Signal › Contributors

Vanessa Torres

I was a litigation associate at Cravath, Swaine & Moore for four years, from 2016 to 2020. For those unfamiliar with BigLaw, Cravath is one of the most prestigious law firms in the world, and being an associate there means billing 2,200+ hours per year, which translates to roughly "you live at the office and your apartment is where you keep your dry cleaning."

I was a good lawyer. I was also miserable. Not because of the work itself, which was intellectually stimulating, but because of the inefficiency. I once spent 300 billable hours on a document review that a well-trained ML model could have done in 48 hours. The firm billed the client $225,000 for that review. The client paid it because that's how the industry works. The incentive structure in legal services actively punishes efficiency: if you solve a problem faster, you bill fewer hours, and you make less money. This is insane.

I left Cravath in 2020 and joined a legal tech startup called Everlaw as a product manager. Everlaw makes e-discovery software, the technology that helps lawyers search through millions of documents during litigation. Going from practicing law to building legal technology was like going from being a passenger on a bus to seeing the bus's engine for the first time. Everything I'd found frustrating as a lawyer became a product problem I could solve.

After Everlaw, I joined a Series B legal AI company as VP of Product. I won't name them because I left on complicated terms. What I will say: the legal AI space is simultaneously overhyped and genuinely transformative. The technology works, it really does, but the adoption barriers are enormous. Law firms bill by the hour. Legal AI reduces hours. You can do the math on why adoption is slow.

I started writing about legal tech because the coverage was uniformly terrible. Tech journalists don't understand law. Legal journalists don't understand technology. The result is articles that either claim "AI will replace lawyers" (it won't, for structural reasons I've written about extensively) or "lawyers will never change" (they will, but slowly, and only when economics force them to).

I'm based in Miami. I'm still a licensed attorney in New York, I teach a legal technology seminar at FIU Law, and I salsa dance competitively, which is the only activity I've found that's more stressful than oral arguments.

Articles by Vanessa Torres (4)

Travel AEO: When AI Plans Your Trip, Who Gets the Booking?Wikipedia is the number one secondary citation source in AI assistants. If your brand has no Wikipedia presence, you are structurally invisible to the · May 25, 2026Web Components and AEO: When Shadow DOM Hides Your Content from AI CrawlersWebinars get watched once. Transcripts get cited forever. The teams winning AI citation share from their webinar programs treat the recording as raw m · May 25, 2026On-Device AI Search and EdTech Privacy: How Local-First AI Reshapes AEO for Student BrandsPerplexity runs a partially-curated source library directory underneath its citation engine that most publishers do not know exists. The ones that do · May 25, 2026Internal Site Search Is Now an AEO Signal. Algolia + Vector Embeddings Win.The 2023 AngelList Talent rebrand to Wellfound buried half the SEO equity for startup-discovery queries. The half that survived is now feeding ChatGPT · May 26, 2026