Working with Worksheets - Decisional Documentation

What Are Worksheets?

Worksheets are your agent’s data layer - they’re full-featured spreadsheets where agents read inputs, write outputs, and store data.

Why Spreadsheets?

Spreadsheets are the perfect interface for automation. They’re flexible (any data structure), familiar (everyone knows Excel), and visual (see your data instantly).

Creating a Worksheet

Upload Existing
Create from Scratch
Import from URL

Start with your current spreadsheet

Click Upload

In the agent page, go to the Worksheet tab

Select File

Choose a file to upload:

CSV: Simple comma-separated values
XLSX: Excel workbooks (multiple sheets supported)
XLS: Legacy Excel format

Review

Your spreadsheet loads in the editor. All formatting, formulas, and data are preserved.

Files up to 50MB are supported. For larger datasets, consider splitting into multiple files or using database integrations.

The Worksheet Editor

Decisional uses SpreadJS, a professional-grade spreadsheet engine with full Excel compatibility.

Key Features

Formulas

All Excel formulas work: SUM, VLOOKUP, IF, INDEX/MATCH, and more

Formatting

Cell colors, fonts, borders, number formats, conditional formatting

Charts

Create visualizations: bar, line, pie, scatter charts

Multiple Sheets

Work with multiple sheets in one workbook

Filters & Sorting

Filter data, sort columns, group rows

Data Validation

Dropdowns, number ranges, date validation

Freeze Panes

Lock header rows/columns while scrolling

Auto-Save

Changes sync automatically every few seconds

Version History

Restore previous versions if needed

Keyboard Shortcuts

Navigation

Selection

Shift + Arrows: Extend selection
Ctrl/Cmd + A: Select all
Ctrl/Cmd + Space: Select entire column
Shift + Space: Select entire row

Editing

F2: Edit cell
Ctrl/Cmd + C: Copy
Ctrl/Cmd + V: Paste
Ctrl/Cmd + X: Cut
Ctrl/Cmd + Z: Undo
Delete: Clear cell

Formatting

Ctrl/Cmd + B: Bold
Ctrl/Cmd + I: Italic
Ctrl/Cmd + U: Underline
Ctrl/Cmd + 1: Format cells dialog

Structuring Data for Agents

How you structure your worksheet affects how well agents can work with it.

Best Practices

Use Clear Column Headers

One Data Type Per Column

Separate Inputs and Outputs

Sheet 1: “Requests” (inputs)
Sheet 2: “Generated Quotes” (outputs)

Why: Makes it clear what the agent should read vs. write.

Use Consistent Formats

Date format: Always use YYYY-MM-DD (e.g., 2025-01-15)Currency: Always include currency symbol or separate columnStatus values: Use controlled vocabulary (e.g., “Pending”, “Approved”, “Rejected” - not “pending”, “APPROVED”, “Reject”)Why: Consistency helps agents understand and generate correct data.

Add a Status Column

Only process rows where Status is "Pending".
After processing, set Status to "Processed" or "Error".

Include an ID Column

="Q-" & TEXT(ROW()-1, "000")

Why: Makes it easy to reference specific rows in logs and debugging.

Using Formulas

Agents understand and respect Excel formulas.

Common Formulas

Math & Aggregation
Lookups
Conditional Logic
Text & Dates

Sum:

=SUM(D2:D100)

Average:

=AVERAGE(E2:E100)

Count:

=COUNTA(A2:A100)

Multiply:

=C2 * D2

Percentage:

=E2 * 0.08  // 8% tax

VLOOKUP (find value in table):

=VLOOKUP(B2, PriceList!A:C, 3, FALSE)

INDEX/MATCH (more flexible):

=INDEX(PriceList!C:C, MATCH(B2, PriceList!A:A, 0))

XLOOKUP (modern Excel):

=XLOOKUP(B2, PriceList!A:A, PriceList!C:C)

IF:

=IF(E2 > 10000, "High Value", "Standard")

IF with multiple conditions:

=IF(E2 > 50000, "Tier 1", IF(E2 > 10000, "Tier 2", "Tier 3"))

IFS (cleaner for multiple conditions):

=IFS(E2 > 50000, "Tier 1", E2 > 10000, "Tier 2", TRUE, "Tier 3")

Concatenate:

=A2 & " " & B2
// or
=CONCAT(A2, " ", B2)

Text formatting:

=TEXT(C2, "MM/DD/YYYY")  // Format date
=TEXT(D2, "$#,##0.00")   // Format currency

Extract text:

=LEFT(A2, 5)      // First 5 characters
=RIGHT(A2, 3)     // Last 3 characters
=MID(A2, 3, 4)    // 4 characters starting at position 3

Today’s date:

=TODAY()

How Agents Use Formulas

Agents read formula results: If cell E2 has =C2*D2, the agent sees the calculated value, not the formula text.

