Text this: Comparative evaluation of conventional search architectures and large language model-based approaches