Customer Insights Agent

<p><strong>AI Agent for Customer Insights</strong></p><p>This workflow enables an <strong>AI agent</strong> to interact with customer sales data in a conversational way. The agent is powered by an OpenAI LLM (GPT-4.1) and has access to two specialized tools for answering user queries:</p><ol><li><p><strong>Customer Name Search Tool </strong></p><ul><li><p>Performs <strong>fault-tolerant name lookup</strong> using approximate string matching (Levenshtein distance).</p></li><li><p>Handles typos, variations, and partial matches.</p></li><li><p>Returns the most relevant customer record(s) in structured JSON.</p></li><li><p>Also <strong>caches customer data</strong>, making it available for follow-up queries.</p></li></ul></li><li><p><strong>Data Aggregator Tool</strong></p><ul><li><p>Computes <strong>aggregated values</strong> such as totals, averages, minimums, maximums, counts, or medians.</p></li><li><p>Supports flexible <strong>groupby columns</strong>, <strong>measures</strong>, and <strong>aggregation functions</strong>.</p></li><li><p>Can aggregate either the <strong>cached customer subset</strong> (for customer-specific insights) or the <strong>entire dataset</strong>.</p></li><li><p>Returns results as structured JSON for agent interpretation.</p></li></ul></li></ol><p><strong>Flow:</strong></p><ul><li><p>The <strong>CSV Reader</strong> loads the customer sales dataset.</p></li><li><p>The <strong>OpenAI Authenticator</strong> and <strong>LLM Selector</strong> configure GPT-4.1 as the reasoning engine.</p></li><li><p>The <strong>AI Agent</strong> uses the dataset and calls the tools as needed to answer natural language questions</p></li></ul>