Prerequisites: Set up your AI Agent’s brain

Cognigy’s AI Agents utilize the power of Large Language Models (LLMs). You will need to set up two Large Language Models (LLMs) in Cognigy.AI:

1. An LLM for language generation (e.g. OpenAI GPT-4o)
2. An LLM for embeddings generation (e.g. OpenAI ada-002)

To set this up, navigate to the sidebar and select Build -> LLMs. Here you configure the language model, such as Azure OpenAI GPT-4o or others. In case you want to allow image upload in a conversation, the model must support images as input (see chapter “Enable your AI Agent to see images”).

If you want to optionally add knowledge to your AI Agent, you'll also need an embedding model to index and search through your knowledge base (see “Add knowledge to your AI Agent”). For this purpose, the embedding model Azure OpenAI Ada is a popular choice.

Once all models are configured, go to the sidebar and select Manage -> Settings -> Knowledge AI Settings. Here, you can assign the embedding model for knowledge search and the generative language model for answer extraction.

➡️ Now you are ready to create and run AI Agents.