answers.

Getting Started

20.1 Understanding AI in Business Communication

Introduction: Why AI Literacy Matters

You’ve probably already encountered opportunities — or mandates — to use AI in your professional communication. So how do you decide what AI to use and how to use it? What guardrails should you put in place to ensure you’re being ethical and responsible when your job, your relationships, and your personal and company reputations are at stake?

This section introduces a framework for AI literacy that can help you make informed, strategic, ethical, and defensible decisions about when and how to use AI to craft business messages.^[1] As you develop your AI literacy, you’ll also want to consider ethical authorship — a complementary concept that focuses on producing AI-assisted content that is transparent, accurate, audience-centered, and reflective of your values and integrity as a communicator.^[2] Together, AI literacy and ethical authorship guide you in using AIWAs effectively and responsibly in your business communication.

Understanding AI Writing Applications (AIWAs)

A variety of AIWAs are available — free or by subscription — to generate text and images, proofread content, help you research, and more. Your first task in managing AI effectively (and building AI literacy) in your business communication is understanding how to use AIWAs and choosing tools that match your communication tasks. By understanding these tools, you can better select what type of assistance you need to improve your communication while staying within your organization’s ethical and legal boundaries.

Natural Language Processing: Natural Language Processing (NLP) is a branch of artificial intelligence focused on enabling machines to understand, interpret, and generate human language. It includes tasks like text classification, language translation, sentiment analysis, and more.

Large Language Model: A large language model (LLM) is a type of artificial intelligence that can understand and generate human-like text. It works by analyzing vast amounts of written data it’s given as training material (like books, websites, and articles) to learn patterns in language. This training allows the model to predict what words or sentences are likely to come next based on the input it receives. LLMs, like ChatGPT, can answer questions, write essays, summarize information, and hold conversations — all by recognizing patterns and relationships in language. LLMs don’t think the way humans do, but they can simulate conversation by using the information they’ve learned. These models are called “large” because they have billions of parameters (data points) that help them make more accurate predictions and provide more realistic responses.

Training Materials: Training materials for LLMs refer to the large amounts of text, images, or other data used to teach the AI how to perform tasks like writing, answering questions, or creating art. The process of selecting training materials involves choosing data that represents a wide range of topics and language uses. However, training materials are incomplete and biased because it’s impossible to include all perspectives and topics the AI might need to know, and certain content isn’t digitally available. This means the AI might not understand certain niche topics, cultural nuances, or new developments that weren’t part of its training data. As a result, LLMs may produce incorrect or limited information on topics outside their training.

Generative AI (GenAI): This term refers to a broad category of artificial intelligence that can create content — text, code, or images. LLMs like Gemini and ChatGPT fall under this umbrella.

AI Writing Tools (AIWT): This term encompasses software applications that use features of GenAI to assist with the writing process. While some AI writing tools have basic content generation capabilities, their primary focus is on tasks like grammar checking, paraphrasing, or suggesting sentence improvements. Grammarly, for example, is an AI-powered editing tool, but it isn’t categorized as a pure GenAI tool the way an LLM is.

Prompt Engineering: Prompt engineering is the practice of designing and refining the input (or “prompt”) given to an LLM, like ChatGPT, to get the best possible output or response. A prompt can be a question, instruction, or description. Since models like ChatGPT generate text based on the prompts they receive, the way you phrase or structure your prompt can greatly influence the quality and relevance of the response.

AI Hallucinations: AI hallucinations refer to situations where an AI, like ChatGPT, generates information that sounds convincing but is actually false, made-up, or inaccurate. For example, if you ask ChatGPT for a market analysis, it might generate a report claiming a new product has a 25% market share in Europe, even though it hasn’t launched there. Or you might ask Gemini for a bio for a keynote speaker at your professional conference, and it will include a fake citation for a book the person hasn’t written. This happens because the AI doesn’t truly understand the information; instead, it predicts what words or facts are most likely to follow based on its training data. The AI is making educated guesses based on patterns it has learned, not verifying facts. This is why double-checking AI outputs is critical, especially when dealing with important or specialized topics.

