Icon

Customer Name Search Tool

<p><strong>Tool Description</strong></p><p>This tool performs <strong>fault-tolerant customer name lookup</strong> on sales data.<br>It takes a customer name provided by the agent, retrieves the stored dataset from Hub Space, and applies <strong>approximate string matching (Levenshtein distance)</strong> to handle typos, variations, or partial matches.</p><p>The tool then filters results to select the most relevant customer record(s), <strong>caches the matched customer data for later use</strong>, converts the output into structured JSON, and returns it to the agent. This enables the agent to seamlessly answer user questions with accurate customer information, while also allowing subsequent tools to reuse the cached customer data without re-running the full lookup.</p>

URL: exorbyte GmbH http://exorbyte.com/en

Load customer data
Reads the stored dataset from Hub Space.
Provides the reference table of customers for matching.

Tool input: customer name
Receives the query string (customer name) from the agent.
Converts it into a table row for matching.

Approximate Customer Matching

Uses Levenshtein distance to compare the input name
with all customer names from the dataset.
Outputs similarity scores and candidate matches.

Finalize and return tool output

  • Filter matches to keep only the most relevant customer row(s).

  • Cache the customer data for later use.

  • Convert results to JSON so they are structured and easy to parse.

  • Send response back to agent, enabling it to use the data in a natural-language answer.

Tool Input:customer name
String Configuration
Variable to Table Row
Levenshtien Matching
Approximate String Matcher
Read the data from Hub Space
Table Reader
Customer Data
Table to JSON
Filters results to keep only the best matching row(s)
Expression Row Filter
Tool output
Tool Message Output
Cache the customer data
Table Writer

Nodes

Extensions

Links