--- title: Improve Enterprise RAG Accuracy in Data 360 | daily.dev description: Salesforce Data 360 can significantly improve RAG pipeline accuracy through three configuration changes rather than model training or custom code: intelligent... canonical: https://daily.dev/posts/improve-enterprise-rag-accuracy-in-data-360-ubbbcw08d twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Improve Enterprise RAG Accuracy in Data 360 | daily.dev og:description: Salesforce Data 360 can significantly improve RAG pipeline accuracy through three configuration changes rather than model training or custom code: intelligent... og:url: https://daily.dev/posts/improve-enterprise-rag-accuracy-in-data-360-ubbbcw08d og:image: https://api.daily.dev/og/posts/uBbbcw08D.png og:image:alt: Improve Enterprise RAG Accuracy in Data 360 og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Improve Enterprise RAG Accuracy in Data 360 **[Salesforce Developers Blog](https://daily.dev/sources/salesforcedevs)** · 8 min read · 1 upvotes · 0 comments ## Summary Salesforce Data 360 can significantly improve RAG pipeline accuracy through three configuration changes rather than model training or custom code: intelligent parsing for rich documents (charts, tables, diagrams), enriched indexing that generates metadata and anticipated question chunks, and tuning chunk size, retrieval depth, and generation prompts. Combined, these changes raised knowledge-base QA accuracy from 62.5% to 92.5% and complex enterprise PDF accuracy to 86.8% on public benchmarks. An upcoming Knowledge Graph feature (GA November 2026) aims to handle multi-document reasoning, and auto-tuning is expected to launch the same month to automate configuration selection. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How much can tuning chunk size and retrieval depth improve RAG accuracy without changing the model? Configuration tuning alone can improve accuracy by roughly 30 percentage points on knowledge-base question answering, with no code changes or custom models. Increasing chunk size from 512 to 2048 tokens, raising retrieval depth from 10 to 30 chunks, and switching from a generic to a domain-tuned generation prompt raised accuracy from 62.5% to 92.5% on the WixQA benchmark using the same platform and data. _Developers optimizing retrieval pipelines can track RAG configuration techniques like these on daily.dev._ ### What causes RAG pipelines to return wrong answers even when the correct document exists? Errors compound across four stages: content processing (charts and tables losing structure during extraction), content splitting (procedures cut at arbitrary token boundaries), search (vocabulary mismatch between queries and documents), and generation (the LLM misusing correct retrieved context). In one benchmark analysis, remaining errors after optimization broke down as 40% generation, 26% content processing, 18% search, and 16% content splitting. _Teams debugging RAG accuracy issues can follow pipeline optimization approaches on daily.dev._ ### What is enriched indexing and how much does it improve RAG retrieval accuracy? Enriched indexing automatically generates rich metadata (titles, keywords, entities, summaries) and question chunks (anticipated natural-language questions) for every indexed chunk, letting queries match against raw content, metadata, and anticipated questions simultaneously. Measured impact was an additional 2.4 percentage points on top of intelligent parsing alone, raising end-to-end accuracy from 84.4% to 86.8% on the OHR-Bench v2 benchmark. _Engineers evaluating enterprise search enrichment strategies can follow retrieval accuracy techniques on daily.dev._ ## Similar posts on daily.dev - [Introductory Walk-through of Progress Agentic RAG Dashboard](https://daily.dev/posts/introductory-walk-through-of-progress-agentic-rag-dashboard-tjbxf5fx3) · Telerik · 1 upvotes · 0 comments --- Tags: [#llm](https://daily.dev/tags/llm), [#rag](https://daily.dev/tags/rag), [#vector-search](https://daily.dev/tags/vector-search), [#salesforce](https://daily.dev/tags/salesforce) [View this post on daily.dev](https://daily.dev/posts/improve-enterprise-rag-accuracy-in-data-360-ubbbcw08d) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Improve Enterprise RAG Accuracy in Data 360","url":"https://daily.dev/posts/improve-enterprise-rag-accuracy-in-data-360-ubbbcw08d","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/improve-enterprise-rag-accuracy-in-data-360-ubbbcw08d"},"datePublished":"2026-09-02T17:02:10.246Z","dateModified":"2026-09-02T19:21:02.591Z","description":"Salesforce Data 360 can significantly improve RAG pipeline accuracy through three configuration changes rather than model training or custom code: intelligent...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/f4faf8479f2aa2a4846869db2c634aa1?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/f4faf8479f2aa2a4846869db2c634aa1?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Salesforce Developers Blog","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Salesforce Developers Blog","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/662fb7dc7cd14ee59747ed58d75cf7c3","url":"https://daily.dev/sources/salesforcedevs"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/improve-enterprise-rag-accuracy-in-data-360-ubbbcw08d","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"llm,rag,vector-search,salesforce","timeRequired":"PT8M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Salesforce Developers Blog","item":"https://daily.dev/sources/salesforcedevs"},{"@type":"ListItem","position":3,"name":"Improve Enterprise RAG Accuracy in Data 360"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/improve-enterprise-rag-accuracy-in-data-360-ubbbcw08d#faq","mainEntity":[{"@type":"Question","name":"How much can tuning chunk size and retrieval depth improve RAG accuracy without changing the model?","acceptedAnswer":{"@type":"Answer","text":"Configuration tuning alone can improve accuracy by roughly 30 percentage points on knowledge-base question answering, with no code changes or custom models. Increasing chunk size from 512 to 2048 tokens, raising retrieval depth from 10 to 30 chunks, and switching from a generic to a domain-tuned generation prompt raised accuracy from 62.5% to 92.5% on the WixQA benchmark using the same platform and data. Developers optimizing retrieval pipelines can track RAG configuration techniques like these on daily.dev."}},{"@type":"Question","name":"What causes RAG pipelines to return wrong answers even when the correct document exists?","acceptedAnswer":{"@type":"Answer","text":"Errors compound across four stages: content processing (charts and tables losing structure during extraction), content splitting (procedures cut at arbitrary token boundaries), search (vocabulary mismatch between queries and documents), and generation (the LLM misusing correct retrieved context). In one benchmark analysis, remaining errors after optimization broke down as 40% generation, 26% content processing, 18% search, and 16% content splitting. Teams debugging RAG accuracy issues can follow pipeline optimization approaches on daily.dev."}},{"@type":"Question","name":"What is enriched indexing and how much does it improve RAG retrieval accuracy?","acceptedAnswer":{"@type":"Answer","text":"Enriched indexing automatically generates rich metadata (titles, keywords, entities, summaries) and question chunks (anticipated natural-language questions) for every indexed chunk, letting queries match against raw content, metadata, and anticipated questions simultaneously. Measured impact was an additional 2.4 percentage points on top of intelligent parsing alone, raising end-to-end accuracy from 84.4% to 86.8% on the OHR-Bench v2 benchmark. Engineers evaluating enterprise search enrichment strategies can follow retrieval accuracy techniques on daily.dev."}}]} ```