Business Challenge story
The hotel business serves as a prime example of how two inexorable forces - competitive labor cost pressures and rising customer expectations - can make change inevitable. Like many service industries, hotels are under increasing pressure to control their labor costs. As a fundamentally brick-and-mortar business - with housekeeping duties to perform, room service to deliver and guests to greet - hotels are highly reliant on their human staff. As such, any efforts to cut back on staffing levels tend to have a rapid and adverse impact on the quality of the guest experience. In an already highly competitive industry, the rising prevalence of hotel guest reviews in social media - along with the emergence of a whole new set of competitors for services such as Airbnb - has made customer satisfaction an even more important metric. The fact that young, smartphone-wielding snake person travelers have grown accustomed to near-instant gratification puts further pressure on hotels to deliver a top-notch, convenient solution. A California-based startup called Go Moment Inc. recognized that these intersecting forces put an especially tight squeeze on concierge services. After all, front-desk concierges are the figures guests turn to for just about every issue they may have after checking in. The questions or requests are typically not complex: What’s my wifi password? Where is the pool? Can I get fresh towels? But they take time, a commodity in extremely short supply for the average harried front-desk worker. In terms of customer service, being in the center of the action can put these employees in a lose-lose situation. Answering the phone can mean that new check-ins wait in line, while putting guests on hold for minutes can risk trying their patience and undermining satisfaction. Go Moment saw a win-win opportunity to use technology to relieve the pressure on the front desk while providing faster resolution and better convenience to guests. Its vision was to deploy a virtual concierge capability to complement its existing front-desk staff. The company envisioned a solution that would be immediate, automatic and easy for guests to use, without requiring them to learn a new interface, scan a code or download an app. Like human concierges, the service would not only answer basic questions but also trigger action on real requests. As such, the solution would also require the capability to deduce the intent of guests’ service requests and answer them directly or route them to the right department for action.On the strategic level, the solution would also have the ability to analyze guest requests for insights into how to improve service.
The traditional processes of the hospitality industry are under pressure on two fronts: ever-higher quality expectations from guests and the need to keep staffing costs down. The hotel front desk - the place where guests go for answers - is the pressure point where these forces converge. Go Moment saw the near ubiquity of smartphones and text messaging as a ready-made platform for reducing front-desk pressures while raising service quality. The company created a virtual concierge service - known as Ivy - that lets guests submit questions or requests through simple free-format text messages. Through connections to cloud-based natural language processing (NLP) services, Ivy recognizes and classifies these queries and responds in an instant with answers or actions. Ivy can also detect from the tone of the message when a guest is unhappy and automatically escalate the response to speed problem resolution. To deliver the NLP capabilities that would be the backbone of its virtual concierge service, Go Moment opted to design it with application programming interfaces (APIs) to cognitive services in the IBM Watson Developer Cloud portfolio. The Ivy service enables guests to send requests in the form of Short Message Service (SMS)-based text messages through their mobile phone. Upon receiving the request, the solution uses the IBM Watson Natural Language Classifier Service to interpret the guest’s intent, classify it and return values that will trigger corresponding actions, such as texting a wifi password or alerting housekeeping to deliver towels. The solution currently serves guests in English, Spanish, Chinese and Japanese. What makes the solution extra smart is its ability to look deeper into the text of messages for the signs of a potentially dissatisfied or irate guest. Roughly one hour after check-in, the Ivy service sends a text to guests asking them how they like their room. The text of that message is then fed into the linguistic analysis models of the IBM Watson Tone Analyzer Service. If the analysis detects an unhappy guest, the solution automatically escalates the service priority for that guest and alerts the most appropriate member of the staff. Using the system’s ability to detect dissatisfaction at its early stages, hotels can take quick action to resolve issues and restore goodwill.
Enabled increases of up to 20 percent in problem resolution satisfaction metrics for hotels deploying the Ivy solution; Enabled as much as a 35-rank jump in TripAdvisor ratings, reflecting increased overall satisfaction; Increases the productivity of hotel front-desk staff by answering up to 90 percent of customer queries automatically, enabling staff to focus more time on issues requiring direct human intervention to resolve