← All episodes 
Episode
AI in Law: The Firm That Wants to Produce More, Not Hire More
Why are so many law firms chasing AI right now? Spoiler: it's not to replace lawyers. Caio and Marina break down how AI attacks the operational waste surrounding the lawyer, show who's already winning this billion-dollar race, and tackle head-on the risk nobody can ignore: law doesn't accept 'looks right.'
In this episode
01 Opening: The Question Has Changed
- Caio opens with the real pain showing up at firms: volume is growing but hiring more doesn't always fix it, because the problem isn't a lack of lawyers, it's a lack of process
- Marina pokes with the classic question: 'is AI going to replace lawyers?' and Caio flips the script: the real question is how much of the repetitive work still needs to pass through a person
- Drop the central thesis right away: AI in law doesn't replace the lawyer, it replaces the pile of tasks that keeps the lawyer from thinking
- Opening stat to anchor it: more than 90% of legal professionals surveyed by Wolters Kluwer already use at least one AI tool day to day, and 62% report saving 6% to 20% of their weekly time
02 Where AI Already Does Real Work (And It's Not a Robot Doing Hearings)
- Caio lists the concrete, boring uses that are actually the most valuable: case summaries, contract review, version comparison, drafting, deadline tracking
- Marina asks: 'but isn't that intern stuff?' and Caio uses it to make the point — that's exactly the high-volume, low-differentiation work that bottlenecks the team
- Practical examples you can picture: pulling data from PDFs, building document checklists, generating client reports, an internal firm knowledge base
- Insight: none of this is sci-fi, it's work that already runs today, the shift is to stop doing by hand what the machine already does well
03 The New Lean Firm: Growth Stopped Meaning Hiring
- Contrast the two growth models: traditional is more clients, more cases, more people, more cost; the new one is more clients, better workflows, more productivity per person
- Strong line from Caio: growth used to require hiring, now growth requires organizing knowledge
- Marina brings the human side: 'so a small firm can look big?' and Caio confirms — AI lets the small compete and the mid-sized stop hiring just to cover repetitive tasks
- Important caveat Caio insists on nailing down: AI alone doesn't fix a disorganized firm, if the process is bad it just scales the mess faster
04 The Billion-Dollar Race: Who's Already Doing It
- Caio maps the global players without it becoming a memorized list: Thomson Reuters with CoCounsel on top of Westlaw, Harvey already used by PwC, Dentons, Reed Smith and Deutsche Telekom
- The numbers that show the size of the bet: Legora raised 550 million dollars at a 5.5 billion valuation, Clio valued at 5 billion
- Symbolic case Caio highlights: Kirkland & Ellis, the largest firm in the world by revenue, partnered with Palantir to build its own platform, replicating senior partners' knowledge across more than a thousand lawyers
- Interesting twist with Wordsmith: targeting companies' in-house legal teams, the idea of keeping work in-house and depending less on outside firms, already serving Canva, Financial Times and BT with 70 million raised
05 The Risk Nobody Can Ignore: Law Doesn't Accept 'Looks Right'
- Caio grounds the excitement: AI hallucinates case law, invents citations, and produces convincing but wrong filings — in law that's not a detail, it's a risk
- Marina asks if it's actually happened and Caio brings the real case: courts have already sanctioned lawyers for fake AI-generated citations in filings, with penalties for hallucination and lack of transparency
- Brazil context few people know: the CNJ regulated responsible AI use in the judiciary through Resolution 615 of 2025, requiring governance, auditing, traceability and human supervision
- Stat that sets off the alarm: 45.8% of courts already use generative AI, but 57.6% of the bodies that use it rely on personal email accounts and 50% have no internal guidelines
- Closing message for the block: in law, AI can't be autopilot, it has to be an auditable copilot
06 Practical Wrap-Up: Turning Legal Knowledge Into a System
- Caio summarizes the implementation roadmap without hype: start with an operational diagnosis, map where the team loses time, then automate intake and reporting workflows
- The most valuable step in his view: internal AI trained on the firm's context, filing patterns, contract templates and procedures, plus a governance layer with data privacy, access control and logs
- Reminder about people: a lawyer doesn't need to become a programmer, but they need to be trained to use AI without dumping sensitive data into the wrong tool and without blindly trusting the output
- Strong closing from Caio: the new competitive edge won't just be knowing the law, it'll be turning legal knowledge into a system, and the firm that wins isn't the one that signs up for the famous tool, it's the one that redesigns the operation