Case study · Legal risk management

AI applied to the litigation of a national power utility

How Vert revolutionized legal risk management at a Brazilian power generation and transmission company — with automated capture, machine learning, NLP and a 360-degree view of litigation.

Client: Eletrobras Solution: CyndIA
Challenge

Legal risk hard to measure and expensive to provision for

  1. High cost of capital — court deposits and contingencies eating into cash.
  2. Scattered court publications — courts with no automated capture and no standardization.
  3. Anomalous data in the legal ERP — undermining the reliability of provisioning.
  4. Manual provisioning — rule-based, with no model for the probability that a claim is granted.
What we did

Six transformation workstreams

Machine learning

Calculating the probability that each claim is granted, enabling data-driven provisioning.

Automated capture

A web crawler monitoring the courts, integrating and cleaning data into the legal ERP.

NLP & text analysis

Analysis of relevant case updates and automatic classification by context.

Time series

Monthly and quarterly forecasting of litigation volume for financial planning.

Cross-analysis

Cross-referenced testimony, court deposits, and data-cleansing and integration costs.

Legal cockpit

Survival modeling, root-cause analysis and AI-assisted settlement optimization.

Results

A 360-degree view of litigation

  1. Automated integration of court publications — with no manual work.
  2. Lower cost of capital on litigation, by reassessing contingencies and court deposits.
  3. Predictive reassessment of risk across cases, correcting provisioning.
  4. Identification of anomalous data in the legal ERP, with statistical cleansing.
  5. AI-assisted settlements with quantitative and financial forecasting based on time series.
Benefits

A data-driven legal operation

Efficient calculations

Safer, more auditable provisions.

Deposit recovery

Proactive identification of recoverable court credits.

Settlements with AI

Assisted settlement with impact forecasting.

Overall statistical view

District, court, judge, hearings and outcome — all cross-referenceable.

Focus and priority

Action driven by probability × value × time to judgment.

Financial forecasting

Quantitative and qualitative, based on time series.

Want the same in your legal department?

Explore CyndIA — Vert's platform that runs legal operations end to end with AI.

Talk to Vert View all case studies