A Startup Lawyer Vibe-Coded an AI Version of Himself

Why It Matters

The initiative shows in‑house legal teams leveraging generative AI to streamline contract negotiations, potentially reducing costs and accelerating deal cycles. It also signals a growing ability for non‑technical professionals to build custom AI agents, reshaping the legal‑tech landscape.

Key Takeaways

• Synthesia built a legal AI avatar in two weeks.
• Avatar handles early contract negotiation and FAQs.
• Uses internal policies, templates, and OpenAI’s ChatGPT.
• Illustrates “vibe coding” – non‑engineers creating AI tools.
• Could reduce workload for small in‑house legal teams.

Pulse Analysis

Synthesia, the $4 billion video‑avatar platform, has turned its own technology inward by giving its general counsel a digital counterpart. Gabe Stern spent roughly two weeks assembling a prototype that blends the company’s low‑code video builder with a customized version of OpenAI’s ChatGPT. The resulting avatar, dubbed Willow, can greet prospects, field routine contract questions, and walk users through Synthesia’s master subscription agreement—all while pulling from a curated library of internal policies, template clauses and negotiation playbooks. This rapid build showcases how “vibe coding” lets non‑engineers translate business needs into functional AI agents without a traditional development cycle.

The legal function stands to gain the most from such automation because contract negotiations often stall on repetitive administrative steps rather than substantive legal analysis. By delegating scheduling, FAQ handling and initial risk framing to an avatar, in‑house teams can accelerate the path to a signed deal and free senior counsel to focus on complex judgment calls. Yet credibility remains a hurdle; the current model relies on scripted tone rather than reading body language or exercising gut instinct, a limitation that could temper adoption among risk‑averse enterprises.

Willow’s debut hints at a broader shift where corporations build bespoke AI assistants tailored to their own policy frameworks, bypassing third‑party legal tech vendors. If the technology matures to interpret visual cues and adapt in real time, it could become a standard front‑line tool for procurement and legal departments, driving down transaction costs and reshaping the competitive dynamics of the legal‑services market. Competitors such as HeyGen and Colossyan may soon add similar capabilities, while traditional law firms could be compelled to integrate comparable AI front‑ends to stay relevant.