AI Literacy and Ethical Authorship

Lentz (2024) considered three ethical frameworks (Aristotle’s Virtue Ethics, Kant’s Categorical Imperative, and Mill’s Utilitarian Ethics) to create a definition of ethical authorship specifically for business communicators.^[3] She also proposed a series of reflective statements to help guide communicators toward ethical authorship, which we’ll look at in a moment. She defines ethical authorship in the context of AI-assisted writing as:

The authorship of business discourse in ways that positively reflect an author’s values and that create ethical, clear, complete, transparent, and audience-centered communication. In addition, ethical authorship requires that an author be aware of and mitigate the risks of using AI-generated content, including but not limited to the use of AI hallucinations (false data) and the use of copyrighted material. (p. 597)

This definition fits well with the framework we’ll use for developing AI literacy for business communication.^[4] This framework consists of four capabilities: application, authenticity, accountability, and agency (p. 277). Combined, these capabilities make it possible to be an ethical, AI-assisted author of business communication. In other words, AI literacy enables ethical authorship.^[5]

Application

Professionals need to be familiar with AIWAs’ capabilities and limitations, and how to align them with specific tasks.^[6] The widespread use of these applications by college students and professionals suggests they’re relatively easy to use. However, to maximize their effectiveness, professionals must learn how to refine prompts and adjust them for better results (e.g., modifying tone, style, or level of detail).

The proliferation of AIWAs will outpace any textbook publication, so giving you an exhaustive explanation of available AIWAs is impossible. Instead, you’ll need to listen to and read about what’s happening in your industry and explore and experiment with applications. As you do this, consider some guiding questions to improve your application capabilities:^[7]

• Based on their capabilities and limitations, which AIWA should I use?
• What are the best practices for optimizing my use of this AIWA? (e.g., use of commands, prompts, or queries)
• What underlying data set informs the AIWA? What are the strengths and weaknesses of this dataset, and how will they affect the AIWA’s output?

Authenticity

Professionals must prioritize genuine, personalized communication when using AIWAs (Deptula et al., 2024).^{[8] [9] [10]} Despite the growing capabilities of AIWAs, AI-generated messages won’t reflect your unique voice or be tailored to the specific needs of your recipients. AI-mediated communication is seen as less authentic (less sincere and caring) by professionals, though the messages are still seen as professional and achieving their instrumental purpose.^{[11] [12]} This reinforces the need to consider your relational and identity goals when working with AI — not just your instrumental goals. It may not matter to a receiver if AI generates a summary of product reviews for an online boutique, but it will matter and damage relationships in scenarios that carry more significance, such as crisis communication and delivering bad news.^[13]

To focus on producing genuine, human-centered communication, consider the following questions:^{[14] [15]}

• To what degree have I inserted my own voice, personality, and style into the message?
• Does the message meet my identity goals and reflect who I am and want to be as a professional?
• To what degree have I ensured the message focuses on my receiver’s needs and relational goals?
• To what degree have I built trust with my receiver through this message?

Accountability

Professionals must take responsibility for the accuracy and appropriateness of AI-generated or influenced content used in their communication.^{[16] [17] [18] [19]} This includes using AIWAs in a fair and equitable manner and developing strong information literacy (the ability to find, evaluate, and use information effectively).^[20] This means making sure applications are used in ways that treat everyone equally and don’t give one group an unfair advantage over others. Accountable communicators must also ensure content doesn’t reinforce biases or discrimination against certain groups of people, and that all content is verifiable. AI mistakes in your writing tend to be judged more harshly than human mistakes, highlighting how important it is to maintain high standards of reliability in AI-mediated communication.^[21]

To keep your accountability in mind, consider the following reflective questions:^{[22] [23]}

• Have I verified the content of my message as factually correct?
• Is the logic of the message solid and coherent?
• Does the message contain depth? What perspectives may have been left out?
• Do my stakeholders have equal access to the AIWAs I’ve used?
• Is credit and attribution given to the original authors of content in my message when I’m quoting, summarizing, or paraphrasing ideas that aren’t my own?
• Have I respected my organization’s terms for using AI at work?
• Have I checked my communication for bias and corrected the reflection of that bias in my work?
• Have I done all I can to make sure I’m not misleading my receivers?

Agency

Retaining control of AI-mediated communication means being the “human-in-the-loop,” ensuring AI is used as a tool to enhance your decision-making, not replace it.^[24] You must remain actively involved and be the ultimate decision-maker in the communication process, even when using AI. An AIWA might suggest or draft something, but you’re the one who reviews it, edits it, and makes sure it fits your purpose.

