Skip to main content
All CollectionsAdvanced
OpenAI Request Multi-Field Responses
OpenAI Request Multi-Field Responses

Create more detailed and concise AI generated content with each of your prompts.

Justin Yan avatar
Written by Justin Yan
Updated over a week ago

This article is a feature and an extension of the OpenAI Request custom action. To learn about the fundamentals of this action and its setup, click here.

Beyond the standard setup of using OpenAI responses into your content blocks, there is new way you can enrich the content you're generating by segmenting the responses with Objects.

AI object responses function similarly to paragraphs in writing, breaking up content for clarity and readability. They help organize information and improve user engagement by structuring responses into digestible chunks.


Examples

Compare these two responses where "roofing" is the input for the prompt below:


​Prompt:

Generate a profitable business idea in the {{ form.industry }} industry.


​Text Response:
​


​Object Responses:


Where the text response is more fluid and conversational, the object response is more concise and gives more pertinent details relating to their respective fields. Each response type has its advantages, however, in the context of this business idea app generator, maintaining content that is focused and organized is definitely more important! Read on to learn about the setup of these objects.

Try it out! Add our Business Idea Generator to your app suite by downloading the app template today!


Setup

To setup your response objects, in your OpenAI Request modal, set the Response Type to Object.

This selection will open a set Key / Description pairs. For each key, set a description that relates to the main prompt in some way. The "problem" key shown above for example has the description, "The problem this business solves".

Like other Frontly functions that leverages the key system, prescribing a key with each description will enable you to access its generated content in your later action steps. Therefore, instead of entering your variable as {{actionSteps.Step.response}} (which is what you would for a text response), you would instead enter {{actionSteps.Step.CustomKey}}, as shown below:

Feel free to revisit the original OpenAI Request article (linked above) to learn how to display these outputs through a Google Sheet action or Update Local State one. Enhance your AI-generated content now using this advanced generation method!

Did this answer your question?