Agents can write to formula cells: If you want the agent to override a formula, that’s fine. The formula gets replaced.

Use formulas for validation: Add a column like =IF(SUM(D2:D10) = E2, "Valid", "Error") to check agent output.

Working with Multiple Sheets

Organize complex workflows across multiple sheets in one workbook.

Common Patterns

Input/Output Separation
Reference Data
Error Handling
Multi-Stage Pipeline

Read customer requests from the "Requests" sheet.
Write generated quotes to the "Quotes" sheet.
Link each quote to its request via Request ID.

Process invoices from the "Invoices" sheet.
Look up vendor details from the "Vendors" sheet.
Classify by category using the "Categories" sheet.

Process rows from "Main Data".
If any error occurs, log it to the "Errors" sheet.
Continue processing remaining rows.

Data Validation

Constrain what agents (and humans) can enter into cells.

Setting Up Validation

Select Cells

Click the column header or select the range you want to validate

Open Data Validation

Click Data → Data Validation in the toolbar

Choose Validation Type

List: Dropdown with allowed values
Number: Min/max range
Date: Date range
Custom formula: Advanced rules

Set Error Message

Define what message shows if validation fails

Validation Examples

Dropdown List

Number Range

Use case: Quantity must be between 1 and 1000Settings:

Type: Number
Minimum: 1
Maximum: 1000

Result: Agent can’t enter 0 or 1001.

Date Range

Use case: Order date must be within last 90 daysSettings:

Type: Date
Minimum: =TODAY()-90
Maximum: =TODAY()

Result: Only recent dates allowed.

Custom Formula

Use case: Email must be valid formatSettings:

Type: Custom
Formula: =AND(ISNUMBER(FIND("@", A2)), ISNUMBER(FIND(".", A2)))

Result: Must contain @ and . to be accepted.

Conditional Formatting

Automatically format cells based on their values.

Use Cases

 Highlight Status
 Highlight High Values
 Expiration Warnings
 Duplicate Detection
Scenario: Color-code status valuesRule:If Status = “Completed” → Green background
If Status = “Error” → Red background
If Status = “Pending” → Yellow background
How to:Select status column
Format → Conditional Formatting
Add rule for each status value
Scenario: Flag high-value ordersRule:If Total > $10,000 → Bold red text
How to:Select total column
Add rule: Cell value > 10000
Format: Bold, red text
Scenario: Warn about upcoming deadlinesRule:If Due Date is within 7 days → Orange background
If Due Date is past → Red background
How to:Select date column
Rule 1: =A2 <= TODAY()+7 → Orange
Rule 2: =A2 < TODAY() → Red
Scenario: Highlight duplicate invoice numbersRule:If invoice number appears more than once → Yellow background
How to:Select invoice column
Add rule: =COUNTIF($A:$A, A2) > 1
Format: Yellow background

Charts and Visualizations

Create charts to visualize your data.

Creating a Chart

Select Data

Highlight the data range you want to chart (include headers)

Insert Chart

Click Insert → Chart in the toolbar

Choose Chart Type

Column/Bar: Compare values across categories
Line: Show trends over time
Pie: Show proportions of a whole
Scatter: Show correlation between two variables

Customize

Add title
Label axes
Change colors
Add legend

Chart Examples

Sales Over Time

Data: Date | RevenueChart type: Line chartUse: Track revenue trends by day/week/month

Product Mix

Data: Product | Quantity SoldChart type: Pie chartUse: See which products are most popular

Regional Comparison

Data: Region | Sales | ProfitChart type: Grouped column chartUse: Compare metrics across regions

Price vs. Volume

Data: Unit Price | Units SoldChart type: Scatter plotUse: Find correlation between price and demand

Importing and Exporting

Export Options

Download
Share Link
Sync to Integration

Click File → Download and choose format:

XLSX: Excel workbook (preserves formulas, charts, formatting)
CSV: Plain data (one sheet only, no formatting)
PDF: Static snapshot for sharing

Import from External Sources

API Import
Web Scraping
Email Attachments

Use integrations to pull live data:

