With the release of Amazon Q Apps, customers may now create their own generative AI applications.
AWS also gave a sneak peek at Amazon Q Apps when they introduced Amazon Q Business in April 2024. Using the data from the company, customers can develop generative artificial intelligence (generative AI)-powered apps using Amazon Q Apps, a feature of Amazon Q Business. Natural language can be used by users to create apps, which can then be safely published to the organization’s app library for public use.
Amazon Q App
Amazon Q Apps are now generally available. During the preview, they gathered your comments and ideas. Additionally, they’re introducing several additional features that weren’t present in the preview, such an API for Amazon Q Apps and the option to define data sources down to the individual card level
Let’s first examine how to use Amazon Q Apps before you go into further detail about the new capabilities.
Convert discussions into reusable applications
Users can create an app from their interactions with Amazon Q Business using Amazon Q Apps. With the help of Amazon Q Apps, users may create custom apps by capturing the context of conversations in an intelligent manner. Watch it in action, please!
While they was writing this essay, AWS considered asking Amazon Q Business for assistance in creating a product summary for Amazon Q Apps. Ultimately, the purpose of Amazon Q Business is to increase worker productivity. then used the Amazon S3 adapter for Amazon Q Business to add the product messaging documents as a data source after uploading them to an Amazon Simple Storage Service (Amazon S3) bucket.
AWS has realised after started talking that others in the company could benefit from seeing a product overview that is based on a product description. To make a sharing and reusable app, you go with Create Amazon Q App.
You receive an automated prompt from Amazon Q Business to create an Amazon Q App, which you may review and modify as needed:
You decide to proceed with the app’s creation by selecting Generate. With four cards, it generates an app called Product Overview Generator. Two output cards that show the product overview and salient features, and two input cards that collect user input.
In addition, you can inspect and alter the prompts for each individual text output card because they are generated automatically. You select the Product Overview card’s edit icon to get the side panel prompt.
You can also choose the approved data sources or large language model (LLM) knowledge as the source for the text output card in the side panel. You can choose one or more of the authorised data sources that are set up specifically for this Amazon Q Business application. You choose the data source that you had set up for this app’s creation, Marketing (Amazon S3).
You’ll see that you was able to create a fully working app right from the discussion, without needing to modify any of the base prompt or the individual text output card questions.
By selecting Publish, you can now add this app to the organization’s app library. Let us examine an alternative approach to developing Amazon Q apps prior to releasing the app.
Make natural language apps with generative AI
You may choose Apps and describe the app you want to create using my own words, rather than starting with a chat in Amazon Q Business. Alternatively, you may experiment with the suggestions from a pre-made sample.
Distribute apps among your group
After you are satisfied with the layouts and prompts, you can publish the app to a common app library so that anyone using this Amazon Q Business application environment may access it.
The strong security and governance measures that Amazon Q Business has in place are carried over to Amazon Q Apps, guaranteeing the preservation of data sources, user permissions, and boundaries. As a result, only the data that is accessible to them in the underlying data sources is displayed to other users when they launch the application.
You can select Publish for the Product Overview Generator app that you made. It offers the ability to choose up to three labels and shows the app’s preview. Labels aid in the classification of the apps according to organisational departments or any other categories. You can select the labels and then click Publish once more on the preview popup.
The application will be made instantly available for use, duplication, and development on top of by others in the Amazon Q Apps library. To search the Amazon Q Apps Library and locate my Product Overview Generator app, you can select Library.
Adapt the apps in the app library to your unique requirements
With Amazon Q Apps, users can swiftly expand their productivity as individuals or as a team by customising and modifying shared apps to meet their unique requirements. Rather of beginning from zero, users can examine pre-existing programmes, utilise them in their current form, or make changes and add their own versions to the app store.
There’s an app called Document Editing Assistant that checks documents for grammar errors. The programme should be updated to include a document summary as well, in your opinion. Let’s investigate your options.
The original App Prompt that was used to create this app is visible to you. You can take a copy of the prompt and use it as a model to make an app that is identical. you only need to update it with a description of the capabilities you want to add, then let Amazon Q Apps Creator handle the rest. Or you could keep changing this version of the application.
It is possible to amend or remove already-existing cards. By selecting the edit icon on the card, for instance, you can change the prompt on the Edited Document text output card.
You can add more cards with additional functions, including a file upload, text output, user input, or an administrator-configured plugin. To further hone or improve the responses to your inquiries, you can utilise the file upload card, for instance, to supply a file as an additional data source. One way to use the plugin card is to open a Jira ticket for any action item that has to be completed as a follow-up.
What have they added since the sneak peek?
As you previously stated, they have added new features based on your comments and recommendations from the preview time.
List the data sources for each card. You can designate the data sources from which you would like the output to be created, as you have demonstrated while developing the programme. This feature was included in an effort to increase response accuracy.
You can configure more than one data source in your Amazon Q business instance.
Nevertheless, depending on the use case, you could just require a portion of these data sources to construct an app. For each of your app’s text output cards, you can now select different data sources. Alternatively, you can set the text output cards to employ LLM knowledge rather than any data sources, if your use case calls for it.
Q Apps API for Amazon Now that APIs for managing apps, app libraries, and app sessions are available, you may create and maintain Amazon Q Apps programmatically. This enables you to include all of Amazon Q Apps’ features into the tools and programmes of your choosing.
