What AI agents do you use?

In our organization we employ several types of AI agents depending on the task:

1. Rule‑based agents – simple if‑then logic for deterministic workflows.
2. Supervised learning agents – classification or regression models (e.g., logistic regression, random forests) that predict outcomes from labeled data.
3. Reinforcement learning agents – used for sequential decision problems such as recommendation or robotic control.
4. Generative agents – transformer‑based models (GPT‑style) for natural language generation and summarization.
5. Retrieval‑augmented agents – combine a knowledge base with a language model to answer queries.

Each agent type is chosen based on data availability, latency requirements, and interpretability needs.