Legal / Litigation AI (Brazil) · Deep dive

Enter

The São Paulo legal-AI company running mass litigation for Brazil's biggest banks and airlines — Latin America's first AI unicorn, 32 months after founding.

emerging

The question that decides it: Can a company whose disclosed ARR was ~R$50M (~$9–10M) in 2025 grow into a $1.2B valuation before global legal-AI capital (Harvey at $11B+) or Brazil's own incumbents localize the same playbook?

My take ★★ Very interesting ★ Interesting − Maybe ✗ Not interesting

Snapshot

Enter is a São Paulo company that sells AI agents for mass litigation — the tens of thousands of near-identical consumer and labor lawsuits that Brazilian banks, airlines, and marketplaces defend every year. Its platform, EnterOS, ingests cases directly from court APIs, reads the full case file, screens for fraud, models settlement economics, and drafts the defense, with partner-firm attorneys auditing and signing every filing. Founded in September 2023, it processed 300,000+ lawsuits in the year to May 2026 for 40+ enterprises including Nubank, Itaú, Bradesco, Mercado Livre, and LATAM Airlines, and its $100M+ Series B at a ~$1.2B valuation made it Latin America’s first AI unicorn — 32 months after founding. It matters now because it is the cleanest test of whether vertical AI can monetize an emerging market’s structural dysfunction: Brazil’s ~75M pending lawsuits are not a bug to Enter; they are the TAM.

Founding story

The founder biography reads like a screenwriter’s draft. Mateus Costa-Ribeiro entered law school at the University of Brasília at 14, became Brazil’s youngest practicing lawyer at 18, argued and won before the Supreme Federal Court, took a Harvard Law LL.M., passed the New York bar at 20, and practiced capital-markets law at Milbank in New York. He won a full Knight-Hennessy scholarship to Stanford’s MBA and quit after one quarter, reportedly after noticing that whenever he explained the Brazilian judiciary to Americans, they heard an AI market where he had been taught to see a national embarrassment. The technical side came from Wildlife Studios, the $4B Brazilian gaming company: Michael Mac-Vicar was its co-founder and CTO for ~13 years; Henrique Vaz — olympiad medalist, Harvard CS, Tesla engineer — rose from intern to CMO there. Pre-seed investor ONEVC says it “played a key role in building the founding team,” which is the polite way of saying the lawyer and the gaming executives were introduced, not childhood friends. The company launched as Talisman AI in September 2023 and rebranded to Enter — tagline “Just press Enter” — before its public emergence in 2025.

How it works

When a new lawsuit is filed against a client, EnterOS pulls it automatically from Brazil’s fully electronic court systems via API and tags it by product, claim type, and amount. For labor cases it assembles a claimant profile from the client’s HR systems — pay stubs, timesheets, severance-fund (FGTS) records, collective-bargaining agreements. Agents then read the entire case file, including audio and video evidence, and enrich it with external data: in airline-cancellation suits, the system has pulled historical weather records to rebut force-majeure claims. Every case runs through 30+ fraud and abusive-litigation checks — tampered proofs of address, invalid powers of attorney, serial plaintiff-attorney collusion, duplicate suits, even deceased plaintiffs. The platform computes worst-case exposure and an ideal settlement figure (recalculated on every new filing), recommends fight-or-settle, and drafts a claim-by-claim defense with specific legal arguments, reusing expert reports across cases from the same cost center to avoid redundant inspections. Licensed attorneys at partner law firms audit and sign everything — a structure that keeps Enter on the right side of Brazil’s bar rules, which reserve court filings for lawyers. The company says it processes ~20 billion tokens a day (InfoMoney, May 2026) and holds SOC 2, ISO 27001/27701, and zero-data-retention agreements with OpenAI and Anthropic.

Product and business overview

EnterOS is packaged as two litigation products and four industry solutions. Consumer/civil litigation — the original wedge — covers intake through appeals for the mass consumer suits that dominate Brazilian dockets. Labor litigation (five clients as of May 2026) runs “from calculation through hearings,” including hearing prep with witness selection from org charts and post-hearing feedback. Industry packaging targets aviation, finance, technology, and retail. The buyer is the enterprise legal department, not the law firm: Enter’s pitch is provision reduction and win-rate lift, and its published case studies claim +6 points of win rate at Nubank, +30% at LATAM, and 5,000+ AI-drafted defenses at Banco BMG. Growth is expansion-led rather than logo-led — Sacra notes Enter grows by taking a larger share of each client’s litigation book (reportedly ~30% of consumer suits at major clients, targeting 60%) and adding claim types, which is why average contracts reportedly run ~$750K a year.

Business model and pricing

Revenue is hybrid: a recurring technology fee plus success fees tied to case outcomes — roughly 30% of revenue is performance-based (Rio Times, May 2026). That structure matters: it aligns Enter with provision reduction rather than seat counts, and it is something neither Harvey nor traditional legal software runs. There is no public pricing page and no per-case rate anywhere — an enterprise-only sales motion — but LatamList (Sep 2025) pegged average annual contracts around $750K, and the company said in late 2024 it had signed its fifth $500K+ ARR contract. Deployment cost reportedly falls below 20% of a contract’s gross revenue at maturity (Sacra), implying real software margins despite the human-audit layer.

