Expert Kernels (Limited Availability)
Introduction
What are Expert Kernels?
For teams building AI applications, Expert Kernels is an LLM-ready knowledge delivery solution. Unlike other knowledge retrieval options on the market, our product provides exactly the knowledge you need to build a fast, knowledge-powered LLM solution.
Without Kernels, teams building LLM-powered applications have limited options to build knowledge into the ecosystem. With Kernels, Expert provides semantically relevant pieces of pages via API, enabling your team to target the parts of your knowledge base that are highly relevant and useful to the query. Your application can use Kernels as the only knowledge source or in combination with other data to construct prompts for an LLM. Ultimately, Kernels enables your generative AI applications to leverage the knowledge already available on your site.
How does Expert Kernels work?
Given a natural language query, semantically relevant content from an Expert site can be retrieved via an API request. The content retrieved respects the available content to the user making the request, meaning that kernels respects permissions (conditional content is supported only by user type, not by group). In addition to verbatim pieces of pages from Expert sites, kernels contain metadata like the page ID, title, etc. All this helps your team create prompts for LLMs that contain the relevant information to build an exceptional solution.
What are some benefits of Expert Kernels?
- Faster integration: Use Kernels instead of scraping or parsing a full Expert page in order to build an integration more quickly.
- Permissions and access restrictions within your LLM application: Anywhere you use Kernels, trust that the information is permissioned for the authenticated user.
- Improved relevancy: By returning only the most relevant kernels instead of full pages, users are more likely to get pertinent answers to their queries.
- Faster results: Kernels provide answers quickly by eliminating the need to search through full page content.
- Better for AI uses: The ability to retrieve relevant text kernels makes it well-suited for seeding generative AI systems compared to page links.
- Always up-to-date: Kernels auto-update based on source content changes, so the information is always current.
- Context preservation: Kernels contain metadata like page title/URL so the source context of the information is preserved.
How do Kernels Work with Restricted Content?
By default, Kernels permissions work as you would expect: only the users with the access to see content will see kernels from that content. This includes complex restrictions like conditional content and group access.
For use cases where content on private and semi-private pages is extremely sensitive, it is possible to enable content from only public and semi-public pages, reach out to support to enable that setting.
What does Beta Release mean?
Until further notice:
- Structures of Kernel responses are subject to change with no warning.
- Structures of Kernel requests are subject to change with no warning.
- Content of Kernel responses for the same query are subject to change with no warning.
- There is no SLA on response time, though preliminary tests indicate that you can expect ~500ms.
Use Case: Pulling knowledge into an LLM application
Use kernels to build prompts out of Expert knowledge into your own LLM solution. An example of an application you could build might work like this:
Compare Kernels to other content retrieval methods
| Knowledge Retrieval API | Query With | Response Is... | Use When |
|---|---|---|---|
| Search Endpoint | User generated keywords | Links to Expert Pages | Users are looking for content on the site, location is unknown but content exists |
| Page Object Endpoint | Page ID | Full content of the page | Content location is known, all the content on the page is desired and you want to display the content elsewhere |
| Kernels Endpoint | User generated natural language question | Parts of pages from around an expert site | Building AI applications that can parse unstructured data, constructing LLM prompts that need factual, up to date information, or building chat bot solutions |
FAQs
What kind of content can be returned as kernels?
Text, tables and PDFs on and attached to content pages. Content permissions of the user making the Kernels request will be respected, so application builders and integration builders should be careful to accurately provision users.
How often are kernels updated?
Kernels are automatically updated based on page changes.
Is there a limit to how many kernels I can request?
The API supports requesting the specific number of kernels per request. The default is 10 and the maximum is 100. This is set with the limit query parameter.
How should kernels be used in my application?
Kernels is optimized for retrieving relevant text for seeding LLMs and building chatbots. Kernels are not formatted for end user display, but do contain verbatim content from Expert pages.
How do I access Expert Kernels?
Contact your CSM to get started with kernels.
How do I know if I am a good fit for Expert Kernels?
Kernels was built for technical teams building a knowledge-powered LLM solution. Is this you? Contact your CSM!