To draw your attention to maintaining control and making your own choices when using AIWAs, consider the following questions:^{[25] [26]}

• Am I retaining or expanding my personal choices through the use of AIWAs?
• Am I enhancing my knowledge, skills, and human potential while using AIWAs?
• Can I make independent human decisions while using AIWAs?
• Have I considered other choices I could make when I’m tempted to use AIWAs unethically or in ways that aren’t allowed in my workplace?
• Have I made the choice to use AIWAs freely, or because I’m desperate — possibly running out of time or not understanding how to complete a task?

20.2 Skills for Using AIWAs Effectively

The Basics of Prompt Writing

Garbage in, garbage out (GIGO) is a common phrase in technology sectors to emphasize the poor decisions or outputs produced when computers receive poor information and data. GIGO applies to your use of AIWAs in business communication: if you provide an AIWA with a vague, incoherent, or overly simplistic prompt — or if it’s trained on inaccurate, incomplete, or biased information — it won’t produce quality content.^{[27] [28]}

One way to mitigate giving or receiving garbage is to provide better inputs through becoming skilled in prompt engineering. Many methods and frameworks exist for writing prompts for AIWAs, and suggestions continue to be refined through experimentation and practice.^{[29] [30] [31] [32]} Commonly, however, prompt-writing frameworks will direct you to include information about the communication context and an explicit task to accomplish.

The RACE framework (Role, Action, Context, Expectation) is an effective approach for creating prompts that helps you address some of the previously stated issues about AI not understanding your specific communication contexts.^[33] The RACE framework starts by clearly defining the Role, specifying the persona or expertise the AI should adopt — such as a social media marketer or email copywriter. Next, the Action outlines the specific task to be performed, whether it’s crafting content, developing strategies, or analyzing data. The Context provides essential background information that helps an AIWA understand the scope and nuances of the request, ensuring a more tailored response. Finally, the Expectation defines the desired outcome, detailing what success looks like and guiding the AI in delivering a result that meets your communication goals and business objectives. By organizing your prompts with the RACE framework, you can produce content that may require less revision and help you achieve your identity, relational, and instrumental communication goals more efficiently.

Example Prompt using the RACE Framework

• Role: “You are a supply chain expert specializing in sustainable manufacturing processes.”
• Action: “Develop a strategy for sourcing and manufacturing eco-friendly materials for our new line of children’s clothing.”
• Context: “Green Sleeves is a sustainable children’s clothing brand that uses only eco-friendly materials like organic cotton, recycled polyester, and water-based dyes. We aim to reduce our carbon footprint and ensure ethical sourcing while maintaining high-quality, affordable products. The new clothing line should align with our sustainability goals, be scalable for mass production, and appeal to environmentally-conscious consumers.”
• Expectation: “The strategy should include recommendations for sourcing materials from ethical suppliers, options for reducing waste during the manufacturing process, and ideas for using renewable energy in production. Provide an outline for a sustainable supply chain that can meet our growth targets while maintaining Green Sleeves’ commitment to sustainability.”

Other frameworks may suggest providing AIWAs with examples of what a successful output looks like.^[34] Be sure that if you provide AI with sample and model communication, this is within your organization’s policy. In all conversations with AI, prevent privacy breaches by never including personal identifiers (e.g., addresses, social security numbers), financial records (e.g., credit card numbers, bank account numbers), proprietary information (confidential business information such as client lists, business plans, formulas), or other sensitive information (passwords, other private or confidential records) in your prompts.^[35]

To achieve the best results from writing with AI, consider AIWAs as collaborative partners or assistants rather than static tools.^{[36] [37]} By engaging in iterative conversations with AIWAs, you can refine prompts and generated content to better align with your specific needs and goals. Experiment with different phrasings and specifications to yield different outcomes. As you refine your prompts, the AIWA’s ability to generate relevant and tailored content will improve, ultimately enhancing your efficiency and effectiveness.

Using AIWAs in the Business Communication Process

You exist in a world where you’ll always have access to AIWAs, whether in a personal, academic, or professional context. This section explores how AIWAs can support business communication students and professionals in using AI in the writing process and producing specific message types.

