NLG is being used in many real-world contexts. Indeed, the usage of NLG is evolving extremely rapidly, with new use cases appearing every month. What is possible to accomplish in specific use cases is also changing fast, as better technology opens up new possibilities.

As always, the book focuses on fundamentals, not the latest developments. I will first discuss generic issues in applying NLG to real-world use cases. Then I will look at several long-standing use cases of NLG: journalism, business intelligence, summarisation, and healthcare.

Resources: Selected blogs

• AI professionals also focus on change management
• INLG: What real-world NLG users want
• LLM hype brings memories of IBM Watson
• Pain Points in Health NLG: Data, Evaluation, Safety
• Sports NLG: Commercial vs Academic Perspective
• Real-world utility is based on many things
• Why is adoption of AI in healthcare so slow?

Resources: Papers and surveys

• A scoping review of robustness concepts for machine learning in healthcare (useful in other domains as well)
• Generative AI in Journalism
• How IBM Watson Overpromised and Underdelivered on AI
• Promises and pitfalls of artificial intelligence for legal applications

Resources: Web sites

• Generative AI in the newsroom