Launched this week

LYQN AI

Launched this week

AI support that learns your business and hands off to humans

50 followers

AI support that learns your business and hands off to humans

50 followers

Visit website

LYQN is a self-learning AI support agent. It crawls your site, grounds every reply in your knowledge base, hands off to a human in one click, and lets visitors continue the chat on WhatsApp. One-line install. 2 weeks free.

Free Options

Launch tags:Marketing•Vercel Day

Launch Team / Built With

Framer AI AgentsDesign and publish professional sites with AI

Promoted

Maker

📌

Hey Product Hunt 👋 We're LYQN a support chatbot that doesn't embarrass you. Most AI chat tools have the same failure modes: hallucinated answers, painful onboarding, and bills that hit $300/mo before a single ticket is resolved. We've been on the other side of that, and it's not fun. Here's what we did differently: Grounded answers only. LYQN crawls your site, ingests your docs, and learns from past conversations. It won't answer questions it doesn't know. Human handoff in one click. When the AI can't help, a teammate steps in with AI-suggested replies and the full conversation already in front of them. WhatsApp bridge. Visitors scan a QR and continue on WhatsApp. Your team stays in one inbox. One-line install. A single `

Report

8d ago

@joseph_akhatasebhudo The handoff to humans part is what makes this interesting. Most AI support tools try to avoid that moment, but it's often exactly where trust is built or lost. Curious how you decide when to escalate.

Report

2d ago

Maker

@mpanpalli The AI checks if it has a grounded source for the answer, meaning the response has to trace back to something it actually ingested from the site, docs, or past conversations. If it can’t find a confident match, instead of guessing, it triggers a handoff. Practically, this means a few signals trip the escalation: No matching content found in the knowledge base for the question. Low confidence score even if something loosely matches. The visitor explicitly asks for a human (“talk to someone”). Repeated rephrasing of the same question, meaning the AI already failed once. When any of these fire, LYQN hands off instantly with the full chat history attached, so the human teammate isn’t starting cold.

Report

2d ago

@joseph_akhatasebhudo The repeated rephrasing trigger is the smart one. Most tools wait for an explicit "talk to a human" but by then the user is already frustrated. Catching the loop early is the right call.

Report

2d ago