Enterprises are reaping the rewards of incorporating large language models (LLMs) into software applications, but doing so creates challenges that technology leaders haven’t faced before. Use prompt engineering and advanced LLM architectures to mitigate these challenges. By mixing various prompt engineering techniques (e.g., chain-of-thought prompting, agent frameworks) with advanced architectures (e.g., retrieval-augmented generation [RAG], discrete data products, model cascades/stacks), development teams can achieve consistent and accurate outputs, thus building trust in generative AI (genAI)-infused applications.