How the Quizell AI Sales Agent Works

Introduction

The Quizell AI Sales Agent is an AI-powered chat assistant that lives on your website and talks to visitors in real time. It can recommend the right product, answer common questions, and capture leads — all in a natural conversation, around the clock.

What sets it apart from a generic chatbot is that it's trained entirely on your content. It doesn't make things up from the open internet; it works from the quiz, instructions, and FAQ you give it. That keeps it accurate, on-brand, and focused on your business.

This article explains how the agent works conceptually — what it does, how it's trained, how a conversation flows, what it can and can't do today, and how it fits your store. If you're ready to build one, the step-by-step guides are linked at the end.

What the AI Sales Agent does

At a high level, the agent handles three jobs:

Recommends products — it can run a quiz conversationally to understand what a visitor needs, then point them to the right product or option.
Answers questions — it responds to visitor questions using the knowledge you've given it.
Captures leads — it can collect visitor details at the right moment, so high-intent visitors don't slip away.

You choose which of these to emphasize when you create the agent (Grow Sales, Automate support, or Generate leads), and you can shape its behavior in detail afterward.

How it's trained

The agent's "knowledge" comes from three sources you control, all found in the Train section:

Quiz — a Quizell quiz the agent runs inside the chat to guide visitors to a recommendation. This is the agent's structured recommendation flow.
Instructions — behavioral rules that shape how the agent responds: what to stay focused on, what to avoid, when to recommend the quiz, when to capture a lead, and how to handle things it doesn't know.
FAQ — specific question-and-answer pairs that give the agent the facts it needs to answer accurately.

Think of it this way: the quiz is the guided journey, the instructions are the agent's personality and rules, and the FAQ is its reference library. The agent draws on all three together in every conversation.

How a conversation flows

A typical interaction looks like this:

A visitor clicks the agent's trigger button on your site.
The agent opens with your welcome message.
From there, it either runs your quiz — asking questions one at a time, with answer options shown as tappable tabs — or answers questions freely using your instructions and FAQ.
Based on the conversation, it makes a recommendation, answers the question, or captures the visitor's details.

The experience is the same in preview mode and on your live site, so what you test is what your visitors get.

What it can and can't do today (Beta)

The AI Sales Agent is currently in beta, which means a couple of quiz features aren't supported yet:

No jump logic. The agent reads quiz questions in their natural order and ignores conditional jumps. For best results, connect a quiz that doesn't use jump logic.
No custom elements. Images, dropdowns, and other custom question elements don't render. The agent reads your answer options and displays them as simple tabs.

These are temporary beta limitations. The practical takeaway: pair the agent with a clean, straightforward quiz — plain questions, standard answer options, no jump logic.

How credits work

The AI Sales Agent runs on credits — a prepaid balance you top up and spend on live conversations.

Testing in preview mode is free. You can chat with your agent in the dashboard as much as you like without spending anything.
Live conversations use credits. Credits are spent per conversation, and the longer a conversation runs, the more it costs.
Your plan affects the rate. The higher your plan, the lower the credit cost per conversation.

You can check your balance, buy credits, and set up auto-recharge under AI Agent → Billing. For the full breakdown, see Understanding AI Sales Agent Credits & Billing.

Is the AI Sales Agent right for you?

It's a strong fit if you want to:

Grow sales by helping undecided visitors find the right product without leaving the page
Automate support by answering repetitive questions instantly, day and night
Generate leads by capturing visitor details during a helpful conversation

These map directly to the three objectives you pick when creating an agent — so you can tailor it to whichever goal matters most to your store.

Getting started

Building and launching an agent takes five short steps, each with its own guide:

Create your agent and set its goal
Train it with a quiz, instructions, and FAQ
Customize its design to match your brand
Test it in preview mode (free)
Connect it to your website to go live

What's next?

Start with How to Create Your First AI Sales Agent, and you'll have a working agent in minutes.

Updated on: 03/06/2026

Was this article helpful?

Thank you!