14/07/2026
"Our chatbot doesn't just look for a perfect, rigid match; it uses a Two-Layer Smart Search Engine to understand what your customers are trying to say, even if they use long sentences or make typos."
Layer 1: The Substring & Word Grabber (Instant Match)
What it does: It looks for broad overlap. If a customer types a whole sentence like "I want to buy a nissan" or just searches a short term like "nissan", the bot scans your database rows (like "nissan x trail").
The result: If the customer's text contains your keyword, or if your keyword contains the customer's text, it triggers an instant match. It also self-corrects for phrase lengths so that longer, more specific targets (like "technical seo") get served before shorter words (like "seo").
Layer 2: The Typo & Similarity Doctor (Fuzzy Matching)
What it does: If there isn't a direct overlap, the bot uses an algorithm called Levenshtein Distance (we call it a similarity score). It calculates how many single-letter edits (additions, deletions, or substitutions) it takes to turn the user's input into your keyword.
The result: It grades the match on a scale from 0% to 100%. If the score is 75% or higher, the bot says, "Close enough!" and serves the answer anyway. This means if a user types "nisaan" instead of "nissan", the bot handles the typo seamlessly without breaking.
Try free with your own keywords
Launch your chatbot on your website for free — upgrade anytime to unlock more keywords and richer, multi-answer responses