Evaluate AI Response Groundedness with LLM
This workflow takes an AI agent's response and a set of reference documents as input, potentially from a Google Sheet. It then uses a Large Language Model (LLM) to assess if the AI's response is solely based on the information provided in the reference documents, calculating a groundedness score. This helps identify potential AI hallucinations or unreferenced information.
Apps Used:
Google Sheets
OpenAICategories:
How It Works
- 1When a new row is added to a Google Sheet containing AI responses and reference documents.
- 2Use an LLM (e.g., OpenAI) to compare the AI response with the reference documents.
- 3Assess if the AI response is grounded only in the provided documents.
- 4Update the Google Sheet with the groundedness score.
Related Templates
Extract Data from Documents with Dynamic AI Prompts using Google Sheets
This workflow is triggered when a new row is added to a Google Sheet containing a document link. It processes the document (e.g., a PDF resume) by extracting its text content. Using prompts fetched from another Google Sheet, an AI extracts specific fields of information. Finally, the extracted data is updated back into the original Google Sheet row.
AI Resume Screening & Evaluation for HR with AI & Google Workspace
This workflow automates the screening and evaluation of job applications. When a candidate submits their application, their CV is uploaded to Google Drive. The system then uses AI to extract candidate information from the CV, retrieves the relevant job description from Google Sheets and Google Drive, and evaluates the candidate against the job requirements. Finally, it stores the results in Google Sheets and sends conditional notifications via Gmail or Slack.
Summarize Google Drive Files with AI
This workflow manually triggers, allowing you to process a specific Google Drive file. It retrieves the content of the document and then leverages an AI model to generate a concise summary. This is ideal for quickly understanding the core information in long documents stored in Google Drive.
Classify Customer Feedback Sentiment in Google Sheets with AI
This workflow triggers when new customer feedback is added as a row in Google Sheets. It then uses an AI model to classify the sentiment of the feedback as positive, neutral, or negative. Finally, the classified sentiment is updated back into the same Google Sheet.
Compare Claude and Gemini for PDF Information Extraction from Google Drive
This workflow is manually triggered to extract and compare information from a selected PDF document in Google Drive. It uses GenFuse's OCR capabilities to get text from the PDF, then sends this text to both Claude and Gemini LLMs for specific data extraction based on a custom prompt. The workflow provides the outputs from both LLMs for direct comparison.
Daily Stock Market Intelligence with Google Sheets & AI
This workflow provides a daily market intelligence briefing tailored to a user's stock holdings in Google Sheets. It triggers daily, fetches portfolio data, uses AI to research real-time market news, and analyzes its potential impact on holdings. Finally, it delivers personalized insights, risk alerts, and trade suggestions via a messaging app or email.