Sheets Workflows

rtrvr.ai's Sheets Workflow feature provides a powerful way to automate complex tasks directly from Google Sheets. By providing rtrvr.ai with access to your sheets, you can instruct it to perform a series of actions, extract data from the web, generate content, and update your sheets with the results. This capability enables you to create sophisticated, multi-step workflows that can enrich your data with web information or perform direct AI enrichment.

Getting Started

To use Sheets Workflows, you have two options:

→Have Sheets open: Keep your Google Sheets open as the active tab or in your Selected Tabs
→Attach Google Sheets: Click the "Attach Google Sheets" icon to automatically grant rtrvr.ai access to your sheets and add them to Selected Tabs

Once your sheets are accessible, simply provide a prompt describing what you want to achieve. Our Planning agent will analyze your request and automatically set up the appropriate workflows to accomplish your objective!

Core Concepts

The Planning agent intelligently determines which underlying workflow to use based on your needs:

→Explore Workflow: Used when you need to extract data from paginated URLs or perform deep crawling through listings
→Research Workflow: Perfect for data enrichment tasks where you need to gather information from multiple sources per row
→Infer Workflow: Ideal for direct AI processing on your sheet data without web interaction
→Context Recordings: You can attach recordings to provide additional context for guiding rtrvr.ai's behavior
→Multi-Step Processing: The planner can create complex workflows that combine multiple approaches

How It Works

1. Provide Your Objective

Simply describe what you want to achieve with your sheet data. For example:

→"Enrich this list of companies with their employee count and latest funding information"
→"Extract product details from these Amazon URLs"
→"Generate personalized email introductions for each person in this list"
→"Find LinkedIn profiles for these company names and extract key information"

2. Let the Planner Work

The Planning agent will:

→Analyze your sheet structure and content
→Determine the most appropriate workflow type(s)
→Configure the necessary parameters
→Execute the workflow to achieve your goal

3. Review Your Results

rtrvr.ai will update your sheet with the requested information, adding new columns as needed while preserving your original data.

Underlying Workflows

While you don't need to select these manually, understanding the three core workflows helps you craft better prompts:

Explore Workflow

When it's used: When you have URLs that lead to paginated content or listings that need deep exploration.

How it works:

→Opens each URL from your sheet
→Extracts data from the current page
→Can navigate through pagination
→Can open linked pages (like individual product pages from a listing) for deeper data extraction

Example prompts:

→"Extract all products and their details from these category page URLs"
→"Get all article titles and summaries from these blog archive links"
→"Crawl through these search result pages and extract company information"

Research Workflow

When it's used: When you need to enrich existing data by researching across multiple sources.

How it works:

→Takes your primary data (e.g., company names)
→Researches across multiple websites
→Combines information from various sources
→Enriches your sheet with comprehensive data

Example prompts:

→"Find the LinkedIn pages for these companies and extract employee count and industry"
→"Research these startups on Crunchbase and add their funding information"
→"Look up these products on multiple e-commerce sites and compare prices"

Infer Workflow

When it's used: When you need AI processing directly on your sheet data without web interaction.

How it works:

→Processes data already in your sheet
→Applies AI analysis, generation, or transformation
→Works on individual rows or entire datasets
→No web browsing required

Example prompts:

→"Generate a product description for each item based on the features listed"
→"Classify these customer reviews as positive, negative, or neutral"
→"Create personalized email templates using the contact information and company details"
→"Summarize these meeting notes into key action items"

Using Context Recordings

You can enhance any workflow by attaching recordings that demonstrate specific actions:

→Record complex interactions: Show rtrvr.ai how to navigate tricky websites or perform specific actions
→Guide data extraction: Demonstrate which elements to extract when the structure is complex
→Handle authentication: Record login procedures for sites that require authentication

To use recordings:

→Create your recording using rtrvr.ai's recording feature
→Attach it when providing your prompt
→The planner will incorporate the recording into the workflow execution

Example Use Cases

Sales Lead Enrichment

Prompt: "Find LinkedIn profiles for these company executives and add their current role and company size"

→The planner uses Research workflow to find LinkedIn profiles and extract relevant information

E-commerce Competitor Analysis

Prompt: "Extract all product listings from these competitor category pages including price and specifications"

→The planner uses Explore workflow to crawl through paginated product listings

Content Generation at Scale

Prompt: "Write personalized outreach emails for each contact based on their company and role"

→The planner uses Infer workflow to generate content using existing sheet data

Market Research

Prompt: "Research these companies to find their latest product launches and press releases"

→The planner uses Research workflow to gather information from multiple sources

Data Classification

Prompt: "Categorize these customer support tickets by urgency and topic"

→The planner uses Infer workflow to classify data using AI

Configuration Options

While the planner handles most configuration automatically, you can influence the workflow behavior:

→Sheet Structure: Ensure your input data is in clear columns with headers
→Context Columns: Include additional columns that provide context for each row
→Batch Size: The planner will optimize the number of parallel operations based on your data
→Output Preferences: Specify in your prompt how you want the results formatted

Tips for Best Results

→Be specific in your prompts: Clearly describe what data you want and from where
→Provide examples: If you have specific formatting needs, include an example in your prompt
→Start with a sample: Test on a few rows first before processing large datasets
→Use descriptive headers: Clear column headers help the planner understand your data
→Leverage existing data: Include relevant context columns that can guide the enrichment process

Example Output

Here is an example Google Sheet showing various workflow outputs:

Next Steps

→Tool Use