What is Structured Data?

Structured data is information organized in a specific way so that it’s easy to search and process. It’s like the data you find in tables in a database or arranged in an Excel spreadsheet. This kind of organization makes it simple for digital tools to quickly find and use the data, which is especially helpful for tasks such as managing product inventories in retail, navigating course catalogs at universities, or optimizing supply chains in manufacturing. Content Management Systems (CMS) also store data in a structured format, enabling querying and filtering of content like articles, blog posts, and multimedia assets. Digital assistants can leverage this structured approach to accurately answer queries and perform actions based on precise data points.

Optimizing Structured Data with AI

Our platform enhances traditional RAG models with agents that specialize in structured data. By allowing the LLM reasoning engine to pass parameters to an agent specifically designed to handle structured data, the system can quickly filter and retrieve relevant information based on user-defined criteria. Whether the data resides in databases, PIM systems, or CMS platforms, the agents can interact with these backend systems to fetch and filter data. A copilot that can handle structured data is able to streamline data retrieval and boost the relevance of results.

Structured Data in Action

Let’s take a look at an e-commerce example. Imagine a company that sells medical devices, including numerous blood pressure monitors. With such a vast product catalog, finding the right device can be overwhelming for customers. 

Our platform employs agents to interact with backend systems, like a Product Information Management (PIM) system or a CMS, to query and filter data before presenting it to the user. When a customer interacts with a chatbot and seeks blood pressure monitors that fit specific criteria—such as price range, features, or brand—the agent quickly filters the structured data to retrieve relevant products.

The chatbot then presents these filtered results in a carousel, allowing users to select devices they’re interested in by checking boxes next to each feature. There’s also a text box where users can explain what’s important to them, such as portability, accuracy, or additional health metrics.

The selected products and the user’s input are then analyzed by the LLM. This analysis enables the assistant to provide personalized recommendations or detailed comparisons, highlighting how each device meets the user’s specific needs. By combining structured data handling with advanced AI reasoning, our platform delivers a personalized shopping experience.