A new report from EyeforTravel gets under the skin of travel chatbots and reveals how to successfully create and deploy a bot.
Chatbots are getting increasingly sophisticated and capable of mimicking human conversation behaviors, but what should these bots aim to do and how can brands deploy a chatbot successfully? The new Are Bots Worth the Bother? Report from EyeforTravel and Travelaer reveals all.
The good news for brands is that the cost of implementing a relatively simple chatbot is not a huge investment and is coming down as more providers move into the space and natural language processing becomes more sophisticated. The cost of implementing a bot typically ranges from 15,000 to 50,000 euros, according to the report. “These aren’t a significant investment for an airline when they’re spending hundreds of thousands, if not millions, a year on other digital experience products,” says Mike Slone, chief experience officer at Travelaer.
However, travel brands need a foundation of data and the resources to monitor, test and support the bot. Without this, there is potential for the bot to go astray. Data is key as brands first need to identify and understand what are the key pain points for customers first and seek to address these, rather than setting objectives beforehand.
For luxury hotel brand Edwardian Hotels they began by creating apps for staff to record notes on guests, service rooms, check breakfast tables and monitor their work schedules. Icelandair started with their social media team and looked through what came up most frequently in their interactions.
In both cases, these brands looked at what their customer-facing staff were being asked most by their customers and then sought to address these queries, freeing up time for their staff to focus on more complex tasks.
“Instead of saying ‘we’re going to build an acquisition tool’, we look at the customer journey, how a customer interfaces with a firm and its existing digital products,” he says. “We look for a gap where there’s a huge customer need and maybe a chatbot could help. Then we come up with a strategy to apply the chatbot to solve it – that way, we know the chatbots will be much more successful in terms of customer satisfaction and metrics.”
Brands then need to monitor their bot and make sure that it is learning from each interaction in the right way. “The first version was built around understanding key words but had difficulties with the difference between: ‘what’s the luggage allowance for Europe?’ and ‘I lost my luggage in Europe.’ It’s a delicate situation to handle a bot telling the customer who has lost his luggage that the luggage allowance is two bags!” said Gudmundur Gudnason, director of digital business development for Icelandair.
The results speak for themselves as the brands in the report, which include Skyscanner and KLM, were able to drive bookings, boost ancillary sales and increase customer service. Customers have even been fooled into thinking they are talking to an actual person and leaving TripAdvisor reviews or cash tips for their artificial helpers.