AI Sources: Where LLMs Get Their Information and Why It Matters for Brands

When we use AI systems like ChatGPT or other AI assistants, we often focus on the response we receive.

But behind every answer generated by a language model, there's a crucial element: the sources of information, often referred to as sources.

Understanding where this information comes from is increasingly important for businesses and brands, as it helps determine how an AI describes, compares, and recommends a company to users.

How LLMs Build Responses

Large Language Models (LLMs) do not operate like traditional search engines that display a list of links.

When a person asks a question, the model generates an answer by combining information from various types of sources.

The main sources include:

• public content available on the web

• editorial articles

• technical documentation

• corporate websites

• databases and knowledge graphs

• structured content available online

These sources contribute to forming the informational context that the model uses to build its responses.

According to OpenAI, language models are trained using a combination of publicly available data, licensed data, and content created by human trainers.
Source:
https://openai.com/policies/how-chatgpt-and-our-language-models-are-trained/

The Role of Sources in AI Responses

When a user asks, for example:

• “What are the best tools to analyze brand visibility in AIs?”

• “What platforms exist to monitor companies' online presence?”

• “Which tools help understand how a brand appears in AIs?”

the model generates a response based on the information it finds from the available sources.

Thus, the sources influence:

• which companies are mentioned

• how they are described

• which features are highlighted

• which competitors are mentioned.

For this reason, online sources help build the narrative that AI presents to users about a specific market or sector.

Why It's Useful for Brands to Know the Sources

Many companies monitor their online presence through metrics such as:

• website traffic

• search engine rankings

• media mentions.

With the rise of conversational AIs, a new dimension emerges: the sources used by AIs to generate responses.

Understanding which sources are used allows brands to:

• better understand how they are represented by AIs

• identify the sources that influence the industry's narrative

• observe where competitors appear

• spot opportunities to improve their informational presence online.

How GeoSnap Analyzes AI Sources

To help companies understand this new scenario, GeoSnap includes a feature dedicated to the analysis of the sources used by AIs.

The platform simulates real questions users might ask AI systems and analyzes the generated responses.

During this analysis, GeoSnap also identifies the sources of the information used in the responses.

This allows observation of elements such as:

• the most used websites by AIs

• the sources that most influence the responses

• the content that helps to build the industry's narrative.

A New Level of Online Visibility Analysis

Traditionally, companies analyze their visibility through search engines.

Source analysis adds a new perspective: understanding which content and sources feed AI responses.

This type of observation allows brands to better understand:

• how AIs interpret the market

• which information influences the responses

• which sources help build the perception of a sector.