State-of-the-art deep learning is necessary for natural conversations in a conversation intelligence system, but it's not enough alone. The system needs deep understanding to model, generalize, and run analytics using all its knowledge, just like a human brain. Artificial neural networks are the most cutting-edge algorithms right now, representing the structure of a human brain modeled on computers with neurons and synapses organized into layers. Deep learning differs from machine learning in that it uses large amounts of data, takes longer to train, and is computation-heavy, but it's better suited for complex modeling and recognizing sophisticated patterns. However, deep learning in conversation intelligence doesn't match the sophistication of a human brain, as it only learns statistical patterns rather than understanding its meaning in a flexible and generalized way. To achieve this, techniques such as abstract knowledge modeling, inferencing systems, and mathematical approaches for generalization and modeling are needed. A combination of modeling with deep learning can be used to build a system that has deep understanding, not just learning, by acquiring knowledge, broadening inferences based on conversation data, and running analytics from all the knowledge.