A recent survey of ecommerce professionals has revealed natural language generation (NLG) for product descriptions as a top priority for the industry.
In the survey conducted by Automated Insights, a company dealing in automated text generation, 94 per cent of respondents identified product descriptions as important to their brand’s marketing strategy. However, only 21 per cent said that descriptions were consistently written in the tone of their brand.
Automated Insights claim that NLG is the answer to this problem. NLG is a branch of artificial intelligence concerned with turning data from spreadsheets or database tables into natural-sounding language, with an understanding of the rules of grammar and syntax that elevates it above the awkwardness of the typical mail-merged document. Users are given the ability to control their narrative to ensure that product descriptions fit their brand’s identity. NLG is touted as being faster than manual copywriting, as it is able to generate thousands of pieces of unique content in a few hours from a single template.
NLG can also assist in the creation of category landing pages for ecommerce stores, in search engine optimisation, and in generating customised emails for email marketing that sound less like spam and more like personal communications.
Ecommerce site owners might want to be sceptical about some of the exaggerated claims made for the abilities of AI engines to replace human writers at the current stage of development, however. Microsoft recently unveiled Tay, a chatbot that was supposed to learn from interactions with humans in order to be able to hold natural-sounding conversations on Twitter. Within 24 hours, “she” was hurling racist abuse, praising Hitler and boasting about smoking cannabis in front of the police. It appears that sounding like a real human being is not always a good thing.