GPT formula examples for GPT for Sheets

See GPT functions in action with real-world examples using every function in the GPT for Sheets library.

Prerequisites

You have opened a Google spreadsheet and selected Extensions > GPT for Sheets and Docs > Enable GPT functions.

You can use the example template to follow along.

Rewrite and create content

Use GPT functions to rewrite or create content adapted to your goals:

• Rephrase existing text by adjusting length, style, grammar or tone.
• Create content that matches your audience needs using the perfect level of creativity.

Rewrite or rephrase

Shorten product names:

Fix mistakes:

Change tone:

Learn more on GPT and GPT_EDIT.

Create content

Write taglines:

Play with temperature to adjust creativity:

Generate SEO metadata:

Generate keywords:

Learn more on GPT and GPT_LIST.

Translate

Leverage large language models for your machine translation tasks and benefit from:

• High performance: OpenAI models are ranged in the first tier of machine translation capability compared to commercial engines in several language pairs.
• Support for a wide range of language pairs: Even though the list of language pairs is not made public, ChatGPT is known to support over 80 languages.
• Integration of your corporate terminology: For on-brand, consistent translations.

Translate with GPT_TRANSLATE

Translate descriptions:

Translate with specific instructions

Translate with a dictionary:

1. Write your dictionary and your prompt.

   

2. Write your final instructions.

   

3. Translate your texts.

   

Your texts are translated in the target language, and the translations comply with the terminology you provided.

Learn more on GPT_TRANSLATE.

Data preparation

Free up your teams from tedious data preparation so they can concentrate on high-value data analysis:

• Basic data cleaning tasks include formatting data according to known standards or by following examples you set.
• Advanced tasks like named entity extraction, which traditionally demanded a specific pattern-based approach, are also an option.

Format data with GPT_FORMAT

Standardize dates:

Standardize cases:

Standardize phone numbers:

Standardize currencies:

Learn more on GPT_FORMAT.

Extract data with GPT_EXTRACT

Extract email addresses:

Extract phone numbers:

Extract countries:

Learn more on GPT_EXTRACT.

Cleanse data with GPT_FILL

Provide examples of how you want your data to be cleaned up, then apply the same logic to a range of cells. For best results, use GPT_FILL with the gpt-4o model and add global instructions to guide the AI.

Clean email addresses:

info

If you have access to the gpt-3.5-turbo-instruct model with your own API key, use it with GPT_FILL. This model is task-oriented and can be more accurate than the gpt-4o model, which is a chat model and requires instructions in natural language to provide similar results.

Learn more on GPT_FILL.

Analyze

Get instant insights from large volumes of content through automated summarizing, categorization, tagging, and sentiment analysis.

Summarize data with GPT_SUMMARIZE

Summarize customer tickets, even from different languages:

Summarize long content:

Learn more on GPT_SUMMARIZE.

Categorize or classify data with GPT_CLASSIFY

Categorize customer feedback:

Analyze sentiment:

Categorize customer feedback with examples:

Learn more on GPT_CLASSIFY.

Assign tags with GPT_TAG

Tag products:

Learn more on GPT_TAG.

Analyze images with GPT_VISION

Applies a prompt to an image. GPT_VISION is the only function that can process images. For example you can ask a question about the image, ask for a description of it, or generate alt text for images. This function uses the gpt-4o model.

Create content from an image

Write product descriptions:

Extract information from an image

Extract amounts from invoices:

Extract info from an ID:

Explain elements in an image

Identify elements of an image:

Understand an image:

Explain formulas:

Generate alt text for images

Create descriptive alt text:

Learn more on GPT_VISION.