Andri Can Make Mistakes
Andri AI is a powerful legal assistant, but like all AI tools, it can make mistakes. Always review and verify AI-generated content before relying on it.
What's improved
Andri has safeguards against hallucinations—the AI will not invent fake case law citations or reference non-existent legal sources. This type of fabrication has been eliminated.
Where mistakes still happen
While Andri won't make up sources, it can still:
Misinterpret information — The AI might misunderstand the context of a document or draw incorrect conclusions from complex facts
Miss nuances — Subtle legal distinctions or jurisdiction-specific rules may be overlooked
Make logical errors — The reasoning in an analysis may contain flaws, even when citing real sources
Provide incomplete answers — Important considerations or exceptions might be omitted
Never use AI output directly in client work without verification. Review all legal analysis, citations, and conclusions yourself.
How to use Andri safely
Verify citations — Check that case law and statutes actually support the conclusions drawn
Review reasoning — Evaluate the AI's legal analysis critically, as you would a junior colleague's work
Cross-check facts — Confirm the AI extracted information correctly from your documents
Use track changes — When generating documents, request track changes so you can review every edit. See Using Templates and Actions
Give feedback — Use the dislike button when you spot errors to help improve accuracy. See Using the Dislike Button
Think of Andri as a well-informed assistant, not a replacement for your legal judgment. Use it to draft, research, and organize—then apply your expertise to verify and refine.