Brainstorming

AIWAs can assist business professionals in brainstorming by quickly generating a range of ideas, solutions, and creative concepts based on specific input or prompts, helping spark creativity and expand thinking.^{[38] [39] [40] [41]} By providing an AIWA with a clear context — such as a project goal, target audience, or business challenge — you can receive a variety of suggestions, alternatives, or approaches to consider. This can help you explore different perspectives, speeding up the ideation process and helping identify solutions that might not have been initially obvious.

• Sample Prompt: “As a marketing manager for Green Sleeves, a sustainable children’s clothing brand, I need your help brainstorming creative ways to highlight the eco-friendly materials we use in a new line of rainboots. Specifically, I’m looking for marketing strategies or product features that will engage eco-conscious parents, increase brand awareness, and ultimately drive sales for our line of children’s apparel.”

Researching

AIWAs can help business professionals quickly gather and summarize information from a wide range of sources, such as articles, reports, and industry studies in their training data.^{[42] [43] [44]} You can also provide AIWAs with content to analyze and summarize. AIWAs can save professionals time by pulling out key insights and trends, allowing them to stay up to date without thoroughly reading large amounts of data. However, since AIWAs are susceptible to hallucinations, they might generate false or inaccurate information that sounds convincing, and AIWAs won’t necessarily provide you with sources or correct citations for content that is factual.^{[45] [46]} This can be particularly problematic when AI is used for research that informs business decisions. Always verify facts from reliable sources, especially when using AI-generated content for decision-making or client-facing materials.

• Sample Prompt: “I’m an HR manager at Green Sleeves, a sustainable children’s clothing brand, and I’m looking to improve our employee engagement and retention strategies. Using the HR reports and articles I’ve provided you with, summarize the key trends in employee satisfaction and retention within the eco-conscious sector, and suggest strategies that we can apply to foster a positive, long-term work culture at Green Sleeves.”

Proofreading & Editing

AIWAs can quickly proofread written content for grammar, spelling, punctuation, and readability, helping business professionals ensure their emails, reports, and presentations are polished and professional (Cardon et al., 2023a; Dobrin, 2023).^{[47] [48]} Beyond basic proofreading, AI can suggest edits to improve sentence structure, clarity, and tone, making the content more engaging and easier to understand. However, while AI is effective at spotting errors, it may miss subtleties like contextual mistakes or tone issues, so reviewing AI’s suggestions to ensure the content aligns with the intended message and business context is important.^{[49] [50]} Finally, AI won’t know if you’ve made a mistake in your input (remember GIGO).^[51] For example, if you generate a blog post and have incorrect product information in your input, your output may be grammatically correct but factually inaccurate. Always review your input and AI’s suggestions and make sure they align with the intended message.

• Sample Prompt: “I’ve written a white paper on sustainable clothing fibers for Green Sleeves, and I’d like you to proofread and edit it for clarity, punctuation, grammar, readability, and flow. Focus on ensuring the content is engaging and accessible while maintaining a professional tone. Please also check that all technical terms are correctly explained for a general audience.”

Content Creation

AIWAs can assist with creating content like emails, blog posts, and marketing materials by generating drafts based on specific prompts.^{[52] [53]} This can save time and help maintain consistency across communication. However, AI-generated content should always be reviewed for alignment with your company’s brand and your communication goals. AI-generated content can often be generic or lack a personal touch.^[54] When crafting customer-facing messages, always ensure that the content reflects your brand’s unique voice and values. Relying too heavily on AI for content creation can result in messages that feel robotic or lack emotional engagement. Verify that your input doesn’t contain sensitive information.

Using AI to generate content — or even to revise paragraphs — can inadvertently misrepresent or oversimplify language that reflects your diverse cultural, regional, or social identity.^[55] AI models trained on biased data can also perpetuate stereotypes or generate content that is insensitive, marginalizing certain groups or miscommunicating intended meaning. Remember the AI-literacy capabilities of authenticity and agency — don’t let AI erase who you are.^{[56] [57]} Your customers and coworkers want to work with you, not a machine.

• Sample Prompt: “I’m the social media manager for Green Sleeves, and I need your help writing a series of engaging social media posts to promote our new line of eco-friendly rainboots for children. The boots are made from recycled rubber, organic cotton linings, and water-based dyes, offering both sustainability and durability. One post should focus on how the boots are perfect for rainy days, another on their eco-friendly materials and their impact on the environment, and a third showcasing their fun, colorful designs and how they can be styled for everyday wear. Our corporate voice is kind, warm, and approachable.

