Cecilia Ziniti spent nearly two decades as a corporate lawyer at Amazon, Cruise and Anki before a role at Replit gave her early access to OpenAI's GPT models. She saw a disconnect between what general-purpose AI could produce and what her profession actually required. In 2023, she left legal practice to build GC AI, a startup designed specifically for corporate attorneys.

What You Need to Know

GC AI targets the precision and sourcing needs of legal work, differentiating it from generic chatbots. Ziniti's background as both a lawyer and a customer shaped the product requirements. The company enters a market where domain expertise is becoming a critical advantage for AI startups. Co-founder Bardia Pourvakil brings engineering experience from Replit and a background that includes admission to Stanford Law School.

From Legal Counsel to Founder

Ziniti joined Replit as general counsel in 2021, when the coding platform was working with OpenAI on early AI-powered features. She gained access to GPT before ChatGPT launched and began exploring what the technology could mean for lawyers. She taught classes on using ChatGPT for legal research and noticed a persistent problem: the chatbot delivered friendly but imprecise answers that no lawyer could safely use in practice.

The Precision Gap

In one class, Ziniti asked ChatGPT to research legal considerations for entering the Brazilian market. The response covered relevant issues but began with language about visiting Rio and grabbing surfboards. The tone was wrong and the sources were missing. That moment clarified the gap between a general-purpose AI tool and a professional-grade legal assistant.

  • Accurate citations: Every legal claim must be tied to a verifiable source.
  • Source transparency: Users need to see where in the document the AI draws its answers.
  • Professional tone: Responses must match legal communication standards, not casual chatbot language.

Ziniti turned each shortcoming into a product requirement. She teamed up with Bardia Pourvakil, an engineer from Replit who had worked with earlier GPT versions. Pourvakil had once been admitted to Stanford Law School but chose engineering, partly because he expected AI to automate some legal work. Their backgrounds complemented each other: Pourvakil built the technology, while Ziniti knew exactly what it needed to do.

Why This Matters

GC AI represents a growing pattern in which domain experts, not just engineers, are founding AI startups. Ziniti was the ideal customer profile for the product she built, giving her an edge in defining requirements that general-purpose models miss. If the legal industry adopts specialized AI tools, the biggest winners may be founders who understand the profession from the inside. The shift also raises questions about how law firms and corporate legal departments will adapt to AI that can handle research, drafting and compliance work with higher accuracy than today's chatbots. Ziniti's bet is that lawyers will pay for reliability over friendliness.

A Co-Founder Who Chose Engineering Over Law

Pourvakil's decision to pursue software development instead of law proved prescient. He brought deep experience with GPT's earlier generations and a willingness to tackle the hard problems of making AI precise. Ziniti said she would teach classes in the morning and relay product specifications to him in the afternoon. The partnership allowed her to remain the product visionary without writing code.

Ziniti's journey from corporate lawyer to startup CEO shows that deep industry knowledge can be a stronger foundation for an AI company than technical background alone. GC AI is now live and serving corporate legal teams, aiming to prove that AI can be both powerful and trustworthy in the hands of lawyers.