Option A: Claude: one chat assembles and classifies
Best when: First-draft case files in one chat
What it does: Claude assembles and proposes; you confirm
Strong at: Evidence and classification stay together
Limit: Classifier also assembled the evidence; check both
Free plan: Yes
Option B: ChatGPT digests threads; you decide the type
Best when: The conclusion stays your own
What it does: ChatGPT quotes threads; you set the case type
Strong at: Case type set by you, not proposed by AI
Limit: Thread-by-thread uploads repeat; caps apply
Free plan: Yes, limited (uploads and analysis capped)
Option C: Gemini marks criteria; you choose the type
Best when: Evidence gaps listed against your criteria
What it does: Gemini marks criteria; you choose the type
Strong at: Output shows what evidence is missing
Limit: Writing the criteria is manual work
Free plan: Yes, limited