A coach we work with told us about a moment in a recent workshop. Someone on the call said, “Oh, you could just build an agent for that,” and several people nodded.
They nodded too. Then they wrote “agent?” on a sticky note and hoped the conversation would provide a clue.
We know that nod. We have used it ourselves, usually while opening a new browser tab under the table.
The AI vocabulary arrives with a particular tone. “Just prompt it.” “Simply set up an automation.” “Obviously you’d want an agent.” The words themselves aren’t the problem. It’s the “just” and the “simply” and the “obviously” doing the damage, because they suggest a handout got passed around and you missed it.
Nobody got the handout. Trust us. We build websites every week, we use these tools every day, and still one of us keeps an actual dictionary file with the words we had to look up. Seriously! (It’s called “Speak Nerd to Me.” This blog post is the “coach edition.”)
So let’s translate five AI terms coaches are hearing often. We will keep the explanations useful, accurate, and free of unnecessary computer-speak.
LLM, or Large Language Model
Think of it as: one of the main engines behind many AI chat tools.
An LLM is a computer model trained on large amounts of language so it can recognize patterns and generate responses. It works with pieces of language called tokens (roughly, word-parts), then predicts what is likely to come next based on the context it has been given.
That sounds more mechanical than the responses often feel, and that is worth remembering. Tools such as ChatGPT, Claude, and Gemini give us a friendly place to interact with a model, along with other features the company has built around it.
The short version: The chat app is where we talk to the AI. The LLM is one of the engines helping produce the response.
Prompt
Think of it as: the brief.
A prompt is the information and direction we give an AI tool. It might include a question, instructions, background, examples, or a file we want it to use.
We get better starting points when we give the tool enough context to understand the job. A useful prompt often answers four questions:
For example:
I’m a life coach who works with people navigating a career change. Please write three possible opening lines for my About page. Keep them warm, conversational, and under 25 words each. Avoid hype.
That will usually give us more to work with than “Write my About page.” It may still need our judgment and voice, but at least we have given it a decent map.
The short version: A prompt is a brief. The more useful context we provide, the more useful the first draft may be.
Hallucination
Think of it as: a made-up answer that sounds plausible.
Sometimes an AI tool gives us information that is inaccurate or completely invented, while sounding perfectly confident. That is commonly called a hallucination.
The tool is not deciding to deceive us. It is generating a response from patterns, and plausible language can still contain false information. Names, dates, statistics, quotations, sources, and links deserve an extra look before we publish them.
This is one place where our very human habit of pausing and asking, “Is that actually true?” becomes useful.
The short version: Polished is not the same as accurate. We check the details before our name goes on them.
Automation
Think of it as: a set of rules that runs on its own.
An automation starts with a trigger and follows the steps someone set up. When a person books a call, send a confirmation. When someone joins an email list, send the welcome message. When it is 6:30 in the morning, start the coffee maker and improve civilization slightly.
Automations do not have to use AI. Some are simple. Others include several rules and branches. The useful distinction is that an automation follows a process we have already defined.
The short version: An automation carries out preset steps after something triggers it.
Agent
Think of it as: an AI helper that can work through multiple steps.
An AI agent can plan and carry out a task toward a goal, often by using tools. Depending on how it is built, it might search for information, compare options, draft something, review the result, and decide what step comes next.
That makes an agent more flexible than a basic automation. It can also make the outcome less predictable, because the system may make choices as it works. Clear instructions, limited permissions, and human review matter, especially before an agent is trusted with anything public, sensitive, or expensive.
The short version: An automation follows a process we set. An agent can choose among steps as it works toward a goal. Both benefit from clear direction and thoughtful oversight.
These terms can help us understand what the technology actually does. They don’t describe the relationship at the heart of coaching.
AI may help someone organize ideas, draft questions, or explore possible next steps. A coach brings presence, discernment, lived experience, and the ability to notice what is happening with the person in front of them. A pause, a change in tone, or the sentence someone keeps circling may matter more than the words on the page.
That was the heart of last week’s conversation, and it still belongs here. We can learn to use AI without handing over the parts of coaching that depend on human connection, judgment, and care.
Many of us drive cars without knowing how to rebuild the engine. Ours still finds Starbucks with remarkable consistency.
AI can work the same way. We can learn enough to use a tool thoughtfully without becoming computer scientists. We can also decide a particular tool is not useful for our business. Understanding gives us options. It doesn’t create an obligation to use them.
When another unfamiliar word shows up, we can ask, “What does that mean in plain English?” By asking the question, we may be doing someone else in the room a favor too.
Translating technology is part of the work we chose at LifeCoach Elevate. Today it is AI vocabulary. Tomorrow it may be another website acronym wearing a tiny lab coat. And when that happens, we’ll look that one up too.
We don’t have to understand everything before we take action. We simply, in the words of Mary Morrissey, “Do what we can, with what we have, from where we are.”
Get Going.
