How to use Memory

1. Introduction

In this tutorial, we will explore the Memory section of Yala in detail. Yala is a powerful extension that integrates AI into Google Sheets to automate and streamline workflows. The Memory section allows you to provide additional context during processing, making the system’s responses and actions more relevant. Understanding how to use this functionality is crucial to making the most of Yala’s capabilities.

Remember that in your sheet (for example, the Nutritionist tab), you might have columns such as Task, Category, Product, Composition, and Status—alongside Memory Type, Memory Prompt, and Memory Output.

2. How Memory Works

The Memory section is designed to supply contextual information that Yala can use while executing prompts. This is especially helpful when you need the system to consider specific or historical data (e.g., from columns like Category, Product, Composition) while generating responses or performing actions.

Operational Flow

Selecting the Memory Type

You pick the memory scope you want to use, such as Spreadsheet, Sheet, Row, or Custom.

Providing the Memory Prompt

You enter any instructions or additional details that will be combined with the chosen memory scope (for instance, “Obtain the Category, Product, and its Composition information.”).

Processing

Yala takes the Memory Type and Memory Prompt to generate the Memory Output, which can then be used in later stages of the workflow.

Interactions with Other Components

Agent

The Memory Output can be utilized by the AI when it processes the User Prompt, adding extra context to produce more accurate answers (such as how to classify or generate nutritional information for the Product).

Parsing and Action

Details from Memory Output can influence how data is analyzed or which tasks (Action Prompt) are performed (for example, moving to another row or inserting data into another tab).

3. Required Parameters and Inputs

Input Data

Memory Type

Determines the memory scope to be used (e.g., Row to consider the specific row that has columns like Task, Category, Product, Composition).

Memory Prompt

Additional instructions or information that complement the chosen memory type (e.g., “Combine the data from this row to form a quick summary.”).

Prerequisites

Relevant Data

Ensure that any data you want to use is correctly filled in the appropriate columns or tabs (for instance, Category, Product, Composition in the Nutritionist tab).

Proper Configuration

Select a Memory Type that fits the context of your workflow (if you only need data from one row, choose Row; if multiple tabs, choose Spreadsheet or Sheet).

4. Detailed Look at the Memory Columns

Memory Type

Description

Defines the scope of memory that Yala should use in your workflow.

Options

Spreadsheet: Considers data from the entire spreadsheet.

Sheet: Considers data only from the current tab.

Row: Considers data from the current row (for example, the row containing Category, Product, Composition).

Custom: Lets you specify your own custom data or instructions.

How to Use

Select one of these options from the dropdown menu based on the scope you need. For instance, if you want Yala to focus solely on the current row, choose Row.

Memory Prompt

Description

A field where you add instructions or extra details to be used together with the chosen Memory Type.

How to Use

Write the text that should serve as memory context. For example: “Obtain the Product details from this row and summarize them.”

Memory Output

Description

Shows the result generated from combining Memory Type and Memory Prompt.

How to Use

This column is filled automatically once Yala processes the memory. Check it to see how the system interpreted the instructions you provided. You can then use the generated Memory Output in Agent, Parsing, Data, or Action prompts.

5. Practical Examples

Example 1: Using Memory from the Current Row

Scenario

You want Yala to create a personalized message based on data from columns Task, Category, Product, and Composition in the current row.

Steps

Set Memory Type to Row.

In Memory Prompt, enter: “Use the data from the current row to create a message about the Product.”

Set Status to Pending to begin processing.

Result

The Memory Output will contain the combined information from Category, Product, Composition, etc., ready to be used by an Agent or in a Parsing operation.

Example 2: Querying Data from Another Tab

Scenario

You need Yala to gather nutritional summaries from a tab called Nutritional Table.

Steps

Set Memory Type to Sheet.

In Memory Prompt, enter: “Provide a list of all items in the Nutritional Table tab that are labeled as cakes.”

Make sure the Nutritional Table tab exists and is up to date.

Set Status to Pending.

Result

The Memory Output will give you a filtered view of items identified as cakes from the Nutritional Table tab.

6. Best Practices

Choose the Right Memory Type

Pick the memory scope that best fits your context (e.g., Row for data in columns like Product and Composition, or Sheet if the entire tab is relevant).

Clarity in the Memory Prompt

Provide concise, direct instructions so Yala knows exactly what you want.

Check Your Data

Make sure that any referenced tabs or columns (Category, Product, Composition) are correct and up to date.

Use Memory Wisely

Avoid requesting large amounts of data you don’t actually need, as it may affect processing performance.

7. Troubleshooting

Empty or Unexpected Memory Output

Confirm that Memory Type is configured correctly.

Double-check your Memory Prompt for clarity and specificity.

Processing Error

If Status changes to Failure, review the Memory Output for potential error messages.

Verify that the tabs or data you’re referencing (like Nutritional Table) exist and are spelled correctly.

Outdated Data

Update information in the relevant tabs or columns.

Reprocess the row by setting Status to Pending again.

8. Relationship with Status

Pending

After choosing your Memory Type and Memory Prompt, set Status to Pending so Yala can process the memory.

Processing

Indicates that Yala is executing the memory instructions you set.

Success

The Memory Output was successfully generated and is ready for subsequent steps (like Agent, Parsing, or Data).

Failure

An error occurred. Check your inputs as described in the troubleshooting section.

9. References and Additional Resources

Complementary Guides

Check the Agent tutorial to learn how Memory Output can be combined with a User Prompt.

Review the Parsing tutorial to extract specific information from the Memory Output.

Support

If you continue to have questions or issues, contact the Yala support team for further assistance.

By understanding the Memory section, you can significantly enrich the context of your workflows, enabling Yala to provide more accurate and relevant responses and actions. Experiment with different combinations of Memory Type and Memory Prompt—for example, referencing columns like Category, Product, Composition—to see how they can meet your specific project needs. Take full advantage of this functionality to optimize your processes and get even better results!