Overview
The Search Infobase action allows you to search through files and documents that you have uploaded and tagged in your Infobase. It will find the most relevant snippets or sections from those tagged files based on your search query. This is particularly useful when you have large documents or multiple files that you need to extract relevant information from.The Search Infobase action enables efficient retrieval of key information from your knowledge base, which can then be used in conjunction with other actions like Generate Text to synthesize complete answers or outputs.
Usage Examples
Support Documentation Q&A - support documentation can be added to Infobase and tagged. Then you can use the Search Infobase action to create a Workflow that easily searches through Infobase, extracts relevant snippets, and then using those retrieved snippets as a knowledge base to generate high-quality answers/outputs with a language model.
Sales Playbook Search - Upload your company’s sales playbooks and training materials to Infobase. Use the Search Infobase action to quickly find specific strategies, objection handling techniques, or product benefits based on a sales rep’s query. This can help in providing quick guidance during a sales call or meeting.
Marketing Content Curation - If you’ve stored a repository of past marketing campaigns, case studies, or content pieces in Infobase, the Search Infobase action can be used to extract relevant snippets for a new campaign. For example, search for “successful case studies in healthcare” to pull out key points that can be repurposed in a new marketing proposal.
Competitive Analysis - Store competitive analysis reports in Infobase and tag them accordingly. Use the Search Infobase action to retrieve specific insights about competitors, such as pricing strategies or market positioning. This allows your team to quickly access and leverage competitive information when planning sales pitches or marketing strategies.
Inputs
Search Query - This is the main query to search across the tagged Infobase.
Infobase Tag - An Infobase Tag is a repository or collection of files/documents that you have uploaded, tagged, and want to be able to search through.
Limit - These are the number of snippets to return.
Outputs
The primary output of the Search Infobase action is a set of the most relevant snippets from files or documents that have been tagged with specific keywords in the Infobase. The number of snippets returned can be configured, with the default being five snippets.
For each snippet, the following information is provided:
The name of the Infobase item (file or document) that the snippet was extracted from.
The actual text snippet that was deemed most relevant to the search query.
The snippets are extracted from larger documents or files in the infobase that have been tagged with keywords relevant to the search. The action breaks down these larger items into smaller snippets and then identifies and returns the most relevant snippets based on the search query.
The Search Infobase action is a key component of a retrieval-augmented generation (RAG) process, where relevant information is first retrieved from a large document store or Infobase, and then this retrieved information is used to generate a final answer or output.
Troubleshooting
No relevant snippets found - If the search Infobase action does not return any relevant snippets, it could be because there are no documents tagged with the specified tag in the Infobase, or the query is not matching any content in the tagged documents. Double check that you have uploaded documents and tagged them correctly in the Infobase.
Snippets are not relevant - If the returned snippets are not relevant to your query, it may be because the query itself is not specific enough or does not match the content well. Try rephrasing the query or adding more context. You can also adjust the number of snippets returned to see if more/fewer results are more relevant.
Related Actions
Generate Text - The Generate Text action is used to synthesize outputs by leveraging the snippets retrieved from the Infobase. After retrieving relevant information, you can use the Generate Text action to create summaries, answers, or further elaborations based on the snippets.
Extract Data From Text - This action allows you to extract specific data elements from the text within the snippets retrieved from the Infobase. It is useful for pulling out structured information like dates, names, or figures from the unstructured text.