Seven steps that turn Claude from a basic chatbot into an assistant that knows you and reaches into your real work. You can do all of this in an afternoon.
Most people use Claude like a search box. They type a question, get an answer, and stop there.
Once it’s set up, Claude knows who you are, works inside your real tools and finishes whole jobs. Here’s the order I’d do it in.
01
Pick your model
Click the model name next to the send button. Each model is good at a different kind of work.
The free plan includes Sonnet and Haiku. Pro adds Opus, and Fable is paid with usage credits. Max includes Fable up to half your weekly limit. (Claude pricing)
In that same menu, click Effort to tell Claude how hard to think. There are five levels, from Low up to Max, and each model’s recommended one is marked Default. Turn it up for hard problems and down for quicker answers that use less of your limit. (How effort works)
02
Turn on memory
Claude can remember you across chats. Your name, your style, your projects. It’s on by default for Free, Pro and Max, so just check it.
On Team and Enterprise, an owner has to allow memory for the organization first. Then each person turns it on for themselves.
03
Build a Project
A Project is a workspace that holds your files and instructions in one place, so every chat inside it starts with the right context.
Make one Project for each big thing you’re working on. The free plan gets up to five.
04
Connect your apps
Connectors plug Claude into tools like Gmail, Google Drive, Notion and Slack, so it can pull from and act on your real work.
You sign in through each app’s own login, so Claude only sees what your account can see. Start with read-only work like "summarize my unread emails from clients" before you let it send or edit anything.
05
Use artifacts
Ask Claude to build something and it makes the real thing, open right beside your chat. A doc, a slide deck, a dashboard, a small working tool. You can edit it, come back to it from the Artifacts tab in the sidebar, and share it.
Just ask. For example: "build me a simple weekly meal planner I can edit." If nothing opens, go to Settings, then Capabilities, and turn on Code execution and file creation (in the newer layout it’s called Create and edit files).
06
Turn on skills
A skill teaches Claude a workflow once, so you stop retyping the same long prompt. You don’t have to call it up. Claude uses it on its own whenever your request fits.
Skills work on every plan, including Free.
07
Hand off whole tasks
This used to be its own mode called Cowork. Since September 2026, Anthropic has been folding Cowork into regular chat, starting with Pro and Max, with Team and Free after. You don’t pick a mode anymore. Ask, and Claude decides whether you need a quick answer or a whole job done.
Give it a multi-step task, like "pull last month’s invoices from my Drive and build a spreadsheet of who still owes me," then walk away. On Pro and Max, these tasks run in the cloud, so they keep going after you close your laptop. Anything on your own computer, like local folders or Claude driving your browser, needs the Claude desktop app open.
Task work is on paid plans, and big jobs use more of your limit than a quick question, so save it for the big stuff. (Cowork help article)
The Real Win
Do these seven once and Claude stops being a chatbot you ask questions. It becomes an assistant that knows you, lives in your real tools, and can finish whole jobs on its own.
Want the 10-minute version first? Start with my 5-step beginner setup.