Frequently Asked Questions
Yes, the AI seller can match products based on images thanks to Visual AI technology. This is especially effective for industries where visual choice is key: clothing, furniture, home decor, jewelry, and building materials.
If you don’t have an automated catalog, you can upload data manually or via a spreadsheet (CSV): To train the AI employee, you can also use documents (PDF, DOCX) or simply provide links to product pages on your website.
Automatic updating via RSS or Google Merchant Feed operates as follows:
To synchronize an AI seller with a catalog and maintain data relevance, use automatic connection methods in the Price Lists section: To set up, simply specify the link to your XML file or Google Feed in the platform interface.
The Binom AI employee improves its skills through self-learning mechanisms and manual response corrections. If an employee encounters an unknown question, the system automatically records it and creates a draft knowledge card for subsequent completion by a human. In the Chats section, an operator can correct an inaccurate answer in Read more…
To ensure high accuracy and prevent hallucinations, Binom AI utilizes the following mechanisms:
To ensure answer accuracy, the Binom AI employee follows a strict information retrieval protocol. Search sequence when processing a request: Data conflict priority (from highest to lowest): If verified information is missing across all sources, the system will not invent an answer; instead, it will state that data is unavailable Read more…
To instantly update data without editing the entire knowledge base, use the following tools:
Yes, the Binom AI employee features Visual AI capability, which enables it to process not only printed text, but also images and handwritten documents. Key recognition capabilities: This functionality allows the employee to understand the content of uploaded files and answer customer questions based on the visual or handwritten information Read more…
To use a shared knowledge base without uploading files multiple times, use the Knowledge Sharing feature: