Skip to main content
Smart Spreadsheet Assistant Jp The Smart Spreadsheet Assistant template lets you interact with your spreadsheet data using natural language.
Upload Excel or CSV files, query data, apply filters, and generate instant summaries or insights, all powered by LLMs. This advanced version also includes a Prompt Optimizer Flow, enabling automated prompt refinement for better query accuracy and analysis quality.

​
Prerequisites

  • A valid Anthropic API Key
  • Spreadsheet file (Excel or CSV format)
  • Access to the Prompt Optimizer Flow (optional but recommended)

​
Key Features

  • Upload spreadsheets (Excel/CSV) for real-time querying.
  • Use natural language or SQL queries to explore data.
  • Connect to Anthropic’s Claude models for reasoning and summarization.
  • Automatically optimize and test prompt variations for accuracy.
  • Ideal for data exploration, reporting, and analytics.

​
1. Prompt Optimizer Flow 

Purpose: Enhance your spreadsheet interaction by fine-tuning natural language prompts for clarity and precision.

​
Flow Overview

  1. User Task – Define the data query or analysis you want to improve (e.g., “Show me the top 10 customers by sales”).
  2. Enhanced Prompt Optimizer – Generates, tests, and scores variations of your prompt.
  3. Results – Displays performance rankings of each optimized prompt.
Usage:
  • Start from the User Task node.
  • Run the Results component to evaluate performance.
  • Apply the top-performing prompt to your data analysis workflow.

​
2. Load Spreadsheet

Use the Load Spreadsheet component to import and prepare your data.

​
Steps

  1. Click Select Files to upload your spreadsheet (Excel or CSV).
  2. Configure the sheet name, chunk size, and indexing options.
  3. Enable caching for large files (optional).
  4. Output is automatically converted into a searchable DataFrame.

​
3. Spreadsheet Query Tool

Search, filter, or aggregate your uploaded spreadsheet data.

​
Supports:

  • Natural language queries (e.g., “Find all orders above $500”)
  • SQL queries (SELECT * FROM sales WHERE amount > 500)
  • Custom filters and sorting options
Outputs: Clean, query-ready data for the LLM to interpret.

​
4. Agent (Claude Integration)

Connects your query logic to Anthropic’s Claude Sonnet model for intelligent reasoning and conversational analysis.

​
Setup

  • Choose Model Provider: Anthropic
  • Select Model Name: claude-sonnet-4-20250514
  • Add your Anthropic API Key
  • Enter short Agent Instructions (e.g., “You are a helpful data analysis assistant that answers questions about spreadsheets.”)

​
5. Chat Interface

Use the Chat Input and Chat Output components to ask questions directly and view responses in real time.

​
Example

User: “Which region had the highest revenue last quarter?”
Assistant: “The West region recorded the highest total revenue of $120K between April and June.”

​
Example Use Cases 

  • Sales reporting (“Summarize total sales per month.”)
  • Expense analysis (“Find top 5 expense categories.”)
  • Research data filtering (“Show all entries with missing values.”)
  • Trend summaries (“Compare Q1 and Q2 growth rates.”)