Luris LogoLuris

AI Policy

Last Updated: 29 June 2026

1. Purpose of This Policy

This AI Policy explains, in plain language, how Luris's artificial intelligence features work, what they are built on, and — just as importantly — what they cannot do. We publish this separately from our Terms and Conditions because we believe lawyers evaluating an AI legal tool deserve a clear account of its methodology, not just a liability disclaimer.

2. Our Approach: Grounded Analysis, Not Open-Ended Generation

Generic AI chatbots answer legal questions from whatever they absorbed during training, with no guarantee that a cited case is real, decided the way claimed, or still good law. This is the principal source of AI "hallucination" in legal contexts.

Luris is built differently. At the core of our platform is a proprietary databank of Indian Supreme Court and High Court judgments. Our AI Features are designed so that large language model reasoning operates on top of this verified databank rather than answering from the model's own unverified training memory. The databank functions as a verification layer: the AI is directed to ground its analysis in retrieved, real judgments rather than recalling case law from memory.

3. The Judgment Databank

Our databank consists of Indian Supreme Court and High Court judgments, sourced from public, official repositories of court records. As of the date of this Policy, it includes tens of thousands of Supreme Court judgments, with High Court coverage being added on an ongoing basis.

The databank is a living resource: it is being continuously expanded and refined, including efforts to improve case title accuracy, semantic search capability, and coverage of additional courts and years.

4. The Role of Large Language Models

We use large language models (currently Google's Gemini) to perform reasoning, synthesis, and natural-language explanation over material retrieved from our databank and the inputs you provide. The model's role is to analyze and reason — for example, to assess how a fact pattern compares to precedent, or to explain an agreement clause in plain language — not to independently recall case law from its own training data without reference to our databank.

5. Human Intervention by Design

Our AI Features are deliberately designed to operate with human intervention where required, rather than as a fully autonomous decision-maker. This is reflected in how the Company itself describes its objects in its constitutional documents, and in product design: outputs are framed as analysis to inform a human decision-maker — you, or the advocate you consult — not as a final determination.

6. Limitations — What Our AI Cannot Do

We believe an honest account of limitations is as important as an account of capability:

  • No outcome guarantee. Case Prediction provides a probabilistic, analytical assessment based on historical patterns in similar cases. It cannot account for factors that only emerge during actual proceedings — witness credibility, judicial discretion, procedural developments, or settlement dynamics.
  • Databank coverage is not absolute. While extensive, our databank does not yet include every judgment from every court in India, and newly decided judgments take time to be added, verified, and indexed.
  • Not a substitute for professional judgment. Our AI Features support legal research and analysis; they do not replace the judgment, strategy, or advocacy of a qualified advocate.
  • Document analysis is not a substitute for legal review. The Agreement Analyzer highlights clauses and provides analysis, but does not guarantee identification of every legally significant issue in a document.

7. Your Responsibility When Using AI Features

We encourage you to treat AI Feature output as a starting point for analysis, not a final answer. Independently verify case citations and legal positions before relying on them, and consult a qualified advocate before making decisions based on any output from the Platform. See our Terms and Conditions, Section 5, for the formal disclaimer governing AI Feature use.

8. Data Used in AI Processing

Information about what data is submitted to AI Features (including third-party processing via the Gemini API) is set out in our Privacy Policy, Sections 4 and 7.

9. Reporting Inaccurate Output

If you believe an AI Feature has produced inaccurate, outdated, or misleading output — particularly a misattributed or incorrect case citation — we want to know. Report it via our Contact Us page. Identifying and correcting databank or methodology issues is an ongoing part of how we maintain accuracy.

10. Updates to Our Methodology

As our databank grows and our AI methodology evolves, we will update this Policy to reflect material changes, so that the description here remains an accurate account of how the Platform actually works at any given time.

11. Relationship to Other Policies

This Policy supplements, and should be read together with, our Terms and Conditions (Section 5 in particular) and our Privacy Policy.

12. Contact

Lex Coverage and Advisory Private Limited H No. 6, LGF, Vikram Vihar Extension, Lajpat Nagar (South Delhi), New Delhi – 110024, Delhi, India Email: support@luris.in