We have so much information online that over the years we’ve relied on search engines. But now, by introducing large language models (LLMs) and multi-query, we’re at the start of a significant change in how we search and find information.

Understanding large language models (LLMs)

LLMs, like OpenAI’s GPT series, are advanced AI models trained on a lot of content and data. They can understand and generate human-like text, making them super valuable in search technologies.

Users can now move from keyword-based searches to inputting a question. The LLM will understand the context of the user’s query as well as their intent. This leads to a response that is summarized so the user doesn’t have to click on search results to try to find the answer to their question.

The multi-query concept

The multi-query concept involves using multiple queries (think of a query as a question or a statement) or variations of a query to get a broader range of results. It’s like when you’re in a real-life conversation and you’re asking several related questions instead of just one to get a comprehensive answer.

Multi-query expands the scope of search beyond single queries. It provides a more diverse set of results.

Example multi-query search: sustainable energy

Suppose you’re researching “sustainable energy.” A traditional single-query search might just use the term “sustainable energy” as the query. However, with the multi-query approach, it would expand your search to include a variety of related queries. Take a look:

  1. Primary Query: “Sustainable energy.”
  1. Related Queries:

Each of these queries goes into a different area of sustainable energy. By using this approach, search engines can provide a more holistic view of the subject, to cover various aspects of the user’s interest or inquiry.

Integrating LLMs with multi-query

Today: when we search, we’re presented back with a list of results, which we then must go through.

The future: combine the following and the search experience will dramatically change:

An LLM that understands natural language


History of what the user has searched for prior (or even which pages they were on in your website) to create the best variants of a question


Multi-query that can return a variety of results to the LLM to answer the question


Powerful search results that are accurate and comprehensive to what the user is searching for (including a summarized response)