Crafting Effective Business Messages with AI

Written Messages

To effectively use AIWAs for crafting routine business messages like emails, letters, memos, and reports, provide clear and concise instructions in your prompt, including specific formatting requirements.^[58] For instance, specify the desired font, font size, line spacing, and margin settings. If you have specific templates or style guides, provide them to the AIWA as reference. This guidance on writing routine business messages can be applied to non-routine documents as well. However, non-routine documents often require more creativity, critical thinking, and a deeper understanding of the specific context.

You can also use AI to generate templates tailored to your specific needs. Provide the AIWA with a brief description of the template, including the message type (email, feasibility report, performance review), content structure, and any specific formatting requirements. The AI can then generate a basic template that you can customize further.

Here are some sample, generic, and brief prompts to help you imagine how you might begin your interaction with an AIWA to produce a message. You should make your prompts more elaborate, following the RACE or another framework for prompt engineering:

Sample Email Prompts:

• “Draft a formal email to a potential client, introducing our new product line and its key benefits. Here is a description of the product and the benefits….”
• “Provide me suggestions on how to improve my persuasive appeals in this email to a team member, encouraging them to attend an upcoming training session.”
• “Compose a concise email to a supplier of [specific product or service], requesting a quote for [specific product or service].”

Sample Business Letter Prompts:

• “How should I format a formal business letter to a potential partner, proposing a collaboration opportunity.”
• “Assemble these project timeline notes into an email for a client, highlighting the successful completion of the project.”
• “Compose a concise business letter to a vendor, requesting a revised invoice with corrected pricing.”

Sample Business Memo Prompts:

• “Draft a memo to the marketing team, outlining these key strategies for the upcoming product launch…”
• “Draft an email to the finance department, requesting an update on the quarterly budget.”
• “How should I format a memo to all employees, announcing this new company policy regarding remote work…?”

Sample Routine (Short) Business Report Prompt:

• “Assemble a concise report on the sales performance for the last quarter, including these key metrics and trends.”
• “Please improve the clarity and conciseness of this detailed report on the impact of the recent marketing campaign, keep my data analysis and recommendations.”
• “Create a template for a progress report on the ongoing software development project, highlighting key milestones and challenges.”

Visual Presentations

Use AI tools like Gamma, Beautiful.ai, Canva, or PowerPoint Designer to generate design suggestions, content ideas, and even entire slides based on your input. AI can also help structure your presentation logically by suggesting a clear introduction, body, and conclusion. It can help you write concise and impactful bullet points, limiting the amount of text on each slide. AI can further assist in finding relevant images and videos to enhance your message and suggest creative slide transitions and animations.

Finally, practice your delivery and seek feedback to refine your presentation. AI tools like Resemble AI can help you practice by providing voice training and analysis. You can also use AI voice cloning tools like Speechify to create realistic virtual presenters (avatars) and record practice sessions. By analyzing these recordings, you can identify areas for improvement, such as pacing, tone, and body language.

Here are some sample, generic, and brief prompts to help you imagine how you might begin your interaction with an AIWA to develop a presentation. You should make your prompts more elaborate, following the RACE or another framework for prompt engineering:

Sample Presentation Prompts:

• “What are some ideas for catching the attention of my audience in an introduction for a 10-minute presentation on the impact of artificial intelligence on applicant recruitment?”
• “Suggest a visually appealing color scheme and font pairing for a presentation on sustainable energy. Consider a minimalist design approach.”
• “Based on my proposal for remote work I provided you, create 3 bullet points for a slide on the benefits of remote work, focusing on increased productivity and employee satisfaction.”

Resumes

To develop a resume with AIWA assistance, begin by gathering relevant information about your work experience, education, skills, and accomplishments, focusing on specific job titles, company names, dates of employment, key responsibilities, and notable achievements. Avoid sharing sensitive information, such as your name, salary details, or confidential project information.

Use an AIWA to generate initial drafts, customizing the content to highlight your unique qualifications. While AI is familiar with general resume conventions — such as the use of action verbs, quantifiable achievements, and a clear and concise format — providing specific guidance is essential to ensure the generated content aligns with your actual experience and skills as well as your desired style and format.

