Tokyo Startup Law Office boosts email inquiries 7.6x with a generative AI legal-consultation chatbot

Tokyo Startup Law Office struggled to convert website visitors into inquiries. Staffing a live-chat team to offer 24/7 human consultation was considered but seen as too costly and resource-intensive, and many prospective clients felt too anxious or unsure to reach out to a lawyer at all, causing serious cases to drop off before ever making contact.

The firm deployed BOTCHAN AI, a generative AI chatbot (combining ChatGPT/GPT-4) built by wevnal, on its website to handle initial legal consultations and route users to an inquiry form. The vendor tested for hallucination risk across roughly seven risk categories before launch, and trained the model in three tiers: public website "knowledge information," internal workflow/manual "knowledge base information," and transcribed, GPT-4-rewritten recordings of the firm's own empathetic client conversations, so the bot could respond with a warm, human-like tone (e.g., offering condolences on an inheritance inquiry) rather than mechanical scripts.

Email inquiries flowing from the AI-chat-to-form funnel increased 7.6x, shifting volume away from costlier, resource-intensive phone calls and producing better-organized inquiries that were easier to act on. A banner-copy test found that framing the CTA as "Try asking the AI for advice" roughly doubled click-through rate versus a generic "Get free advice" CTA. The chatbot also surfaced previously invisible ("silent") user concerns even when they didn't convert, which the firm used to iteratively refine the bot's responses and creative.

• Framing the CTA around "asking the AI" rather than a generic offer lowered the psychological barrier to a legal inquiry and roughly doubled click-through rate.
• Tiered training data — public site content, then internal manuals, then transcribed empathetic staff conversations rewritten by GPT-4 — was key to producing a human-feeling tone suited to a sensitive legal context.
• Routing inquiries through AI chat to email (instead of phone) reduced staff/resource strain while improving inquiry quality and completeness.