From zero to 142 captured leads in one week: how Mintwave built a 24/7 AI sales rep.

Mintwave builds embedded finance infrastructure for Southeast Asian e-commerce platforms. Their website attracted 4,200 monthly visitors — strong traffic for their stage — but conversion to demo requests was under 1.2%. A Hotjar analysis showed that 73% of visitors scrolled through the pricing page and left without acting.

The problem was timing: Mintwave's two founders handled all sales calls personally, but they were based in Singapore (GMT+8) while most inbound traffic came from the US and Europe during their off-hours. Visitors arriving at 2 AM Singapore time had no way to get answers to pre-sales questions and left before the team woke up.

Hiring a BDR was off the table at their runway stage. They needed a solution that could engage, qualify, and capture visitor intent at any hour without requiring headcount.

Mintwave's deployment was the fastest in this study: the full configuration took one afternoon.

They connected UnifiedRAG to HubSpot for lead capture, uploaded their product documentation, pricing FAQ, and a library of technical integration guides. The persona was configured as a knowledgeable "solutions engineer" — able to explain API concepts and answer technical pre-sales questions without sounding like a generic chatbot.

A lead qualification flow was built: the AI was prompted to collect company name, use case, and expected transaction volume before surfacing the "book a demo" CTA. Leads above a transaction volume threshold were routed to the founders via Slack alert with full conversation transcript.

The entire configuration — integrations, persona, qualification flow, Slack routing — took 4.5 hours from first login to production.

In the first seven days, 142 visitors engaged with the AI assistant in a qualifying conversation and provided contact details. Of those, 38 met the transaction volume threshold and received a same-day Slack alert to the founders. 11 became paid pilots within 30 days.

Off-hours conversion — visitors from US/Europe time zones — went from near-zero to a measured 3.4%. The AI's ability to answer technical questions (API rate limits, supported payment rails, compliance questions) distinguished it from simpler lead-capture widgets.

The founders reported that inbound leads from AI conversations converted to paid at 29% — higher than their outbound rate of 18% — because prospects had already been educated by the AI before the first human call.