Turning Web Chat Into a Real Lead Channel: Jack Hunt Gold and Silver's AI Assistant Converts Nearly 1 in 25 Conversations
Overview
Jack Hunt Gold and Silver's web traffic kept growing, but it wasn't turning into transactions. The Incept team built and trained a member of their AI workforce to meet visitors with real answers the moment they landed on the site, instead of a static contact form.
The Opportunity
Growing web traffic wasn't converting into transactions. Precious metals buyers and sellers want to know what their gold or silver is worth, whether the store buys what they have, and whether they can trust the answer, before they ever walk in the door. A chatbot that gives vague answers, or worse, promises a price it can't back up, does more harm than good in that kind of business.
The Move
Incept built a purpose-made AI chat assistant for Jack Hunt Gold and Silver's website, trained specifically on how precious metals buyers and sellers shop. The assistant answers common buying and selling questions, builds enough trust to keep a visitor talking, and knows exactly when to hand a serious inquiry to a person instead of guessing. Clear guardrails keep it from ever quoting a firm price or promising a value that only an in-person appraisal can give.
The Win
- 57,823 unique visitors engaged by the chat assistant
- 6,274 messages exchanged
- 988 real conversations started
- 39 leads delivered to the CRM, a conversion rate right around 4%
Strategy
A purpose-built AI chat assistant, mapped to how precious metals buyers actually shop and tested against real conversations before launch.
Conversational AI Build
We defined exactly what the assistant needed to do for Jack Hunt Gold and Silver: answer common buying and selling questions, build enough trust to keep a visitor talking, and know when to hand a serious inquiry to a person instead of guessing. From there we mapped the real questions visitors ask a gold and silver dealer, from current pricing to how appraisals work.
Testing & Guardrails
We organized approved content and set clear rules the assistant should never break, including quoting a firm price or promising a value only an in-person appraisal can give. Before launch, it ran through Incept's break-the-bot testing: simple questions, vague questions, and the odd, high-pressure ones a real visitor might ask at 11pm.
Answer clearly before asking for anything, and know the difference between an estimate and a real offer.
Creative Approach
The assistant's job was never to close every conversation. It was to make sure the right ones got through. It answers before it asks, so the visitors who are serious enough take the next step on their own, with no pressure tactics and no premature asks.
Conversational AI
Guardrails & Trust
Lead Qualification
Human Handoff
Unique Visitors Engaged
October 2025 – April 2026
Conversations Started
From 6,274 messages exchanged
%
Conversation-to-Lead Conversion
39 leads delivered to the CRM
That rate matters more than it looks at first glance. A 4% conversation-to-lead rate, with no pressure tactics and no premature asks, is a strong return on nearly six thousand messages, especially for a purchase decision most buyers don't make on impulse.