Unleashing the potential of NLP: a case study

Sciant
Sciant
Published in
3 min readOct 10, 2019

Natural language processing (NLP) is the ability of Artificial Intelligence programmes to understand and respond to natural human speech, whether in spoken or written form. It is currently one of the biggest disruptors in business with its application via chatbots saving potentially tens of billions of dollars by automating processes and cutting staff costs.

It is also revolutionising the travel and hospitality industry. With its ability to simplify and automate anything from booking requests to complaints, queries and customer service interactions in a speedy and round-the-clock manner, it fits perfectly with the increasingly automated and online nature of the travel market. Indeed, so successful are chatbots becoming at interacting with humans that 69% said they would prefer to deal with one than a real person. All of which is probably why some of the biggest players in online travel and tourism base their operations around chatbots run on NLP, such as Expedia, KLM and Booking.com.

One example of how NLP has been used to great effect in the sector is with HotelResBot. The German-based startup uses AI to automate responses to email booking requests — 25% of all hotel bookings — saving time and repetitive work for staff, increasing revenue and providing a better more personalise service to guests. As founder, Brendan May, states of traditional methods: “Booking a hotel by email is a terrible experience for everyone involved. The reservation agent hates answering emails (and it shows). The potential guest needs to wait forever to get an incomplete and unattractive offer. The hotel manager suffers with an expensive and untrackable booking process.”

HotelResBot’s NLP platform, HERA, reads these emails automatically then scans room rate and availability data to provide an automated response with a tailored and bespoke offer within seconds, 24 hours a day seven days a week.

HERA can also understand and reply in multiple languages and deal with complex requests such as multiple bookings. In fact the system is able to correctly recognize 95% of booking requests that hotels receive.

Sciant helped develop HERA into a viable product by taking a generic open-source NLP platform and modifying it specifically for the hotel industry to create a custom programme that outperformed generic platforms. One of the challenges was not just to provide a system that could read and respond to emails, but one which could seamlessly interface with backend processes to scan room rates and availability via an API. Sciant also developed a complex offer selection logic that was customisable to individual hotel selling strategies, ensuring guests receive the best available matching rate and room types.

To ease set up and staff use, Sciant developed an easy-to-use interface via a Microsoft Outlook add-in which is available free from the Microsoft AppSource Store. Finally Sciant helped develop, monitor and optimize key performance indicators (KPIs) to improve email processing for better accuracy. The result was a quick increase in the proportion of emails processed successfully in a fully compliant and on-brand style with full conversion tracking and detailed reporting. “Sciant has been able to understand our requirements as a customer but also our end-customer needs. We’re not just talking about our customers (hotels), but also their customers (guests),” commented May.

It’s just one example of how NLP is changing the game in the travel and hospitality industry. One thing is for certain — there are many more examples to come.

--

--

Sciant
Sciant
Editor for

IT services company specialising in hospitality and travel, logistics and transport.