.png)
How Light runs its litigation with the support of Enter’s AI
+9 p.p.
+53%
+65%
+9 p.p.
+53%
+9 p.p.
“Enter set in motion an important change inside our legal department, one that made us one of the most digital areas in the company.”
+53% expansion in AI use across Light's lawsuits in under 12 months
+9 p.p. in operational efficiency on lawsuits supported by Enter's AI
65% operational efficiency in lawsuits where we flagged violations, versus 20% before Enter
Challenge
01
Scale
Light faces around 60,000 lawsuits per year in Rio de Janeiro, with 40% of them tied to power outages and energy theft. In outage lawsuits, for example, operational efficiency hovered around 20%.
Light already held solid technical inputs: field videos, expert reports, consumption histories and service records. The problem was not missing data. It was the difficulty of turning that data into clear, consistent proof the courts would recognize — at scale.
02
Fraud and violations
Beyond the sheer volume, Light operates in a landscape of abusive litigation and bad faith. In outage and energy-theft lawsuits, litigants routinely presented service tickets, WhatsApp screenshots and administrative records as evidence of a prior attempt to resolve the issue. The question was:
Light reviewed each lawsuit in isolation, with no historical cross-check, which made it impossible to spot violations at scale.
“12 pp more performance and a 25% reduction in average damages paid per lawsuit are results that meaningfully impact both the legal department's and the bank's bottom line.”
Solution
Enter organized and standardized the technical inputs Light had already produced:
- Videos Light technicians recorded during field inspections
- Consumption history, with analysis of individual patterns
- Service ticket records across every channel
- Expert reports
Beyond the technical inputs, Enter's AI agents read the full case file too, pulling out the information that matters for the defense.
“Com a Enter, construímos uma API que extrai subsídios de uma forma muito mais eficiente, rápida e assertiva.
Hoje, o advogado deixa de fazer isso manualmente e consegue aproveitar o tempo dele para analisar quais são os movimentos adequados para cada caso.”
Light suspected litigants were systematically reusing service tickets across separate lawsuits. Enter tested that hypothesis on more than 7,000 lawsuits.
The analysis found that 7% of the lawsuits showed clear signs of litigation fraud. These lawsuits shared identical ticket numbers even though different plaintiffs filed them and they referred to different homes.
The pattern grew even clearer through the concentration of these repeated tickets around the same legal representative, who used identical screenshots of supposed exchanges with Light in lawsuits about different installations.
Once Enter mapped the tickets, its AI compared every new lawsuit against the full history of prior actions, flagging on the spot the ones that had already appeared elsewhere.
In one specific lawsuit, the tickets the plaintiff presented were identical to those used in other actions against the company and had no connection to the outage claimed: third-party tickets, mismatched dates and no link to the event.
“With Enter's AI, we earned real credibility from the courts for the company's defense — confident that we're not filing generic answers, but finding the best input for each lawsuit.”
RESULTS
In under 12 months, Light gained 9 p.p. in operational efficiency on the lawsuits Enter supported, expanding AI use to more than 53% of its litigation.
Spotting replicated tickets let Light get ahead of baseless lawsuits and secure not just dismissals but formal recognition of bad-faith litigation — with a 5% fine on the claim value in one landmark lawsuit. In lawsuits involving bad-faith litigation, Light reached 65% operational efficiency.
