Coveo Launches Passage Retrieval API

Coveo has rolled out Coveo Passage Retrieval API, a headless retrieval method for generative artificial intelligence applications that leverages the Coveo AI-Relevance Platform.

Coveo's Passage Retrieval API delivers precise, trustworthy text to large language models and AI applications, including Salesforce Agentforce, Microsoft Copilot, Amazon Bedrock Agents, and SAP Joule. Organizations can scale multiple generative AI applications using a single retrieval method.

With Coveo's information retrieval method and advanced relevancy and ranking techniques now available to power generative AI applications, organizations can ground results in enterprise knowledge. Coveo's platform centers on a unified index that combines advanced ranking techniques—lexical, vector-based, and behavioral—to securely connect data from systems like CRMs, enterprise resource planning, IT service management, and content management systems using prebuilt connectors. This eliminates redundant indexing and integration work while supporting a wide range of AI applications such as search, recommendations, generative AI, and content discovery.

With hybrid semantic and lexical indexing, machine learning models like Automatic Relevance Tuning (ART), and personalized ranking, as well as control over weighting factors and business rules in the query pipeline, Coveo continuously improves retrieval results.

Top use cases for Coveo's Passage Retrieval API include the following:

• Custom genAI experiences: Improve the reliability and security of any generative AI application by providing LLMs with the most relevant passages from unified enterprise content sources.
• Chatbots and virtual agents: Improve the relevance of the answers from chatbots, co-pilots, and agentic AI applications for personalized and context-rich conversations.
• Internal knowledge management to empower employees with personalized, accurate answers grounded in enterprise knowledge.
• Customer-facing support and case deflection, providing precise self-service answers cited from support knowledge.

"Enterprises today are challenged with turning vast amounts of fragmented data into accessible and relevant knowledge for generative and conversational AI," said Laurent Simoneau, co-founder, president, and chief technology officer of Coveo, in a statement. "While large language models (LLMs) hold immense potential, their success hinges entirely on the quality of the data they access—a growing concern as many enterprises develop their own genAI applications. Many organizations are adopting retrieval-augmented generation (RAG) and quickly realizing that the hardest part—the 'R'—lies in retrieving precise and relevant information from scattered enterprise systems. This critical step ensures secure, accurate, and relevant AI responses, yet its complexity often hinders progress."

Related Articles

Coveo Launches Relevance-Augmented Passage Retrieval API

Coveo's Relevance-Augmented Passage Retrieval API allows companies to integrate their own LLMs with Coveo's advanced retrieval technology.