General AI assistant
ChatGPT
OpenAI's general-purpose AI assistant for drafting, document work and research, not a legal database.
Last verified:
Categories
AI tools for drafting and document automation
These tools draft memos, minutes and other legal documents. The category also covers document automation, assembling documents from templates, though none of the tools we profile so far documents it. The list ranks them by documented data handling, workflow fit and estimated price. Legal task quality, drafting included, is scored on each tool's profile from published benchmarks, which test US law rather than Finnish documents.
- Last verified:
- Last updated:
Ranked by Security and data handling, then Workflow fit, then Estimated price, within each type of tool. How we rank
We rank only by criteria assessed for every tool on the list, and only tools of the same type against each other. No placement can be bought.
Showing 4 of 4 tools.
General AI assistants
-
No. 1 of 2
-
No. 2 of 2
General AI assistant
Claude
Anthropic's general-purpose AI assistant, not a legal database, profiled for law firms and legal teams in Finland.
Last verified:
Legal AI platforms
-
No. 1 of 2
Legal AI platform
Legora
A legal AI platform with an agent, legal research, drafting in Word and bulk document review, for law firms and in-house teams.
Last verified:
-
No. 2 of 2
Legal AI platform
Harvey
A legal AI platform with agents, legal research, drafting and bulk document review, for law firms and in-house legal teams.
Last verified:
What counts as a drafting tool here?
A tool is in this category when its vendor documents that it drafts or edits legal documents, or assembles them from templates. General AI assistants are prompted to draft; legal AI platforms describe drafting inside Word and against the team’s own precedents.
How are these tools ranked?
Within each type of tool, by data handling first, then workflow fit, then estimated price, cheapest first: the criteria assessed for every tool on the list. Draft quality is part of legal task quality, which each profile scores from a named, linked published benchmark.