Using AI Summary on Knowledge Central
Overview
When searching for information on Knowledge Central, you can enable the AI Summary functionality to generate an answer based on information from multiple articles across the Knowledge Central site. This can help you find best practices or give you a quick how-to.
Enable AI Summary
By default, AI Summary is turned off. To turn it on, perform a search from the home page. On the search results screen, click the AI Summary toggle, shown below. The AI Summary generates a summary based on your search term. After the summary, you can click "Source Articles" to view articles related to the response.
Note: Source articles originate from a pre-filter built into the Knowledge Central platform. The backend reads your search query and matches it with "kernels," the platform's content algorithm, then sends the results to the AI Summary. Therefore, there is not necessarily a 1:1 relationship between the articles listed and the AI Summary itself.
Ask a Question
Because AI Summary has a Large Language Model backend, you can ask questions or create prompts to improve its responses.
Note: In the above screenshot, "Generating from" indicates "Users Guide >> Product Catalog." You can use search filters to drill down for a more specific or relevant response. If you were to filter the results through "Developers Guide," the response would focus on API calls.
Localization
Once the AI Summarizer generates a response, you can use Google Translate to view the response in the language of your choice.
You can also specify that language you want the response in without using Google Translate. Unfortunately, AI Summary does not understand all languages.
Important Notes
There are a few sections explicitly excluded from the AI Summary index, listed below:
- Aria6—Because this is an older version of Aria, indexing that documentation would cause the AI Summary to include erroneous information.
- Add-Ons—Because add-ons are not part of Aria Core, but perform various Aria Core functions, the AI Summary would confuse add-on and core functionality, diminishing the value of the responses.
- Release Notes—Because release notes might refer to prior iterations of functionality, the quality and accuracy of the AI Summary responses would be diminished.
Note: This LLM index exclusion does not affect the non-AI-Summary search results.