Avatar

In today’s fast-paced digital landscape, finding precise and relevant information is crucial. At DevNet, we continuously innovate to enhance the user experience and streamline how our users interact with our platform. Today, in collaboration with the Meraki API team, we are excited to announce the launch of our new AI-driven Semantic Search and Summarization beta feature for Meraki API documentation. This powerful new tool leverages advanced Artificial Intelligence (AI) and Large Language Models (LLMs) to revolutionize how users search and retrieve information, making it more accurate, efficient, and intuitive.

This launch brings substantial improvements and benefits, showcasing the capabilities of the platform built by the DevNet group. The Meraki API team conducted an extensive user survey, highlighting several challenges and enhancements requested by users for API searches. These inputs have been integrated into the new UX improvements and semantic search feature development. Users will now enjoy support for natural language queries, enabling more intuitive and conversational interactions. Our search engine matches results based on the meaning and context of the queries, rather than just keywords, ensuring more relevant and precise results. Additionally, the feature includes summaries of search results, providing concise and actionable information at a glance, and an improved user experience with enhanced navigation and access to related resources.

In this blog post, we’ll explore the features, benefits, and challenges that the new Semantic Search and Summarization feature addresses. We will delve into the technical details in follow-up blogs.

What is Semantic Search?

Semantic search is a groundbreaking technology that goes beyond traditional keyword searches. It focuses on the meaning and context of search queries instead of just matching words. Unlike conventional searches, which return results based only on specific terms, semantic search interprets the intent behind the queries to deliver more relevant and accurate results.

Our semantic search engine uses advanced Artificial Intelligence (AI) and Large Language Models (LLMs). These tools help it understand complex queries and provide precise, contextually relevant answers. It leverages Natural Language Processing (NLP) and vector search techniques. This means it can understand the nuances of human language and find information based on meaning. Even if the exact keywords are not present, the search engine can still identify and retrieve relevant content by understanding the underlying meaning. This results in a more intuitive and effective search experience.

Key Features of the New AI-Driven Semantic Search

Our new semantic search comes packed with several innovative features designed to enhance your search experience and address common challenges:

  • Contextual Understanding and Advanced Query Interpretation: Our AI-driven semantic search engine interprets the meaning behind your queries by leveraging advanced algorithms and large language models (LLMs). This enables the search engine to grasp the nuanced intent behind your searches, ensuring that the results are contextually relevant. By understanding complex user queries and delivering the most pertinent information, this feature addresses the challenge of irrelevant search results, providing you with more precise answers.
  • Support for Natural Language Queries: Users can now search using natural language, making interactions more intuitive and conversational. This feature allows you to ask questions in a more human-like manner, and the search engine will understand and process these queries effectively. When you search using natural language queries, our semantic search engine leverages contextual understanding and advanced query interpretation to match your queries against the indexed content. We index the content using similar techniques, ensuring that the search results are not just based on keywords but on the actual meaning and context of the queries. This is achieved through vector search, which finds results based on semantic similarity.
  • Enhanced User Experience: The search results now display more detailed information about APIs, making it easier to locate and understand where an API fits within the broader context. As you type, our search engine provides intelligent suggestions that include the API method, path, and a brief description, helping you find what you need faster. This detailed view allows you to quickly grasp the functionality and integration points of the API, streamlining your development workflow and reducing the need to navigate through multiple documents.

Generative AI Overview

Our new semantic search feature leverages advanced Generative AI techniques to enhance the user experience by providing AI-generated summaries of the top search results. We achieve this through the capabilities of Large Language Models (LLMs), which not only understand the context and intent behind user queries but also generate concise, relevant summaries that capture the essence of the information.

  • AI-Generated Summaries: The Generative AI capabilities of our LLMs allow the search engine to analyze and condense large amounts of text data into meaningful summaries. These summaries provide users with quick and actionable insights, reducing the need to sift through extensive documentation. By presenting the most important information upfront, users can make faster, more informed decisions.
  • Contextual Relevance: Our LLMs understand the nuanced intent behind queries and the context of the content being searched. This ensures that the generated summaries are not only accurate but also highly relevant to the user’s needs. Whether you are looking for a specific API method or a general overview, the AI-generated summaries will provide you with the most pertinent information.
  • Sample Code Extraction: When the search results include relevant sample code, our Generative AI capabilities extract and present this code as part of the AI-generated responses. This feature is particularly useful for developers who need quick access to code examples. By integrating sample code into the search results, we streamline the development process and enhance productivity.AI Generated Summarization

By leveraging these advanced Generative AI techniques, our semantic search engine not only improves the accuracy and relevance of search results but also provides valuable insights and tools that enhance the overall user experience. Please note that while the AI strives to provide accurate and helpful responses, the generated content may contain errors, incomplete information, or biases. Users should review and verify the information before using any code or recommendations in a real environment. The AI-generated content serves learning and informational purposes and should not be used in production environments without thorough review and testing.

Additional Related Contents

Beyond just document search results, we now showcase related items such as Code Exchange, Learning Labs and Sandbox in a tab view, giving you a comprehensive view of available resources. This integration helps in reducing information overload by providing a structured way to access diverse resources

Conclusion

The new AI-driven Semantic Search and Summarization features at Meraki significantly enhance your search experience by leveraging advanced AI and LLM technologies. These features provide more accurate and contextually relevant search results, support natural language queries, and offer intuitive navigation and detailed API summaries. Additionally, the integration of Generative AI techniques ensures that users receive concise, actionable insights and relevant sample code, streamlining the development process and enhancing productivity.

We invite you to try out these beta features at https://developer.cisco.com/meraki/api-v1/ and experience the improvements firsthand. Also check out video podcast as part of DevNet Decoded series.  Your feedback is invaluable to us as we continue to refine and enhance our offerings. By exploring and using these new tools, you will be at the forefront of this innovative technology, helping us shape the future of API documentation and search experiences.



Authors

Neelesh Pateriya

Principal Engineers

Cisco Developers Platform Engineering Group