Massive linguistic models, like ChatGPT and GPT4, are revolutionizing the domain of natural language understanding and AI because of their adaptive capabilities and broad applicability. Yet, these models, being opaque in nature, sometimes struggle to consistently retrieve accurate information or generate false information ("hallucination"). At the other end of it, a robotic templatized response does not allow for a free-flowing conversation to happen although it does provide surety on the accuracy of responses. We suggest a middle ground where we figure out the right intent of the user and guide the user through defined flows to arrive at the right answer. By integrating a classifier to identify user intent, followed by a retrieval-augmented generation process, we can merge the adaptability of massive linguistic models with the precision of template-based systems. This ensures both a dynamic conversational experience and reliable information delivery, bridging the gap between fluidity and accuracy in AI-driven communication.