Start on the web: ChatGPT and Claude for everyday work

Start where it costs you nothing to be wrong: a web chat doing the everyday writing and thinking you already do, briefed the way you would brief a sharp junior.

What rung one actually is

Rung one is a model in your browser, used for real work. Open ChatGPT or Claude and put your actual tasks into it. Paste in a messy client brief and ask for a clean summary. Hand it your rough proposal and ask it to sharpen the argument. Drop in a call transcript and ask for the three decisions and who owns each. There is nothing to install and nothing to configure; this is the rung almost everyone reaches, and most people stop here because it already feels like a win.

It is worth knowing the two tools are close cousins with slightly different feels. ChatGPT is the one most clients have heard of and is excellent for quick drafting and brainstorming. Claude tends to hold a longer, more careful thread and is the model you will lean on as you climb, because the later rungs in this playbook run on it. Pick one to live in, keep the other open for a second opinion, and do not agonise over the choice; the skill that matters is the same on both.

Why this rung earns its place

It would be easy to dismiss rung one as toy use, but it is the foundation the entire ladder is built on, and skipping it is why people stall later.

The everyday wins are not trivial. Drafting, summarising, reformatting, and research are most of the dull middle layer that eats your week. Clearing them in a web chat buys back real hours immediately, with zero setup risk. More importantly, this is where you learn to brief, and briefing is the single skill that every higher rung depends on. A freelancer who can brief well on the web gets dramatically more from Cowork, from coding agents, and from a standing agent. One who cannot will find every rung disappointing and blame the tool.

The exact steps to a usable answer

The difference between a casual user and a leveraged one is entirely in how you ask. Follow this shape every time.

1. Give it a role. Tell the model who it is. "You are me, a freelance brand designer." This sets the voice and the assumptions.
2. Give it the context. State the situation, the constraints, and anything it cannot guess. Who is the client, what has happened, what tone is right.
3. Give it the shape of the output. Word count, format, number of options. "Under a hundred words, two versions, warm but not needy."
4. Hand it the raw material. Paste the brief, the transcript, the draft, the notes. The model is only as good as what it can see.
5. Read, then refine. Treat the first answer as a draft. Tell it what is off ("too formal, cut the apology") and let it revise. The second or third pass is usually the keeper.

So instead of "write a follow-up email", you write: "you are me, a freelance designer, writing to a client who has gone quiet for two weeks on a project that is half paid; keep it warm, under a hundred words, and give me two versions." Treat the chat like briefing a sharp junior who has no memory of your business. Tell it who it is, what it is working with, and what good looks like. Get that habit right here and every later rung gets easier, because the whole path is built on knowing how to brief.

Add research to the mix

One more everyday win belongs on this rung: research. When you need to scope a new industry, check a claim, or gather current facts with sources, Perplexity answers like a chat but cites where each fact came from, so you can verify rather than trust. It replaces the half-hour of manual searching that nobody ever pays you for, and it hands back links you can stand behind in front of a client.

A worked example: the proposal turnaround

A freelance copywriter I know used the web chat to fix a bottleneck that was costing them work.

• Challenge. Each new proposal took about two hours to draft from a blank page, so they could only respond to a handful of leads a week and lost the warm ones to faster competitors.
• Approach. They built one detailed brief in Claude describing their voice, their packages, and their pricing logic, then for each new lead pasted the enquiry and asked for a first-draft proposal in that exact shape.
• Result. Drafting dropped from two hours to about twenty-five minutes of editing, and they now reply to every qualified lead the same day. Their close rate on warm leads rose because the proposal arrived while the prospect was still keen.

The example is the step itself: a strong reusable brief plus the raw material equals a usable draft. Anyone reading this can copy it today.

Pitfalls

• Vague prompts. A vague brief gets a vague answer. If the output is generic, the fix is almost always more role and more context, not a different tool.
• Accepting the first draft. The first answer is a starting point. Push back, refine, and the quality jumps.
• Pasting confidential client data carelessly. Strip names and sensitive details you would not want stored, or use a paid tier with the appropriate privacy settings, especially for regulated clients.
• Trusting facts blindly. A chat model can state something wrong with confidence. For anything factual, verify, and use a cited tool like Perplexity when sources matter.

Handoff

The web chat has one ceiling you will hit fast: it forgets, and it cannot touch your actual files. The next chapter, Level up with Claude Cowork, is the rung that breaks the box open and points the model at the real work your business is made of.