Work Smarter, Not Harder: Top 5 AI Prompts Every Legal Professional in Brazil Should Use in 2025

Brazilian legal professionals can no longer treat AI as a novelty - investments in AI and generative projects are projected to top BRL 13 billion, and Bill No. 2,338/2023 plus LGPD and ANPD guidance are reshaping duties around transparency, risk assessments and data use (see Artificial Intelligence 2025 – Brazil).

Precise prompts turn sprawling case files and contract libraries into LGPD‑safe, audit‑ready summaries and redlines, and real deployments - like WideLabs' Justice Intelligence platform that helped prosecutors tackle Brazil's massive court backlog - show how good prompts can cut months of work to minutes.

Learning to write effective, defensible prompts is now a practical compliance and efficiency play; operators can build those skills in courses such as the AI Essentials for Work bootcamp to produce reproducible research prompts, contract checklists, and DPIA inputs that regulators and clients will trust.

“The system can find inconsistencies in huge databases in a matter of minutes,” Sidou said.

Selection prioritized prompts that map directly to Brazil's emerging AI guardrails: the risk‑based obligations and algorithmic impact assessment (AIA) duties in Bill No.

2,338/2023, LGPD obligations and ANPD expectations for transparency and logging informed every choice. A three‑step filter was applied - regulatory relevance (must generate audit‑ready outputs for high‑risk uses and DPIAs), client value (must cut routine tasks like contract redlines and case‑law synthesis), and defensibility (must produce explainable, well‑documented outputs) - using sources such as a practical summary of the Bill and penalties on Securiti's analysis of Bill 2338/2023 risk-based AI framework and the sector‑by‑sector context in Chambers & Partners' Artificial Intelligence 2025 – Brazil trends and developments.

Prompts were also cross‑checked against the technical legislative analysis to ensure they surface LGPD risks, required explainability and contestability, and remediation steps - because in Brazil a missed compliance signal can translate into enforcement actions and fines up to R$50,000,000, so precision isn't optional.

Prompt 1 is a reproducible research prompt that turns messy filings, regulatory resolutions and enforcement reports into an LGPD‑focused case‑law synthesis: ask the model to compile recent ANPD decisions, cite relevant LGPD articles and ANPD resolutions (for example Resolution CD/ANPD No.

15 on incident reporting, No. 18 on the DPO and No. 19 on international transfers), flag enforcement precedents such as the preventive measures against Meta and X, and produce an audit‑ready summary that lists legal bases, required DPIAs, breach‑notification timelines and practical remediation steps with pinpoint citations to statutes and guidance (useful because ANPD has already compelled DPIAs and stopped AI training on scraped Brazilian data).

The output should also tier risks (high/medium/low), surface potential fines (up to 2% of annual revenue, capped at BRL 50 million per infraction) and produce a short, plain‑language note for counsel to share with clients.

For regulatory grounding and examples to feed the prompt see ANPD enforcement reporting and sector activity in ManageEngine's roundup and a practical LGPD primer at DLA Piper to ensure statutory accuracy and citation formatting.

“We are not waiting for the AI legal framework in Brazil to begin regulating. Article 20 of the LGPD already requires us to address automated decision‑making. For now, we are focusing on this article. When the law is enacted, it will bring additional obligations for regulating AI in Brazil, and that will lead to further regulatory actions.”

Prompt 2 turns contract review into a repeatable, LGPD‑aware workflow: ask the model to redline agreements against the Civil Code and Consumer Protection Code (CDC), identify LGPD triggers (legal basis, sensitive data, DPO obligations, retention), and produce an itemized risk checklist that calls out mandatory clauses for international transfers - including the ANPD's Standard Contractual Clauses (SCCs) and the Aug 23, 2025 implementation deadline - so counsel can patch contracts before audits or enforcement escalates.

The prompt should require citeable evidence (article and paragraph citations), surface whether a DPIA is advisable under Art. 38, flag breach‑notification language that meets the three‑working‑day rule and required recordkeeping, and estimate potential administrative exposure (fines up to 2% of Brazilian revenue, capped at BRL 50 million) if controls are weak.

A practical redline output will include ready‑to‑insert SCC language, suggested technical and organizational measures, a short plain‑language client note, and a prioritized remediation plan - because a single missing transfer clause can turn a routine outsourcing contract into a regulatory emergency.

Read the DLA Piper LGPD guidance and the Mayer Brown ANPD SCC implementation summary for templates and timing.

Prompt 3 - Analogous Case Analysis should ask the model to pull Brazilian restructuring precedents and map each decision to the statutory test in the amended Bankruptcy Law (arts.

