Conversational systems should generate diverse language forms to interact fluently and accurately. In this context, natural language generation (NLG) engines convert meaning representations (MRs) into sentences, directly influencing user perception. These MRs usually encode the communicative function (e.g., inform, request) via dialogue acts and enumerate the semantic content with slot–value pairs. In this work, our objective is to analyse whether providing a task demonstrator to the generator enhances the generated sentences of a fine-tuned model. This demonstrator is an MR–sentence pair extracted from the original dataset that enriches the input at training and inference time. The analysis involves five metrics that focus on different linguistic aspects and four datasets that differ in multiple features, such as domain, size or lexicon. All experiments are conducted with GPT-2 models, so the findings should be interpreted within this model family. To the best of our knowledge, this is the first study on dialogue NLG, implementing a comparative analysis of the impact of MRs on the generation quality across domains and corpus characteristics. Our key insight is that the proposed enriched inputs are effective for complex tasks and small datasets with high variability in MRs and sentences, and also for zero-shot settings. Moreover, the analysis of the metrics shows that the semantic metrics capture the generation quality more accurately than lexical metrics. In addition, those trained with human ratings can detect omissions and other subtle semantic issues. Finally, the evolution of the metric scores indicates that the generative models adapt quickly to different tasks.