Agent instruction:
Fetch customer list from Salesforce API
Write to "Customers" sheet
Refresh daily at 6 AM

Pull data from websites:

Agent instruction:
Scrape pricing from competitor website: example.com/pricing
Extract product names and prices
Update "Competitor Prices" sheet

Extract data from emailed spreadsheets:

Agent instruction:
Monitor Gmail for emails with subject "Weekly Report"
Download attached CSV
Import data into "Weekly Data" sheet

Advanced Features

Named Ranges

Create named references for frequently used ranges: Example:

Select cells B2:B100
Name it “CustomerNames”
Use in formulas: =VLOOKUP(A2, CustomerNames, 1, FALSE)

Pivot Tables

Summarize large datasets:

Select Data

Click anywhere in your data table

Insert Pivot Table

Click Data → Pivot Table

Configure

Rows: What to group by (e.g., Product)
Values: What to summarize (e.g., Sum of Quantity)
Columns: Secondary grouping (e.g., Month)

Freeze Panes

Lock header rows/columns while scrolling:

Select Cell

Click the cell where you want the freeze to start (e.g., B2 to freeze row 1 and column A)

Freeze

Click View → Freeze Panes

Performance Optimization

For Large Worksheets

Use Filters Instead of Multiple Sheets

Instead of: Separate sheets for each statusDo this: One sheet with a Status column, use filters to view subsetsWhy: Easier for agents to manage one dataset

Limit Formula Complexity

Avoid: Nested formulas with 10+ levelsDo this: Break complex calculations into multiple columnsWhy: Faster calculation, easier debugging

Archive Old Data

Instead of: One sheet with 100,000 rows going back 5 yearsDo this: Monthly or yearly archive sheetsExample: “Active Data” (current month) + “Archive 2024” + “Archive 2023”Why: Faster loads and agent processing

Use External Databases for Very Large Datasets

If you have: Millions of rowsDo this: Store data in PostgreSQL/Snowflake, use worksheet as a “view” or “cache”Agent pattern:

Query database for relevant rows
Write summary to worksheet
Process in worksheet
Write results back to database

Best Practices Summary

Keep it simple: Don’t over-engineer your data structure. Start simple and add complexity as needed.

Use clear names: Descriptive column headers, sheet names, and named ranges help everyone (including agents).

Validate inputs: Use data validation to prevent bad data from entering your worksheet.

Separate concerns: Use different sheets for inputs, outputs, reference data, and errors.

Document your structure: Add a “README” sheet explaining what each sheet/column does.

Version control: Download backups regularly or sync to external storage.

Next Steps

Working with Agents

Learn how agents read and write to worksheets

Working with Sources

Combine worksheets with document context

Advanced Data Processing

Advanced patterns for complex data workflows

API Reference

Programmatic access to worksheets

Getting Started

Integrations

User Guides

Reference

​What Are Worksheets?

Why Spreadsheets?

​Creating a Worksheet

​The Worksheet Editor

​Key Features

Formulas

Formatting

Charts

Multiple Sheets

Filters & Sorting

Data Validation

Freeze Panes

Auto-Save

Version History

​Keyboard Shortcuts

​Structuring Data for Agents

​Best Practices

​Using Formulas

​Common Formulas

​How Agents Use Formulas

​Working with Multiple Sheets

​Common Patterns

​Data Validation

​Setting Up Validation

​Validation Examples

​Conditional Formatting

​Use Cases

​Charts and Visualizations

​Creating a Chart

​Chart Examples

Sales Over Time

Product Mix

Regional Comparison

Price vs. Volume

​Importing and Exporting

​Export Options

​Import from External Sources

​Advanced Features

​Named Ranges

​Pivot Tables

​Freeze Panes

​Performance Optimization

​For Large Worksheets

​Best Practices Summary

​Next Steps

Working with Agents

Working with Sources

Advanced Data Processing

API Reference

What Are Worksheets?

Creating a Worksheet

The Worksheet Editor

Key Features

Keyboard Shortcuts

Structuring Data for Agents

Best Practices

Using Formulas

Common Formulas

How Agents Use Formulas

Working with Multiple Sheets

Common Patterns

Data Validation

Setting Up Validation

Validation Examples

Conditional Formatting

Use Cases

Charts and Visualizations

Creating a Chart

Chart Examples

Importing and Exporting

Export Options

Import from External Sources

Advanced Features

Named Ranges

Pivot Tables

Freeze Panes

Performance Optimization

For Large Worksheets

Best Practices Summary

Next Steps