←back to docs

Rerankers

Explore the capabilities of the reranking algorithm in ChatBotKit, designed to enhance the accuracy and relevance of AI responses by employing a sophisticated second-level sorting mechanism for retrieved records.

Rerankers improve the quality of search results from your datasets by applying an intelligent second-pass sorting step. After a dataset store retrieves an initial set of matching records, the reranker scores and reorders them so the most relevant results reach your AI first. This feature works with Retrieval-Augmented Generation (RAG) workflows.

How Reranking Works

  1. Initial retrieval - The dataset store searches your records and returns a candidate set based on the query.
  2. Reranking - The selected reranker model scores each candidate for relevance to the query and reorders them.
  3. Improved results - The AI receives the reordered list, so the best-matching content is prioritized in its response.

Available Reranker Models

ModelDescription
Cohere Rerank 4 FastSpeed-optimized tier of Cohere's Rerank 4 generation, tuned for lower per-query latency and higher throughput. This is the default reranker model.
Cohere Rerank 4 ProQuality tier of Cohere's Rerank 4 generation - Cohere's strongest reranker, aimed at enterprise search and RAG pipelines where ranking accuracy matters most.

The reranker considers up to 20 candidate records from the store on each search. How many records are ultimately returned is controlled by the dataset's own search settings, not the reranker model.

Setup Instructions

Reranking is configured per dataset.

  1. Navigate to your Dataset Settings in the ChatBotKit dashboard.
  2. Find the Reranker field.
  3. Pick a reranker model from the model picker.
  4. Optionally click the options icon to adjust the Region.
  5. Save your changes.

To disable reranking for a dataset, clear the reranker field.

Configuration Options

After selecting a reranker model, you can adjust the following via the options icon:

  • Region - The region where the model runs. The reranker models are currently available in the us region.

The number of candidates considered before reranking is a fixed platform setting, and the number of results returned is governed by the dataset's search settings.

Best Practices

  • Start with Cohere Rerank 4 Fast - It is the default and covers most use cases with fast response times.
  • Try Cohere Rerank 4 Pro for enterprise or high-precision needs - It is designed for applications where accuracy outweighs latency.
  • Test with representative queries - Enable reranking, run typical user queries, and compare result quality before and after.
storagererankersearchRAG