Skip to Main Content

AI Tools in Library Databases

A description of the AI tools available in the library's databases, and their capabilities.

ai

AI Tools from Statista

Research AI Tool and Chat is available in: Statista


Research AI

"Research AI" interprets a natural language query and generates a research summary based on the statistical content of the database.

  • Click on the "Research AI" link in the top menu to take you to an informational page about Research AI.

statistic Home Screen


On the Research AI informational page, click on the three highlighted buttons to see examples of natural language queries that could be entered.

Scroll down for much more information on the capabilities of Research AI.

  • To begin using the tool, click on the "Go to Research AI" button.

Research AI information screen


On the Research AI "chat" search page, enter a natural language query into the search box.

  • Example:  "Do people still read books? Summarize what book readership looks like, where it has been, and where it is going. Who reads print books? Who reads ebooks? Who listens to audiobooks?

research AI home page


On the Research AI results page, a brief report is generated summarizing the information from several sources.

  • The "TL;DR" section provides a brief summary.
  • Statistical details are displayed below, and the statistical sources are listed in the right sidebar.

Research AI results page


Hover over one of the numbered source references to bring up a box with details about the source.

research AI sources


"Next Steps," below the research summary provides search suggestions for additional research directions, and at the bottom of the page you can enter a new question.

research ai next steps page


More information on "Research AI"


Under the Hood

What powers Statista's AI tools?

Statista’s AI tools, including Research AI and Research AI Chat, use a blend of advanced Large Language Models (LLMs). Specifically, they leverage Claude 3 Sonnet (from Anthropic) for generating summaries and context, and Cohere’s LLMs for semantic search and query interpretation. The platform employs Retrieval-Augmented Generation (RAG), ensuring all AI outputs are grounded only in Statista’s own curated database, never pulling answers from the open web. This architecture allows for nuanced, transparent, and reliable statistical insights while minimizing hallucinations.

AI-assisted, human-verified text