Artificial Intelligence (AI) is the buzzword in tech nowadays. Many executives are embracing it and organizations using it.
The smarter your system, the better the disruption it will cause. But how much does it cost really? In the hospitality business, do we even need to use cloud infrastructure to get the most out of it?
As an IT Hospitality guy, the team I’m part with is always presented with solutions that can allegedly make our lives easier :). I am amazed by the advancement and the possibilities based on the solutions that were presented to us. So I asked myself, what does it take to build it?
The first thing I need to do is to find data, lots of it. So I turned to customer reviews of one of our competing hotels. I developed data collection script to mine customer reviews from TripAdvisor, Agoda and Booking.com. I optimize my script so that it can collect huge amount of data in a very short time by implementing multi-threading.
Using Python’s Natural Language library, I then created a Sentiment Analysis using Vader. Vader process any statement and classifies if it’s positive, negative or neutral. I originally intended to use Naive Bayes classifier but I cannot seem to find a proper data set for hotel reviews (please message me if you know one).
After I classified the comments, I realized I need to evaluate all negative comments and re-classify them according to the hotel processes or teams. This can certainly help hotel GMs in understanding customers pain points. It will also assist the GMs in knowing which teams need some accountability and proper boost.
So I created a Predictive Analyzer script which classifies the comments into ‘Housekeeping’, ‘Reception’ and ‘Dining’. These are 3 major areas the customers deal with during their stay. I then developed my own custom library for these 3 areas so I can train my AI program to make the best classification.
Present the Data
After churning out and getting the results. I created a simple UI using Google Chart and HTML5 Table.
This will make it easy for any management (GMs or Executive) to do the proper assessment and make the necessary action.
The scripts I created have less than a thousand lines combined. Python libraries certainly helped me in keeping it short. It can ran on a Raspberry pi and can definitely run on a cloud environment. The script can also be used not just for evaluating on-premise experience comments but also online experience comments, you just need to create another library appropriate for such tasks. You can even use it for competitive intelligence where you can check how your competitors fare. The possibilities are huge.
Creating an AI initiative doesn’t have to be expensive and complex. Hotels just have to know which part of their processes need it and strategize how to achieve it.