Here are some sample, generic, and brief prompts to help you imagine how you might begin your interaction with an AIWA to develop resume content. You should make your prompts more elaborate, following the RACE or another framework for prompt engineering:^[59]

Sample Resume Prompt:

• “I’m applying for a marketing internship. Based on my resume and coursework, and the internship requirements I have provided you, suggest key skills and accomplishments to highlight.”
• “Generate a professional summary highlighting my 5 years of experience in human resources and marketing, emphasizing project management skills and leadership experience.”
• Create a strong skills section, including technical skills such as Python, Java, and SQL, as well as soft skills like problem-solving and communication.

Cover Letters

To create a customized cover letter, begin by gathering specific details about the job you’re applying for — such as the company culture, the role’s key responsibilities, and the required qualifications — and your skills, experience, and career goals. You’ll use this information to draft specialized prompts for your letter.

Use an AIWA to generate initial drafts, customizing the content to highlight how your qualifications align with the specific job requirements and company culture. For example, if the job description emphasizes strong communication skills and teamwork, the AIWA can generate content that highlights your experience in collaborative projects and your ability to effectively convey complex ideas, as long as you prompt it to do so. Carefully proofread and edit your cover letter, ensuring it represents your genuine interest in the job and your actual skills and qualifications.

Here are some sample, generic, and brief prompts to help you imagine how you might begin your interaction with an AIWA to develop a cover letter. You should make your prompts more elaborate, following the RACE or another framework for prompt engineering:

Sample Resume and Cover Letter Prompts:

• “I’m interested in the marketing internship position at Green Sleeves. Generate a draft cover letter that highlights my relevant skills and enthusiasm for the opportunity based on my resume and their job announcement.”
• “What keywords should I include in a cover letter for a data analyst position?”
• “How should I explain how my previous experience in [previous field] has equipped me with the necessary skills and knowledge to transition to a data-driven role?”
• “Write a compelling paragraph highlighting how my passion for sustainable fashion and my internship experience with the Sustainable Farming Initiative where I gained hands-on experience in sustainable agriculture practices aligns with Green Sleeves’ commitment to eco-friendly practices. Emphasize my ability to innovate sustainable solutions and my dedication to promoting a more sustainable future.”

20.3 Ethical and Strategic AI Decision-Making

Disclaimers and Copyright Concerns in AI-Generated Content

In terms of copyright, AI-generated ideas and suggestions are typically not copyrightable since they’re usually based on commonly known concepts or widely used phrases (see resources from the United States Copyright Office https://www.copyright.gov/ai/). However, when AI generates specific content — such as particular text, creative expressions, or phrases based on proprietary or copyrighted sources — it could potentially infringe on copyright. Businesses should always check the origin of content generated by AIWAs and ensure that it doesn’t infringe on intellectual property rights.

Businesses (and you) should consider attributing the AIWA as the source of content when appropriate.^[60] Many companies now use disclaimers to provide transparency to clients and stakeholders and clarify when content is AI-generated. For instance, an email or report might include a note like:

Disclaimer: “This content was generated with the assistance of AI and is intended for informational purposes only. Please verify any factual details before relying on the information provided.”

Conclusion

Artificial Intelligence Writing Applications (AIWAs) are disrupting business communication practices, and you’ll likely use them in your professional and academic work if you haven’t already.^{[61] [62] [63]} While they offer potential benefits like increased efficiency and consistency, using them responsibly and ethically is essential to enhance your abilities and avoid costly mistakes in decision-making and damaged reputations (yours and your business’s). To do this, you must be an ethical user of AI and develop AI literacy — a skillset that enables individuals to understand AI’s capabilities and limitations (application), retain their unique voice and enhance their relationships (authenticity), and exercise human judgment, control, critical thinking, and creativity in the communication process (accountability and agency).^[64]

To use AIWAs effectively, providing clear and specific prompts is important.^[65] AIWAs can be prompted to assist in various tasks, including brainstorming, research, proofreading, content (message) creation, and presentation development.^[66] However, remember that AIWAs can’t actually think. Your input, control, and understanding of what’s needed from a message is necessary to ensure the quality and relevance of the output.

