Module 1 · Lesson 1 of 13AI in Plain LanguageFree preview
AI in Plain Language: What It Is and How It Works
A jargon-free explanation of today's AI tools, how they produce answers, and a simple mental model for using them well at work.
On this page
You have probably used AI today, maybe without noticing: the suggested reply in your email, the "summarize" button in a meeting app, the chatbot on a vendor's website. You may also have heard two very different stories from colleagues. One drafted a proposal in five minutes with an AI assistant. Another nearly sent a client a "fact" the AI made up. Both stories are true, and both come from the same kind of tool. This lesson gives you a clear, hype-free picture of what these tools are, so you can use them with confidence and know where the edges are.
Learning objectives
- Explain in a sentence or two what generative AI is and how it produces answers
- Recognize the AI tools and features you already encounter at work
- Use the "capable new colleague" mental model to decide how much to trust AI output
- Run a first, safe experiment with an AI assistant
What people mean by "AI" today
Artificial intelligence (AI) is a broad term for computer systems that do tasks we usually associate with human thinking: recognizing speech, spotting patterns, making predictions, writing text. AI has been in business software for years. Spam filters, fraud alerts on your credit card, and "customers also bought" recommendations are all AI.
What changed recently is generative AI: AI that creates new content (text, images, audio, or computer code) in response to a request written in everyday language. That request is called a prompt. Tools like ChatGPT, Claude, Gemini, and Microsoft Copilot are AI assistants built on generative AI, and many of the apps you already use now include generative AI features as well.
The systems behind text-based assistants are called large language models (LLMs). You don't need to know how to build one. You do need a rough idea of how one works, because that explains both why it's so useful and why it makes mistakes.
How it works, in one paragraph
An LLM learned from an enormous amount of text: books, websites, articles, and other written material. From that, it learned patterns, meaning which words, ideas, and structures tend to go together. When you type a prompt, it builds a response by repeatedly predicting a likely next piece of text, based on those patterns and on what you've told it. That's why it can write a polished email in seconds: it has absorbed the patterns of countless emails. It's also why it can be confidently wrong. It produces what sounds right, which is not the same as looking up what is right.
Example: Ask an AI assistant for "three tips for writing a clear meeting agenda," and you'll get sensible, well-written tips, because that kind of advice appears constantly in what it learned. Ask it for "the return policy at our Dallas warehouse," and it may produce a plausible-sounding policy that doesn't exist. Unless your organization has connected the tool to your internal documents, it has never seen that policy. It is filling the gap with a guess.
Some AI tools can also search the web, read files you upload, or connect to your company's systems. Those features can improve accuracy, but they don't remove the need to check the answer.
The mental model: a capable new colleague on day one
Imagine a new colleague who is fast, articulate, well read on almost every topic, and eager to help. This colleague also:
- started this morning and knows nothing about your company, customers, or unwritten rules
- rarely says "I'm not sure" unless you ask
- sometimes fills gaps with confident guesses
- does exactly what you ask, even when what you asked was vague
How would you work with that person? You'd give clear instructions and useful background. You'd hand them first drafts, not final decisions. You'd check their work before it reached a customer. And you wouldn't give them the keys to the safe on their first day.
That is the right way to work with AI. Keep this picture in mind for the rest of the course.
| AI works well as... | AI is risky as... |
|---|---|
| A first-draft writer | The final author of anything important |
| A summarizer of material you provide | A source of facts you haven't checked |
| A brainstorming partner | The decision-maker on anything that affects people |
| A patient explainer of unfamiliar topics | A substitute for legal, medical, or financial expertise |
| An editor for tone, clarity, and length | A place to put confidential information |
Where you'll meet AI at work
- Chat assistants you open in a browser or app, such as ChatGPT, Claude, Gemini, or Microsoft Copilot.
- Built-in features in tools you already use: "draft with AI" in email, meeting summaries in video calls, AI search in a knowledge base, suggested replies in help desk software.
- Customer-facing AI, such as a chatbot on your organization's website or an automated phone assistant.
- Behind-the-scenes AI that routes support tickets, flags unusual transactions, or ranks sales leads.
Not every AI tool is approved for every kind of work at your organization. Module 3 shows you how to tell the difference. Until then, a simple rule: experiment only in a tool your organization has approved, and only with public, non-sensitive information.
Why this matters for your job
AI doesn't make your judgment unnecessary. It changes where you spend your time: less on blank-page drafting and reformatting, more on deciding, checking, and working with people. The people who get the most from AI are the ones who know what to hand it, how to ask clearly, and when to stop and verify. That's exactly what this course teaches, in about two and a half hours, with no technical background required.
Try it
Time: about 6 minutes. Use an AI assistant your organization has approved, and use only public, non-sensitive information.
- Pick a routine writing task from your job that involves nothing confidential. For example: an out-of-office message, an agenda for a generic team check-in, or a plain-language explanation of a common term in your field.
- Paste this prompt into your AI tool and fill in the brackets:
I work in [your department] at a [type of organization]. Write a [type of document] for [audience]. Keep it under [number] words, use a [friendly / formal / direct] tone, and format it as [bullet points / short paragraphs / a table].
- Read the result as if a new colleague wrote it. What's good? What's generic? Is anything wrong or invented?
- Now test the limits. Ask a question about something specific to your organization that isn't public, such as "What is [your organization]'s policy on [an internal topic]?" Unless your tool is connected to your company's documents, watch whether it admits it doesn't know or produces a confident guess.
- Write down one task where you'd happily use this help again, and one thing you would always check.
Key takeaways
- Generative AI creates new content from a plain-language request called a prompt.
- It works by predicting likely text from patterns, so it can sound right while being wrong.
- Treat AI like a capable new colleague on day one: give clear instructions, use it for drafts, and check its work.
- AI already appears in many tools you use, and not all of them are approved for every kind of information.
- Your judgment, your knowledge of your organization, and your accountability are what turn AI output into good work.
Enjoyed this preview?
Unlock all 13 lessons, 2 templates, and your certificate for $99.