Traction over time

• Nov 2024 — fifth $500K+ ARR enterprise contract; ~$10M contracted ARR claimed in the Sequoia-era narrative.
• Mar 2025 (seed announcement) — 13 enterprise customers, including Nubank, C6, SulAmérica, BMG, Agibank, Vivo, Light.
• Sep 2025 (Series A) — 250,000+ lawsuits managed; ~15 public clients; average share of a client’s cases up more than 4x since onboarding.
• 2025 full year — R$50M ($9–10M) annualized ARR, described by InfoMoney as 13x growth for the year.
• May 2026 (Series B) — 300,000+ lawsuits a year; client base tripled to 40–45+ enterprises; revenue up “more than 10x” year-over-year per the company; ~100–115 employees (85%+ technical), hiring toward 200.

Every number above is company-sourced or company-adjacent; no independent revenue verification exists.

Market analysis

Brazil’s judiciary is the product’s tailwind and its raison d’être. The official CNJ census counted 62.2M pending state-court cases (Dec 2024) plus ~5M labor cases; company materials round the total to 75–80M — roughly 8x the US caseload with two-thirds the population, at ~$30B a year in judicial operating cost (~1.6% of GDP). Labor filings hit ~2.1M in 2024 and are projected above 2.3M for 2025, the highest since the 2017 labor reform. Three structural forces feed the machine: courts went fully electronic and free to file, plaintiff-side AI is now increasing filing volume (Rest of World, 2025), and the judiciary’s own AI adoption is speeding dockets — all of which grows defense-side demand. The regulatory overhang is real: Brazil’s AI bill (PL 2338/2023) passed the Senate with EU-style risk tiers and fines up to R$50M; legal AI could plausibly be classified high-risk, and Enter’s CEO has already testified against parts of it.

Competitive intel

The competitive set is in the frontmatter table; the shape of it is three-sided. Domestically, Finch Soluções (out of Brazil’s largest mass-litigation firm, ~1M lawsuits under management) and Softplan (which builds the courts’ own case-management systems) are the incumbents with installed bases and older automation stacks; Jusbrasil owns the deepest legal-data corpus in the country but monetizes lawyers, not enterprise defense. Down-market, Jusfy ($15M Series A with Thomson Reuters Ventures) arms individual lawyers with agents — including the plaintiff-side attorneys filing against Enter’s clients. Globally, Harvey at $11B+ and ~$300M ARR is the loaded gun: no confirmed Brazil operation today, but its capital and model access dwarf Enter’s, and localization is a hiring problem, not a physics problem. Enter’s real moats are its court-API integrations, fraud detection tuned to Brazilian scam patterns, the attorney-audit structure that satisfies OAB rules, and success-fee contracts that shift risk onto itself.

History and evolution

• Sep 2023 — founded in São Paulo as Talisman AI; ONEVC pre-seed; company claims cash generation in year one.
• 2024 — enterprise contracts stack up quietly; Sequoia closes the seed (announced Mar 2025 — its first Brazil check since Nubank, 2013).
• Feb 2025 — OpenAI approves Enter for zero data retention.
• Sep 24, 2025 — $35M Series A at $350M, co-led by Founders Fund and Sequoia; billed as Latin America’s largest AI round to date.
• 2025 — CEO testifies before Brazil’s Senate against provisions of the AI bill.
• May 5, 2026 — $100M+ Series B at ~$1.2B led by Founders Fund; Ribbit and Kaszek join; valuation up ~3.4x in eight months; first team-wide employee secondary at Series B stage in Brazil, per the company.
• H2 2026 (planned) — international expansion into undisclosed markets; stated ambition (unchallenged in press) to be a “monopolistic force” in Latin American legal AI.

No pivots, layoffs, or public crises found as of August 2026 — which after 32 months mostly means the record is short.

What people say

The case for. Named legal executives at Nubank, LATAM, SulAmérica, and Banco Mercantil supply specific, quantified praise — +6 points of win rate and a 2.5x-faster-than-planned rollout at Nubank (“Enter did in days what would have taken us months,” per Nubank’s dispute-resolution director), +30% win rate at LATAM, structured extraction from 500,000 documents at SulAmérica. The investor case — Sequoia returning to Brazil after twelve years, Founders Fund leading twice — treats Brazil’s litigation volume itself as the moat. The team-wide secondary at the Series B bought real employee goodwill.

The complaints. The loudest finding is silence. There are no G2, Capterra, Trustpilot, or Product Hunt reviews, no matching Glassdoor profile, zero Hacker News hits, and no Reddit threads — meaning there is no independent user validation of any company claim, and the press coverage is uniformly promotional (Rio Times let the “monopolistic force” line pass without a follow-up question). The structural criticism sits one level up: Brazil’s consumer bar argues the entire “predatory litigation” framing Enter’s fraud-detection pitch rests on is a defense-industry invention, and a 2025 CNJ-linked study found anti-abusive-litigation measures mainly harm vulnerable claimants. Enter is squarely on the corporate side of that contested debate — a reputational and regulatory watch item, especially with an AI bill in the Chamber of Deputies.