AI technologies are rapidly advancing, and as they continue to evolve, issues like AI hallucinations, biased training data, data privacy, and human understanding of and overreliance on technology must be addressed by AI developers, business communication educators and students, and employee training and development professionals.^{[67] [68]}

Ongoing research and development are needed to improve AI models and ensure they’re trained on diverse and unbiased datasets. Building a culture of critical thinking and information literacy will empower students and professionals to evaluate AI-generated content and make informed ethical decisions. As AI takes on routine tasks, skills such as emotional intelligence, interpersonal skills, and effective verbal communication will become increasingly valuable in the workplace.^[69] These skills will enable individuals to build strong relationships, collaborate effectively, and navigate complex social and emotional situations, setting them apart in an AI-powered workplace.

Case Studies of AI in Business Communication

Case Study 1: Maya’s Marketing Dilemma

Maya, a junior marketing associate, was tasked with writing a blog post on the latest industry trends. Feeling overwhelmed by the deadline, she turned to an AI writing tool to generate a draft. The tool produced a well-written and informative piece, saving her significant time and effort. However, Maya hesitated to disclose the AI’s involvement, fearing it might reflect poorly on her abilities. What should she do next and why? What should she consider in her decision?

Case Study 2: Clayton’s Confidential Chatbot Conversation

Clayton, a project manager, was struggling to brainstorm innovative solutions for an upcoming project. He decided to use a popular AI chatbot to bounce ideas off of. As he discussed project details, he realized he may have inadvertently shared confidential information about the company’s strategic plans and upcoming product launches. What should he have considered before beginning his AI-powered conversation? What should he do next and why?

Case Study 3: Amina’s AI Initiative

Amina, the owner of a thriving online boutique, is considering AI’s role in her business’s future. A recent surge in social media competition has made it increasingly difficult to maintain a consistent content calendar. As she and her team struggle to keep up with the demand for fresh, engaging posts, Amina is exploring AI tools as a potential solution. However, she’s cautious about their ethical implications and potential impact on brand authenticity. She’s considering developing an AI-use policy for her social media marketing team. What should she do next and why? What should she consider as she makes her decision?

Case Study 4: Chan’s Artificial Apology

Chan, an account manager at a construction management firm, receives an email from a long-time client filled with frustration and disappointment over their recent commercial construction project. The project had been delayed, over budget, and compromises were made on the original design due to a scheduling mistake. Chan feels embarrassed over the clear fault of his firm and knows he needs to respond promptly. Uncomfortable with writing such an apology, Chan turns to Google Gemini. He inputs the prompt: “Write an apology letter to an angry customer.” The AI generated a generic apology, expressing regret for the inconvenience and promising to rectify the situation. Chan copies and pastes the AI-generated response into a reply email to the customer and is about to hit “send.” Before sending this email, what should Chan consider? What should he do next and why?

Closing Case: The Statistic That Wasn’t — Revisited

It’s Sunday evening, and Noa Villaseñor-Katz is sitting at her kitchen table with a legal pad, a cold cup of coffee, and a problem she can’t solve alone. Tomorrow’s all-hands meeting is twelve hours away. She has spent the weekend drafting the bones of an AI-use policy, but the more she writes, the more she realizes the policy needs to address not just what happened with Brin’s press release but the deeper tensions the incident exposed.

Start with the AI literacy framework from §20.1. Each team member’s relationship to AI maps to a different capability — and a different gap. Brin’s hallucinated statistic is a failure of accountability: he didn’t verify the AI’s output, and the fabricated number reached a client and a trade journalist. But his prompt — vague, context-free, with no instruction to flag unverifiable claims — is also a failure of application. He used the tool without understanding its limitations. Reeve, by contrast, scores high on authenticity (he refuses to let AI touch his voice) but low on application (he hasn’t explored what AI can do well, like summarizing research or reformatting content). Suki models agency: she uses AI for internal brainstorming but rewrites everything before it reaches a client, keeping herself as the decision-maker. Tomás has strong application skills with image generators but faces an accountability gap around disclosure and copyright.

Now consider the prompt engineering concepts from §20.2. Brin’s prompt — Write a press release announcing Bridger Valley Provisions’ growth in the organic jerky market in the Mountain West region. Make it sound professional and newsworthy — violated nearly every principle of the RACE framework. It assigned no specific role, gave no context about the company’s actual data, and set an expectation (“newsworthy”) that encouraged the AI to generate impressive-sounding claims. A RACE-structured prompt would have constrained the AI to work with verified information and flag any claims it couldn’t source. The hallucination wasn’t random bad luck; it was a predictable consequence of a poorly engineered prompt. This insight matters for the policy: training the team on prompt engineering isn’t optional — it’s the primary safeguard against the kind of error that nearly cost them the Bridger Valley account.

Finally, apply the ethical and strategic decision-making principles from §20.3. Noa needs to decide how to handle disclosure. Should Meridian & Lark tell clients when AI assists in producing their content? The disclaimers discussion suggests yes, but the degree of disclosure matters. A blanket disclaimer on every email might alarm clients unnecessarily; no disclaimer at all leaves the agency vulnerable to the kind of surprise that hit Margaux Prentiss. The four case studies in the section mirror the agency’s own dilemmas: Maya’s reluctance to disclose AI involvement echoes Brin’s failure to flag it; Clayton’s confidential information leak is the privacy risk Noa hasn’t addressed yet; Amina’s policy-drafting process is exactly what Noa is attempting; and Chan’s generic AI apology is a cautionary tale about what happens when AI handles high-stakes, relationship-dependent communication without human revision.

Noa sees three paths forward for tomorrow’s meeting:

Option A: The Firewall. Adopt Reeve’s position and ban AI from all client-facing content. Internal use only — brainstorming, research summaries, agenda templates. This protects quality and eliminates disclosure concerns, but it doesn’t address the competitive gap that cost them Gallatin Outfitters. Reeve would support it. Brin would see it as punishing the whole team for his mistake. Suki would call it an overreaction that ignores AI’s real value.

Option B: The Green Light. Adopt Brin’s position: AI is fine everywhere, the team just needs better prompt training. This maintains speed and competitiveness, but it relies on individual judgment — the same judgment that already failed. There are no structural safeguards, no verification requirements, and no disclosure standards. If another hallucination slips through, the agency has no defense beyond “we trained people better.”

Option C: The Tiered System. Build on Suki’s proposal: create categories of AI use (internal drafting, client-facing with human revision, prohibited uses), require verification protocols for any AI-assisted content that reaches a client, establish disclosure guidelines, and mandate RACE-framework training for the entire team. This balances speed and safety but requires buy-in from both Reeve (who must accept that AI has a place) and Brin (who must accept that AI output requires verification). It also requires clear definitions — what counts as “client-facing”? Does Tomás’s mood board? Does an internal strategy memo that gets forwarded to a client?

There is no single right answer. Each option reflects different priorities — quality, speed, trust, competitiveness — and each carries risks the others avoid. What Noa decides will shape not just the agency’s workflow but its identity: the story it tells clients, and itself, about what kind of work it does and why.

Closing Case Discussion Questions

Review Questions

Key Terms Matching

Application Exercises

Discussion Questions

Extended Project

Self-Assessment Revisit

20.4 Additional Resources

Watch Andrej Karpathy’s YouTube video where he explains how large language models work. https://youtu.be/zjkBMFhNj_g?si=zu-b4o7fUo6jSpW0

Read the article “Leveraging AI in Business: 3 Real-World Examples” by Kate Gibson on Harvard Business School Online. https://online.hbs.edu/blog/post/ai-in-business

Read the article “Research: How AI Helped Executives Improve Communication” by Katharina Lange and Jose Parra-Moyano from the Harvard Business Review. https://hbr.org/2025/02/research-how-ai-helped-executives-improve-communication

Read: Getchell, K., Carradini, S., Cardon, P. W., Fleischmann, C., Ma, H., Aritz, J., … & Stapp, J. (2022). Artificial intelligence in business communication: the changing landscape of research and teaching. Business and Professional Communication Quarterly, 85(1), 7-33. https://doi.org/10.1177/23294906221074311

Check out the International Association of Business Communicators’ Guidelines on the Ethical Use of AI. https://www.iabc.com/about/what-we-do/standards/ethical-use-of-ai

Watch this AI-generated video on “Written Business Communication in the Age of Artificial Intelligence.” https://youtu.be/N92CAkStDNU?si=3xPrCmtJEpln93a5