Claude for lawyers

What Claude is

Claude is the conversational AI built by Anthropic. The current production model is Claude Opus 4. For lawyers, the most relevant features are its long context window, which holds an entire commercial agreement and a hundred pages of correspondence in one conversation, and Claude Projects, which lets you give a persistent set of files and instructions to a single project.

Where Claude is the right tool

Long-form drafting. Document review. Anything where the lawyer needs to pass a long document and ask for analysis that respects nuance. Claude is the model I reach for when the work needs careful reading, not fast turnaround. It is conservative in the right places. It will tell you when it is unsure.

Where Claude is not the right tool

Image generation. Real-time web search (Claude has retrieval but it is deliberate, not the same posture as ChatGPT search). Voice. If the work is mostly throwing quick questions at a model and reading short answers, ChatGPT is faster.

Claude Projects, in a lawyer's terms

A Claude Project is a folder with custom instructions. For a litigator, a Project for a single matter holds the pleadings, the witness statements, and the relevant authorities. Every conversation inside the project starts with that context loaded. You ask "draft a reply to paragraph 14 of the counterclaim" and Claude has the counterclaim already.

Confidentiality posture

Anthropic does not train on inputs or outputs from Claude API or paid consumer accounts by default. The free tier is the exception; do not put client data into the free Claude. The full posture is at /ai/confidentiality.