69‑G to 69‑L): surface cases like Rede Energia, OGX/OSX, OAS, the Viver proceedings (where more than 60 entities were involved), Oi and Americanas, identify whether courts treated consolidation as procedural or substantive, and flag the critical “entanglement” requirement that many rulings have overlooked; for grounding see the detailed treatment of substantive consolidation in Global Restructuring Review and the broader institutional context in Economic Analysis of Jurisdiction: The Case of Brazil.

The prompt should require the model to extract facts (cross‑guarantees, control, identical corporate structure, joint market activity), state whether assets/liabilities were pooled or kept separate, note creditor‑voting outcomes, and summarize how jurisdictional practice (São Paulo's historical permissiveness versus pockets of restraint in Rio) shaped results - because a strategy memo that links a client's facts directly to the BBA's three‑part test can turn a sprawling docket into a focused litigation plan with clear win/loss triggers and remediation steps.

“The Brazilian Judiciary is considered by many people to be dysfunctional.”

Prompt 4 should produce a concise, client‑ready memo that translates Bill No. 2,338/2023 and LGPD obligations into an action plan: a plain‑language summary of the Bill's risk‑based regime (including when an algorithmic impact assessment is required for high‑risk or general‑purpose systems), the three basic rights for affected individuals (explanation, contestability and human review), and ANPD's coordinating role in the SIA, plus a clear list of governance must‑haves - preliminary risk classification, AIA and DPIA (noting they can be prepared together where efficient), logging/documentation, incident reporting and procurement clauses for model suppliers.

so what

The prompt should ask for a one‑paragraph that flags enforcement exposure (fines up to BRL 50 million or 2% of turnover and possible suspension or prohibition of systems), a prioritized remediation checklist for counsel, and ready‑to‑use client language explaining next steps.

Seed the prompt with the Bill overview and technical AIA guidance so outputs cite authoritative grounding (see the Bill summary at Artificial Intelligence 2025 – Brazil and the technical analysis from Data Privacy Brasil for AIA specifics).

Prompt 5 should turn an AI audit into an ANPD‑ready dossier: ask the model to generate a DPIA/AIA that maps processing flows to LGPD principles, enumerates technical and organizational measures, and logs the explainability tests, bias‑and‑fairness checks and counterfactuals used during model development so every decision is traceable for auditors; the ANPD Regulatory Agenda 2025–2026 highlights exactly this push to regulate DPIAs and AI governance, so seed the prompt with those agenda points from the INPLP summary of the ANPD agenda and with statutory requirements from the DLA Piper LGPD guidance to ensure legal bases and DPO interactions are documented.

Require the output to flag high‑risk vectors (biometrics, profiling of children, automated credit or hiring decisions), recommend mitigations, and produce a breach‑response annex that reflects the three‑working‑day ANPD notification window and the five‑year incident recordkeeping rule; tie those items to potential sanctions (warnings, public disclosure, suspension, fines up to 2% of turnover or BRL 50 million) so counsel can prioritize fixes that matter.

A well‑crafted audit prompt creates a single reproducible file - think of it as an audit trail you can hand to ANPD or the board, not a vague checklist - so demand citations, timestamps and machine‑readable logs.

For regulatory framing see the ANPD Regulatory Agenda and the practical DPIA and incident guidance at DLA Piper LGPD DPIA guidance and BreachRx incident response guidance.

Finish strong: translate these five reproducible prompts into firm practice by pairing clear governance with hands‑on training - start by aligning policies with the Brazil OAB AI Guidelines on Confidentiality and Due Diligence (Brazil OAB AI Guidelines on Confidentiality and Due Diligence), adopt a firm‑level GenAI integration framework to close the “AI adoption divide” highlighted in recent industry research, and build prompt‑writing muscle so teams can reliably produce audit‑ready outputs that save time (lawyers who adopt an AI strategy report measurable productivity gains, roughly five hours a week).

Make at‑risk systems subject to the Bill/LGPD checklist in the national practice guide (Artificial Intelligence 2025 - Brazil practice guide on LGPD and AI risks), require DPIAs/AIA logging for high‑risk uses, and train fee‑earners and compliance staff to run and vet prompts rather than outsourcing every analysis - practical training such as the AI Essentials for Work bootcamp (AI Essentials for Work bootcamp - prompt design, AIA/DPIA & reproducible workflows (15 Weeks)) teaches prompt design, AIA/DPIA inputs and reproducible workflows so counsel can hand auditors a traceable file, not a vague checklist.

The next step is operational: codify prompts into matter templates, require human review checkpoints, and start with one pilot matter so wins compound rapidly.