EN
Log in Create a bot
Home / How it works

How it works

Connecting takes two steps. The first lets you see how the bot answers before you install anything. The second takes a few minutes.

Step 1. Site analysis and testing

You enter your store’s address. The BotNow robot opens the site, reads what any visitor can see and follows the rules in robots.txt. More about it on the BotNow robot page.

What the bot learns about your store

What you see after the analysis

A “What the bot learned” screen: how many pages were read, how many products were found and where from, how many products have no price. A chat is open next to it. You ask the bot the same questions shoppers ask and look at the answers before it appears on your site.

About the demo. In demo mode the bot reads a limited sample of pages and products so that it finishes in about a minute. If your catalog has tens of thousands of items, the screen shows how many were studied. In the full version the bot learns the whole catalog.

Step 2. Installation on your site

The bot is embedded with one line of code:

<script async src="https://api.botnow.site/widget.js" data-key="your_key"></script>

How the bot answers a shopper

  1. The bot parses the question and decides where to look: in the catalog or on the site’s pages.
  2. In the catalog it searches with filters: by keywords, by price range, by category, with sorting. That is why requests like “the most expensive set” or “under $50” work as data queries, not guesses.
  3. The answer is built only from what was found. Prices and availability are quoted exactly. Product cards with photos and links appear under the answer.
  4. If the needed data isn’t there, the bot says so and offers to leave contacts.

What the bot does not do

What already works and what is in development

WorksIn development
Site and catalog analysis by address; test chat before installation; product recommendations with price filters; widget with product cards; lead collection with a consent checkbox; installation check and look settings; masking of personal data in messages before they reach the language model (tuned for Russian-language data so far); live operator conversation in the chat window; dashboard with leads and conversation historyCorrecting bot answers from the dashboard; scheduled catalog refresh; full study of large catalogs; multi-currency catalogs

Want to see how it looks on your store? Contact us.