Outlook: the open question

The question is whether revenue can catch the valuation before the window closes. Disclosed 2025 ARR of R$50M ($9–10M) against a $1.2B valuation implies a triple-digit forward multiple that only compounding 10x-ish growth for two more years can justify. The bull case has clean answer-conditions: Enter goes from ~30% to ~60% of case share inside 40+ locked-in enterprises, the labor product scales the way the consumer product did, success-fee economics keep net revenue retention expansion-led, and international expansion proves the machine works on at least one more litigious civil-law market (Mexico, Colombia, Indonesia) before Harvey or a localized rival arrives. The bear conditions are equally specific: growth is concentrated in a small number of ~$750K contracts with sophisticated buyers who can multi-source; the courts or the AI bill restrict automated defense at scale; the predatory-litigation backlash turns Enter’s fraud screening into a political liability; or Brazil’s litigation volume — the entire TAM thesis — finally gets reformed downward. A company built on a national dysfunction owns a real market exactly as long as the dysfunction outlasts the round.

How to attack it

The wedge is the plaintiff’s side of the same docket. Enter is structurally locked into corporate defense — its 40+ clients are the banks and airlines being sued — which leaves the other side of every one of Brazil’s ~15M new consumer suits a year unserved by comparable tooling. A challenger that gives plaintiff firms Enter-grade drafting, evidence enrichment, and settlement modeling (Jusfy is closest, at 1/80th the valuation) doesn’t just build a business; it inflates Enter’s clients’ loss rates and turns Enter’s own success-fee revenue against it. Head-on, the exploitable weaknesses are concrete. First, concentration: ~$750K average contracts across ~40 clients means a handful of logos are a material share of revenue, and those buyers — Itaú, Bradesco, Nubank — are precisely the companies capable of building in-house or dual-sourcing to Finch to compress pricing. Second, the human-audit dependency: partner law firms sign every filing, so a rival that convinces those same volume firms to adopt its own copilot (or a firm like Nelson Wilians building tech internally) controls Enter’s compliance chokepoint. Third, the model layer is rented: Enter runs on OpenAI and Anthropic APIs; it owns workflow and integrations, not foundation models, and Harvey’s $300M-ARR war chest could fund a Portuguese localization plus loss-leader pricing that no $140M-funded company can match. Fourth, the “30+ fraud checks” pitch is a regulatory glass jaw — one high-profile case of a legitimate claimant wrongly flagged, amplified by the consumer bar’s existing campaign against the predatory-litigation narrative, invites both press and the AI bill’s high-risk classification.

Adjacent-segment play

The core asset — court-API ingestion, claim-level document extraction, fraud screening, settlement modeling with a success-fee wrapper — generalizes along three credible axes. Geographically: any civil-law country with electronic courts and mass consumer litigation (Mexico, Colombia, Argentina, Indonesia, India’s consumer forums) is a candidate, and Enter itself has signaled H2 2026 expansion — the open question is whether it can rebuild court plumbing country-by-country faster than local clones appear. By buyer: the same engine sold to insurers as a claims-litigation product is nearly frictionless — SulAmérica is already a client — and a dedicated insurance-claims-defense package would compete with claims TPAs rather than law firms, a larger and softer market. By segment: mid-market Brazilian companies with 500–5,000 suits are priced out of ~$750K contracts; a self-serve or channel version (sold through the volume law firms themselves) would open the long tail, though it would cannibalize the partner-firm relationships Enter depends on for signatures. The least attractive axis is elite legal work — Harvey and Legora already own high-end drafting, and Enter’s advantage is volume infrastructure, not prestige workflows. The most dangerous version of this play is someone else running it: a Jusbrasil or Thomson Reuters-backed Jusfy packaging plaintiff-side or mid-market variants before Enter’s land-and-expand locks the market.

Sources and further reading

• Enter — company site, product and client pages — getenter.ai, fetched Aug 25, 2026
• Enter raises $100M+ Series B, becomes Latin America’s first AI unicorn — Enter, May 5, 2026
• Brazilian AI legal startup Enter valued at $1.2 billion in round — Bloomberg, May 5, 2026
• Founders Fund and Sequoia back Brazilian AI startup Enter at $350M valuation — GlobeNewswire, Sep 24, 2025
• Quem é a Enter, unicórnio brasileiro de IA do setor jurídico — InfoMoney, May 7, 2026
• Enter: Brazilian legal AI unicorn at $1.2 billion — Rio Times, May 7, 2026
• Enter raises $37.5M led by Founders Fund — LatamList, Sep 25, 2025
• Sequoia is back in Brazil, co-investing with ONEVC — BrazilCham/ONEVC, Mar 24, 2025
• Justiça em Números 2025 — executive summary — CNJ (Brazil’s National Council of Justice), Oct 2025
• Legal AI startup Harvey raises $200 million at $11 billion valuation — CNBC, Mar 25, 2026
• O mito da litigância predatória contra bancos: o que dizem os dados oficiais — Conjur, May 2, 2026