Claude? What the Hell Is Claude?

Let's clear this up.

Claude is an AI assistant — same general category as ChatGPT (made by OpenAI) and Gemini (made by Google). It's made by a company called Anthropic, founded by ex-OpenAI researchers in 2021. It's named after Claude Shannon, an American mathematician who basically invented the field of information theory in the 1940s. (No, you don't need to know that.)

If you've used ChatGPT, you already know how to use Claude. Same shape: you type or speak, it responds, you have a conversation. The differences are in how it responds and what it's actually good at.

Why people choose Claude over ChatGPT

In no particular order, and based on day-to-day use rather than hype:

• It's more careful. Claude is famously less likely to confidently make things up than its competitors. When it doesn't know something, it's more likely to say so. For business work where wrong answers are expensive, this matters more than people realise.
• It's better at long, structured tasks. Drafting a Statement of Work from a transcript. Reading a 50-page contract and pulling out the actual obligations. Comparing two documents for the differences that matter. Claude tends to stay on-task and produce more usable output.
• It's less yappy. Ask Claude a yes/no question, you get a yes/no with a one-line reason. Ask the same question of some competitors and you'll get five paragraphs of throat-clearing first. Time is money.
• Long memory. Claude can hold a lot of information in one conversation — useful for "here's our customer history, draft me a renewal proposal that makes sense" type tasks.
• Coding. If you build software, Claude Code is widely considered one of the strongest tools for letting an AI actually edit your codebase. We use it every day.

Where it's weaker

Be honest:

• Image generation. Claude doesn't generate images. If you need a picture, use ChatGPT or Gemini or one of the dedicated tools.
• Web search by default. ChatGPT does live web search well. Claude does it but with less polish.
• Ecosystem. ChatGPT's mobile app and integrations are slicker. The brand has more momentum.
• Voice. Claude's voice mode exists but is less mature than ChatGPT's Advanced Voice or Google's Gemini Live.

Three Claudes, not one

There are different sizes:

• Claude Haiku — fast, cheap, good for high-volume simple tasks. Think: sorting emails, classifying support tickets, summarising meeting notes.
• Claude Sonnet — the workhorse. Most business use cases land here. Good balance of capability and cost.
• Claude Opus — the brain. For high-stakes reasoning, complex extraction, long-document analysis. Pricier per use.

If someone says "we use Claude", they usually mean Sonnet, sometimes Opus for the hard stuff. You don't need to memorise this; the right one to use is usually obvious in context.

Cost (in plain numbers)

For a small business actually using it for work — drafting documents, summarising calls, helping with email — Claude costs single-digit-NZD per user per month. The pro plan is NZD 30-ish/month per person for unlimited day-to-day use.

For heavier programmatic use (e.g., embedding it in your own app), Claude is priced per million tokens, which is a unit roughly equivalent to "about 750 words". Most workflows we build for NZ SMEs cost well under NZD 200/month total in AI fees.

So which should you use?

If you're working on something where being right matters more than the bells and whistles — contracts, customer communications, anything that goes out under your name — Claude is the safer default.

If you're building or coding software, Claude.

If you're generating images or want the slickest mobile experience, ChatGPT.

If you're deep in Google Workspace and want one tool that lives inside Gmail and Docs, Gemini.

And if you're running a business and trying to figure out which of these to deploy at scale — talk to us. That's literally what we do.

Daniel Thomson runs AiTearoa, an AI consultancy in Aotearoa New Zealand. We help businesses pick the right tools and ship them into real workflows.