Sunday, 22 November 2026 Login

Justice. Rights. Recourse.

BREAKING
Tort Settlements

AI Speeds Up Contract Analysis, But Data Mess Persists

AI Speeds Up Contract Analysis, But Data Mess Persists - contract analysis
Catherine Carroll, head of legal operations at iManage, highlights AI’s role in contract analysis.

Legal teams can now use AI to query their contracts and get answers in minutes, according to a report by Catherine Carroll, head of legal operations at iManage.

Carroll explains that connectors built on open standards like the model context protocol (MCP) allow AI agents to directly query governed document stores. This means questions that once required multi-week projects can now be answered almost instantly.

The Pitfall: Messy Data

However, Carroll warns that AI does not clean up messy data. Teams that have already invested in classification, data hygiene, and permissions will be better positioned to scale their AI efforts. Those with disorganized repositories will struggle.

The mess often stems from scattered storage and inconsistent naming conventions. For example, a client’s paperwork might be filed under various names, making it difficult for AI agents to recognize connections. Version control issues, where multiple “final” versions exist, further complicate matters.

Carroll emphasizes that for firms handling privileged client data, disorganized repositories pose not just a housekeeping issue but also a regulatory, contractual, and reputational risk.

Read Also: Avoidable Requisitions Drop to Half Previous Levels

Groundwork Before Gadgets

Carroll advocates for a “groundwork before gadgets” approach. This involves establishing a solid information architecture before implementing AI tools. She outlines four key areas:

  • Single Source of Truth: Designate a central repository for signed agreements with strict data standards.
  • Machine-Readable Classification: Use consistent fields for document status and contract types.
  • Legacy Cleanup: Prioritize cleaning up high-value, live, and risky legacy contracts.
  • AI-Ready Permissions: Implement access controls tailored for AI agents, ensuring ethical walls and audit trails.

Carroll suggests a structured plan to get started. This includes a self-assessment, a focused pilot project, and gradual implementation of data hygiene practices.

Practical Steps for Implementation

To achieve the benefits of AI in contract analysis, legal teams should follow a structured approach. Carroll suggests starting with a self-assessment of the current state of contract data. Could you say today, without commissioning a review, how many live agreements you hold with your 10 largest counterparties, which copy of a given contract is the signed one, and which people and AI tools can access it?

If the self-assessment reveals gaps, the next step is to launch a focused pilot project. This project should address a specific business or client question. In the initial phase, designate a project owner, conduct a security and data protection review, select a risk tier, locate every relevant agreement and flag duplicates and obsolete drafts, having first checked retention rules and legal holds.

Next, implement a consistent naming convention and reconcile counterparty names. Require a status field for all new document uploads and audit access rights to ensure proper permissions. Finally, connect an AI agent to the system, test its responses against source documents, and establish criteria for accuracy and error rates.

Tags:

Leave a Reply

Your email address will not be published. Required fields are marked *