Skip to main content
AI models search the web to add to their training knowledge when answering questions. Knowing what they search for helps you create content that influences their answers. There is a common misconception that AI models just pick the first search result on bing/google, but that’s not true. They pick the most relevant page that answers the question, and they do this multiple times just to answer one prompt. So to understand whats’s going on you need to really follow the AI thought process. HeyAmos visualizes the AI thought process behind every answer:
  • Reasoning steps
  • Web search calls (what queries the AI used)
  • Evaluation factors
  • Final answer synthesis
Here you can for example see all the steps that ChatGPT goes through in order to answer the prompt “What are the best airlines flying from Singapore to Europe”: Image

​
Evaluation factors

In order to answer prompts like “What are the best airlines flying from Singapore to Europe” or “What is the best CRM for small businesses” the AI models choose a set of evaluation factors that they use to compare brands and options. In this example you can see how pricing, service quality and comfort where the main evaluation factors. This analysis helps you understand what part of your offering needs to be highlighted more in your content – in order to increase visibility. Image

​