Retrieval and Answer Configuration Guide

This guide covers the configuration of retrieval strategies, answer generation, and search results in Search AI. These settings determine how content is retrieved from your index and how responses are delivered to users.

Navigation: Configuration menu in Search AI

Retrieval Strategies

Configure the chunk retrieval strategy and corresponding thresholds for finding relevant content.

Retrieval Methods

Strategy	Description	Best For
Vector Retrieval	Uses cosine similarity between query vector and chunk vectors. Scores range from 0 (no match) to 1 (complete match)	Semantic similarity matching, contextual queries
Hybrid Retrieval	Combines keyword-based matching with vector-based scoring, leveraging strengths of both approaches	Balanced precision and recall, improved accuracy

Qualification Criteria

Parameter	Description	Range	Default
Similarity Score Threshold	Minimum similarity score for a chunk to be considered relevant. Higher values require more similarity	0-100	20
Proximity Threshold	How closely retrieved chunks should be to the highest-ranking chunk. Lower values mean closer chunks	0-50	20
Top K Chunks	Maximum number of qualified chunks for answer generation	1-100	20
Token Budget for Chunks	Maximum tokens to send to LLM. Must account for prompt, query, context, and response within LLM’s context window	1-1,000,000	20,000

Default Configuration Summary:

• Retrieval Mechanism: Hybrid Retrieval
• Similarity Score: 20
• Proximity Threshold: 20
• Top K Chunks: 20
• Token budget for chunks: 20,000

Answer Generation

Configure how responses are composed and delivered to users.

Answer Components

Component	Description
Answer Text	The generated response addressing the user’s question
Snippet Reference	Link to source as citation for further reading

Answer Types

Type	Description	Configuration
Extractive	Top chunk retrieved is directly presented as-is without text changes	Configure Response Length (tokens)
Generative	Top chunks are sent to configured LLM, which generates a paraphrased answer	Requires LLM integration and enabled Answer Generation in GenAI Tools

Generative Answer Configuration

Chunk Settings:

Setting	Description	Default	Max
Token Budget for Chunks	Total tokens for chunks sent to LLM	20,000	1,000,000
Enable Document Level Processing	Send full documents instead of just chunks for richer context	Disabled	-
Token Budget for Documents	Maximum tokens when sending full documents	50,000	100,000

Chunk Order Options:

Order	Description	Use Case
Most to Least Relevant	Highest relevance first, then decreasing	Standard prioritization
Least to Most Relevant	Lowest relevance first, most relevant at end followed by query	When recency in context matters

LLM Configuration:

Setting	Description
Select Generative Model	Choose from configured LLM models
Answer Prompt	Select prompt template for answer generation
Temperature	Controls randomness (lower = more deterministic, higher = more creative)
Response Length	Expected answer length in tokens

Feedback Configuration: Enable feedback mechanism to allow users to rate answers. Feedback data appears in Answer Insights analytics.

Response Streaming

Enable real-time token-by-token response delivery for Web/Mobile SDK channels, reducing perceived latency for longer answers.

Note: Streaming is configured via prompt settings and is currently supported only for OpenAI and Azure OpenAI models. Not available for API-based responses.

Search Results

Search results display a ranked list of documents/chunks by relevance, providing broader information compared to direct answers.

When to Use Search Results vs Answers

Use Case	Recommended
Direct, specific questions	Answers
Broad topic exploration	Search Results
Complex queries requiring comparisons	Search Results
Debugging/troubleshooting with multiple sources	Search Results

Configuration Settings

Setting	Description	Range	Default
Number of Search Results	Maximum chunks displayed	1-100	20

Note: When both search results and extractive answers are enabled, the top result is omitted from search results (shown as the answer) to avoid redundancy.

Filters (Facets)

Filters enable users to narrow results based on specific criteria, useful for large result sets.

Filter Types:

Type	Description
Static	Fixed, predefined filter values
Dynamic	Values derived from search results

Filter UI Options:

UI Type	Availability	Selection
Tabs	Static filters only	Single value, string fields only
Single Select	Dynamic filters	One value at a time
Multi Select	Dynamic filters	Multiple values concurrently

Creating Filters

1. Provide unique Filter Name
2. Select Filter Type (Static or Dynamic)
3. Choose Field for filtering
4. Select Filter UI style

Filter Rules:

• Only one tab-style filter can be enabled at a time
• Two filters cannot use the same field concurrently
• Filters apply only if search results contain the specified field

Default Filters

Every application includes default filters that can be updated, deleted, or disabled as needed.

Search Results Access

Currently available via Search API only.

Configuration Checklist

Step	Setting	Location
1	Choose retrieval strategy	Configuration > Retrieval Strategies
2	Set qualification criteria thresholds	Configuration > Retrieval Strategies
3	Select answer type (Extractive/Generative)	Configuration > Answer Generation
4	Configure LLM settings (if Generative)	Configuration > Answer Generation
5	Enable/configure search results	Configuration > Search Results
6	Set up filters	Configuration > Search Results

Quick Reference: Default Values

Parameter	Default Value
Retrieval Strategy	Hybrid
Similarity Score Threshold	20
Proximity Threshold	20
Top K Chunks	20
Token Budget (Chunks)	20,000
Token Budget (Documents)	50,000
Search Results Count	20

---