Data-to-text is a technology that automatically converts structured data into natural language texts. It is a sub-area of Natural Language Generation (NLG), in which data forms the basis for text creation.
Data-to-text systems make it possible to generate, update and localise high-quality texts based on structured data. Instead of generating each individual text, text creation and updating becomes a scalable and even autonomous process thanks to a centralised set of rules.
The advantages of data-to-text include
- Efficiency: Acceleration of text creation and updating.
- Scalability: Enables the production of texts on a large scale.
- Consistency: Ensures that texts are standardised and of high quality.
- Control: Enables the author to retain control over the text results.
- Multilingualism: Enables the creation of texts in up to 110 languages.
- Up-to-dateness: Enables texts to be kept up to date at all times, as changes in the data are automatically converted into new texts.
- Combination with AI: Enables integration with GPT and DeepL for improved text creation processes.
Data-to-text is used in various industries, particularly in e-commerce, where it is used for the automated creation of product descriptions. It enables companies to make their text production more efficient and optimise their product communication.