Phase 1 · ChatGPT · Level 1 · Foundations
What ChatGPT is great for at work
By the end, you'll be able to…
- Match a real work task to one of five things ChatGPT does well: drafting, summarising, rewriting for tone, explaining, and chat with voice and image understanding
- Write a first prompt for each that gives you something usable, not filler
- Judge when image understanding or voice earns its place in your day
Why it matters
A blank chat box is intimidating precisely because it can do so much. This lesson replaces 'what do I even type?' with five concrete, recognisable jobs, each with a worked example on real Fernway material, so you'll spot a task in your own week and know it's a good fit.
The honest short version
ChatGPT is, at heart, very good with words: writing them, reshaping them, explaining them, and thinking out loud with you. It's also increasingly good with voice and images, which widens what "words" can come from. Once you see it that way, the useful jobs stop being a mystery. Here are the five that cover most of a working week, each shown on the fictional Fernway Group so you can see exactly what a good first prompt looks like.
1. Drafting from scratch
The most common first win. You have a thing to write and a blank page; ChatGPT gives you a draft to react to, which is far easier than starting from nothing.
Draft a short, friendly team email from Priya announcing that the August new-starter onboarding stays in the week of 18 August, but the sales-shadowing sessions move to Monday 25 August so the sales team can focus on a big renewal. About 90 words, warm but clear.
Why this works: It supplies the audience, the facts, the tone and the length, so the model produces a real draft instead of a generic template.
You'll get a complete email (greeting, the change, the reason, a sign-off) reading far more smoothly than a first attempt of your own. Then keep going in the same chat: "add one line thanking Maya for reworking the schedule." It edits the existing draft rather than starting over. A draft in ten seconds that you shape in two follow-ups is the whole point.
2. Summarising
Turning something long into something you can act on. Notes, a policy, a thread: ChatGPT pulls out what matters and drops the rest.
From these rough meeting notes, give me: (1) a three-sentence summary a busy manager could read at a glance, then (2) a table of every action with columns Action, Owner, Due date. Mark anything with no clear owner as UNASSIGNED. Notes: [paste].
Why this works: Asking for a specific shape (a short summary plus an owner-and-date table) gives you output you can paste into a task list instead of prose you'd have to reorganise.
The reply surfaces the decisions and hands you a clean action table, including flagging the jobs nobody picked up. That last instruction is doing real work: without it, unassigned actions quietly vanish.
3. Rewriting and fixing tone
Often you already have the words; they're just wrong for the reader. This is where nervous senders get the most value.
Make this warmer and less blunt without losing the point. It's going to a colleague I work with closely: "I still haven't had the numbers for the Harlow renewal. I need them before the next sync."
Why this works: Naming the relationship and the feeling you want lets the model adjust register precisely, rather than guessing how formal to be.
Ask for it "warmer", "half the length", "less corporate", "more direct", one nudge at a time, and it adjusts. It's a drafting partner, not a vending machine, so the first version is where you start.
4. Explaining and learning
ChatGPT is patient and never makes you feel daft for asking, which makes it a good explainer for the things you nod along to but don't quite follow.
Explain what an "auto-renewing lease" means for a small business, as if I've never dealt with commercial property. Then tell me the one thing I'd most want to check before a renewal deadline.
Why this works: Setting your starting point ('as if I've never dealt with one') pitches the explanation at the right level so you actually follow it.
The trick is the follow-up: ask it to explain simply, then ask about the exact bit that confused you, and keep going until it clicks. Because this is a real strength, it's also where over-reliance creeps in; see Common mistakes.
5. Chat, with voice and image understanding
Underneath all of this is plain conversation, and two extra ways of feeding it in.
Voice lets you talk to ChatGPT and hear it reply, hands-free, useful for thinking out loud on a commute, rehearsing a tricky conversation, or capturing an idea while you're away from the keyboard. What you say is turned into text and answered like any other prompt. See the voice mode FAQ.
Image understanding means you can show ChatGPT a picture and ask about it. This is part of it being multimodal, able to work with more than just typed text. At work that's handy: photograph a whiteboard after a workshop and ask for the notes as a typed list; snap a printed form or a slide and ask what it's saying; share a screenshot of a confusing error message or a chart and ask it to explain what you're looking at. It reads and interprets what's in the image, then answers in words.
A crucial caveat carries across both: reading your voice or your photo doesn't make ChatGPT more accurate. It can misread a smudged figure on a whiteboard or a number in a chart just as confidently as it can invent a fact. Treat anything it pulls from an image as a draft to check, especially numbers.
Try it now
Common mistakes
- Treating it as a search engine. It predicts likely text, it doesn't look up guaranteed facts. Fluent is not the same as correct.
- Giving it too little to work with. "Write an email" is thin. Name the reader, the situation and the tone, as every prompt above does.
- Expecting one perfect answer. The strength is the back-and-forth; the first reply is a draft to improve.
- Over-trusting the explainer because it's patient and confident. ChatGPT is a brilliant-sounding tutor that is sometimes simply wrong, and it won't flag which bits. For anything you'll act on (money, legal, health, a decision at work) treat its explanation as a helpful start and confirm the load-bearing facts against a trustworthy source before you rely on them.
Keeping current
What ChatGPT can take in (voice, images, files, live web results) keeps expanding, and the exact controls move around the screen. When a feature here looks different, check OpenAI's official ChatGPT release notes. Accurate as of 13 July 2026.