Automation

AI Contract Management Accelerates Lease Review by 50%

By cutting review times in half, improving compliance visibility, and centralizing documentation, client achieved tangible productivity and governance gains.

Executive Summary

In 2023, client’s Midstream Division implemented an AI-powered contract management platform across its U.S. terminal network to automate the review and renewal of lease agreements. Within six months, the initiative cut lease-review cycle time by 50%, reduced manual legal workload by 35%, and established a centralized digital repository of over 40,000 legacy contracts. The project has since become a model for enterprise-scale document intelligence within client’s global logistics operations.

Background & Challenge

Client’s midstream business operates a large network of crude oil and refined product terminals across North America. Each facility manages a complex set of lease, storage, and throughput agreements, many of which contain unique commercial and regulatory clauses that require regular review and renewal.

Historically, contract review was manual and fragmented:
  • Individual terminals maintained localized copies of lease agreements in disparate systems.
  • Legal and commercial teams manually extracted key terms such as renewal dates, rate escalations, and indemnity clauses.
  • The average end-to-end review took three to four weeks per contract, often delaying renewals or triggering last-minute negotiations.
  • Version control was inconsistent, and metadata errors occasionally led to missed notice deadlines.
With over 40,000 contracts under management and more than 1,200 renewals annually, the process had become a significant operational bottleneck. In late 2022, the Midstream Projects Office commissioned an AI-driven solution to automate and standardize the lease-review process.

Objectives

The project set out to achieve four core goals:

  1. Reduce contract-review cycle time by 50% across all midstream terminals.
  2. Increase accuracy and consistency in key-term extraction and classification.
  3. Centralize legacy documentation into a searchable digital repository.
  4. Enhance compliance with corporate governance and regulatory audit requirements (FERC, EPA).

Solution Implementation

Technology Selection and Design

Clinet partnered with a SubSurfaceOps and deployed a customized Natural Language Processing (NLP) engine trained on midstream lease agreements. The model was fine-tuned to recognize over 120 contractual attributes—such as escalation clauses, throughput obligations, and environmental liabilities.

Architecture Overview:

  • Ingestion Layer: Batch upload of scanned and digital contracts from SharePoint and local drives.
  • OCR and Classification Engine: Multi-language OCR converting legacy PDFs into searchable text.
  • AI Extraction Model: Custom-trained LLM identifying key terms and mapping them to a standardized metadata taxonomy.
  • Validation Interface: Web portal enabling legal analysts to review, approve, or correct AI suggestions before final submission.
  • Integration Backbone: Seamless sync with client’s enterprise contract management system (Icertis CLM) and SAP for financial linkage.

Implementation Phases

  1. Pilot Phase (Q1 2023): Deployed at three terminals covering 5,000 contracts. The AI engine achieved 92% clause-recognition accuracy after iterative retraining.
  2. Scale-Up (Q2–Q3 2023): Rollout expanded to 25 terminals, with full migration of legacy contracts and user training for 70 legal and commercial staff.
  3. Enterprise Integration (Q4 2023): Model linked to renewal calendars and workflow automation, triggering proactive alerts for contracts nearing expiration.

Results & Key Metrics

After full deployment across the U.S. terminal network:

  • Lease-review cycle time reduced by 50%, from 21 days to an average of 10 days.
  • Manual review hours cut by 35%, allowing legal teams to focus on high-value negotiations rather than clause extraction.
  • Metadata accuracy improved by 25%, minimizing renewal errors and audit findings.
  • Renewal alert automation prevented 18 potential late-notice penalties during the first quarter post-launch.
  • Search and retrieval time for legacy contracts dropped from hours to seconds, supporting instant cross-terminal comparisons.
Financial modeling indicated a payback period of under 9 months, factoring in labor savings and avoided penalty costs.