Proactively address needs.
By better knowing their guests, Porto Bay could respond to guest needs and desires in a more proactive, almost magical, manner.
In the spring semester, our team spent four months performing in-depth research on the hospitality industry and the client’s existing products and services. During this time, we employed a wide variety of research methods, including contextual inquiries, service safaris, competitive analysis, diary studies, guest surveys, and guest interviews.
Over 100 hours worth of contextual inquiries were scheduled with eight different hotel departments: marketing, central reservations, IT, front office, food & beverage, maintenance, housekeeping, and entertainment. These observation and interview sessions were crucial for gaining a thorough understanding of Porto Bay’s operations. We were exposed to the company’s culture and philosophies, and, perhaps more importantly, made aware of challenges facing the staff.
Guest surveys and interviews helped us familiarize ourselves with the very people Porto Bay hoped to better engage. We learned about their attitudes and habits, their relationships with hotel staff, and their reasons for choosing Porto Bay. Suggestions on how to improve the guest experiences occasionally emerged from these sessions.
"Interacting with guests is both the best part and the most tiresome part of our job."
"When you are on vacation, you are always in a good mood, so to match the guest, we also need to be in a good mood, even though the work is stressful"
"The main challenge is that you cannot be yourself, you have to change your mood according to the guest and how they are feeling"
To consolidate our research, we organized over 1,400 notes into an affinity diagram, eliminating redundancies along the way. As our affinity diagram slowly took shape, groupings and themes revolving around staff needs, guest needs, and the role of technology formed organically.
Grouping our notes into an affinity diagram led to over 152 separate insights into the portobay expererience. Further abstracting these insights and focusing on the most powerful and impactful items led us to derive the following series of key themes as one method of organizing our findings. At the core of these themes is one central concept, personal Touch, people like to be recognized. And in doing so, Porto Bay can exceed expectations.
We considered many different technologies we could utilize to tackle the opportunities we identified. Early discussions focused on spatial personalization and room customization. We also talked about giving guests the ability to create a personal avatar that would appear on both personal and public screens, provide personalized recommendations, and subsequently enhance the individual experience. Ultimately, we found that personal avatars would not be a good fit for the brand Porto Bay has created for itself. However, the notion of using a non-human party to facilitate personalized interactions remained with us, leading us to explore conversational interfaces such as the Amazon Echo and Apple’s Siri. We realized that an conversational user interface designed specifically for use in hotel rooms could very well accomplish Porto Bay’s goal of better understanding guest needs and thus improving their experience.
After trying out our crazy ideas, we were ready to define our focus and decide on a design idea. We did another round of ideating and eventually fell in love with the idea of using intelligent conversational agents to enhance the hotel experience.
When it came time to validate our design idea, we organized a number of in-room user testing sessions with real Porto Bay guests. In total, we ran tests with 13 guests at two different properties: Vila Porto Mare in Madeira, and Porto Bay Liberdade in Lisbon. In these tests, we asked guests to complete tasks “using the in-room technology.” Tasks included placing a room service order and making a taxi booking for the near future. As guests “conversed” with a physical prototype, our team worked behind the scenes to manually control the responses of the conversational agent and the television screen. The feedback we received from guests in these user testing sessions was overwhelmingly positive. We heard things like “I felt like I was speaking with a human after a very short time” and “I’m not a computer person, but this was very easy to use!”
We also had the opportunity to validate our design with members of the Portobay staff. Our team conducted participatory design workshops with 10 staff members at three different properties: Vila Porto Mare, The Cliff Bay, and Porto Bay Liberdade. During these sessions, we led staff through activities like speed dating, card sorting, and projective sketching. Because staff members play such a crucial role in the guest experience, we recognized the importance of involving them in the design process. The selected workshop activities all relied heavily on staff participation and input. When presented with an early iteration of Eva, a member of the guest relations staff summed up his thoughts with a simple sentence: “I love it.”
Our concept validations encouraged us to move forward with our conversational agent idea, which we named Eva.Eva interacts with guests using two forms: in-room speech and mobile messaging. She also connects directly to the in-room television, and uses it to display information relevant to the guest’s request, when appropriate. For example, a statement like “Eva, I’d like to order some dessert” would surface the room service menu. Eva’s other form is a chatbot available on platforms such as Facebook Messenger or SMS. This allows the guest to interact with Eva while outside the hotel room, even before the guest has arrived at the hotel.
After gathering feedback from all of our stakeholders, we generated these key design principles that our conversational agent, Eva, needs to adhere to.
For Eva’s physical form, we felt it was important to design something that would feel appropriate for all of Porto Bay’s properties - from their contemporary city hotels to more homey resorts. Although early iterations of Eva’s body were inspired by origami and geometric shapes, we eventually found that a gentle curvature would help Eva appear more inviting than traditional electronic devices.
Eva’s visual language, reflected on both the physical device and digital screens, needed to fit in with the Porto Bay brand while still maintaining a distinct identity. After exploring many potential ideas, we decided that intertwined speech bubbles combined with a rounded, script-like typeface - Signerica Fat, one that Porto Bay occasionally uses - conveyed in an elegant manner that Eva was a conversational agent.
To build our prototypes, we utilized API.AI, a platform for developing conversational user interfaces. With API.ai, we were able to create the backend technology stack that powered Eva. This is a snapshot of what happens in the backend when a guest interacts with Eva in the hotel room: Embedded software hidden within the physical device recognizes the wake word “Eva” to capture the guest’s voice commands. Next, the captured audio is relayed to a cloud service, which uses a speech recognition engine to transcribe the human speech to text. The text then undergoes natural language processing. The cloud service then analyzes the intent of the parsed text, detecting user commands and actions. With a request like “Eva, where can I get a haircut?”, the key words “where” and “haircut” would help communicate that the speaker is looking for the location of a salon or barber shop. Once information needed to fulfill the request is gathered, the cloud service uses text-to-speech technology to generate a response that is outputted through the the physical device.