Buyer guide
10 questions before you buy legal AI
Use this checklist when comparing Harvey, CoCounsel, Lexis+ AI, Spellbook, Ironclad, and matter-native platforms like Locke Prime.
Ask for citation excerpts, filename/path matches, and matter-scoped retrieval — not pasted summaries. Locke Prime indexes firm documents with RAG, exact-phrase weighting, and trust footers showing model and timestamp on outputs.
Point solutions excel in one lane. Locke Prime unifies Case Advisor, adversarial Boardroom, War Room, CLM Light, Locke CLM, and CLM Pro on one matter record — so strategy and contracts share context.
Require firm_id scoping on every sensitive route, IDOR tests, and a written data-handling policy. Locke Prime documents tenant isolation in our security overview. Query content is not used to train models.
Licensed-attorney use only; all outputs require review before client or court use. Locke Prime does not practice law and does not create an attorney-client relationship. See Terms of Service.
Many vendors are demo-only or self-serve. Locke Prime requires a consultation before access — we confirm fit and provision the right tier after a guided walkthrough. Request consultation.
Ask about playbook deviation, obligation inbox, portfolio risk, and version diff — not just redlines in Word. Locke Prime ships CLM Light (Starter/Professional) and full Locke CLM + CLM Pro on Enterprise. Contracts overview →
Boardroom runs fixed roles (Co-Counsel, Opposing Counsel, Judge). War Room maps proof gaps and trial readiness. Litigation overview →
Opaque throttling hides economics. Locke Prime exposes usage and estimated cost in Admin and System Monitor — with optional BYOK on Enterprise. Usage transparency →
Locke Prime uses multi-provider routing per task. Enterprise supports BYOK. Legal intelligence overview →
Public tiers via GET /api/entitlements/tiers and pricing. Production uses manual billing until Stripe Checkout ships; enterprise gets dedicated onboarding and API access.