The company claimed that its analytics tool was 80% accurate in identifying sarcastic comments online as it generates reputation reports based on social and traditional media material; for instance if Air France was two hours late and someone tweeted “thanks for getting us to Heathrow two hours late”, the company said the tool would identify the sarcasm inherent in the tweet.

Using a combination of linguistics, semantics and heuristics, it covers 29 languages including Chinese, Russian and Arabic.

Speaking to the BBC, Spotter’s UK sales director Richard May said: “Nothing is fool-proof – we are talking about automated systems. But five years ago, you couldn’t get this level of accuracy – we were at the 50% mark.”