fbpx Smiley face

Artificial intelligence to manage the online reputation of your restaurant


Numerous studies show that improving scores, digital portals opinion, with a star It increases the turnover of a restaurant between 5% and a 9%. They also say that a third of consumers will not eat at a restaurant less than 4 stars.

Numerous studies show that improving your score with a star increases your turnover by 5 a 9%. They also say that a third of consumers do not eat at a restaurant less than 4 stars.

How much does manage Online Reputation? Reviewing the volume portals opinion it has increased in recent years, to a point that is very difficult to manage professionally all comments that restaurant patrons share. The average time to read a review is 35 seconds. Some restaurant chains receive a sock 100 comments a day and needed to read them all 30 hours per month, to which must be added, the average time to answer a review is 2 minutes on average.

Today, less than a 5% the reviews are answered Tripadvisor, and the variety increases given the volume of portals to be monitored as ElTenedor, Facebook, Google My Business, Yelp, Atrápalo and many more depending on the country or the city where the restaurant is located. To add a little more complexity, Instagram has become one of the most popular platforms customers no longer share only text but also photos, videos o stories which are active only a few hours, but can have thousands of visits.

Then, I would like to show you some of the features that have launched in recent weeks in CloudReputation, platform to manage reputation data most advanced on the market.

What is behind an opinion

Most of the online reputation services Traditional handle meta-data: Review structured information such as date, the score, number of Likes, etc. The challenge we have is we put training algorithms able to understand the text as a human being.

Meta-data Trip Advisor review

Meta-data Trip Advisor review

In CloudReputation we use algorithms to understand the feeling of an opinion. train Artificial Intelligence models able to understand the feeling and emotion in customer feedback. Techniques we use natural language processing (NLP) to understand the written text. The vast number of unique languages, contexts and complexity make this an extremely challenging task.

The results are surprisingly good, identifying the feeling of each of the sentences of the reviews and magnitude showing the relevance of the phrase in the text. We also highlight entities to provide the necessary information and quickly find the most important in the text.

Analysis Example feeling Phrase, magnitude and prominent entities

Analysis Example feeling Phrase, magnitude and prominent entities

What's behind a photo on social networks

Each of the photographs that your clients post on social networking sites also include information that can be automatically extracted using Artificial Intelligence models. We have trained two models of image recognition:

  • Demographics of subjects: This model analyzes images to people and gives us information on age, multicultural sex and appearance of each detected face based on facial features. This model is ideal for sorting your audience and customers who appear in social networks: Is your most popular restaurant for young? How many young? Will more for men or women?
  • Recognition food: the model 'Food’ recognizes over 10.000 food pictures to the level of ingredients. Most notably, we believe we are the only ones who have customized this model for Spanish food: we can detect the most common dishes such as paella in Spain, gazpacho or potato omelette.

Example information automatically extracted from an image of Instagram

Example information automatically extracted from an image of Instagram

With this information added, We can provide information on the percentage of your clients postean photographs on food or subjects, What are the most popular photos of food or the average age of the people who appear in photographs.

autoresponders reviews

The answer to positive reviews often it leaves out, as business owners focus on damage control by negative customer reviews. After all, How often do you see a positive review of Tripadvisor goes viral?. Nevertheless, serve customers who are already excited about a restaurant is the perfect way to create loyal followers and a network of solid references. In fact, he 78% consumer believes that news of the establishment management about their reviews makes them believe that the company cares more for them.

To help with this task we have recently launched a system that automatically does the work of writing the response to a review. Neural networks we have created are programmed to learn behaviors through training. In general, This functionality goes well for those restaurants who have no customer service team or has one large and overloaded, and would like the first level of interaction is instantaneous for common problems.

At the moment we think our AI is able to handle 20-70% of incoming reviews and this can mean big cost savings for the companies. No matter how many people are talking to a restaurant or business in social networks, our algorithms can handle all at once.

The vision is that our system is a wizard, automatically be responsible for responding personalized and, appropriate when necessary without any human interaction. From time machine answers are not automatically sent. Instead, Choose among the restaurant suggested by the responses IA CloudReputation and endure answers in Spanish, English and French.

Example suggestion response to a comment on Google My Business

Example suggestion response to a comment on Google My Business

Transforming information into knowledge

With these advanced techniques, Now we collect all the information imaginable review or photo. IA Our models are put to work and automatically all information is labeled and becomes relevant knowledge, actionable, ready to be exposed in our simple interface to always comprehensible reputation of your local. The unknown is now known.

What is more important, It is that our application learns and our models always improve and become more intelligent and accurate. continually we use the most recent data, to ensure they always reflect business reality today, your restaurants providing the information you need to make smarter decisions and build competitive advantage.

according to Forrester, in 2018, bots have replaced and / or increased 311 000 office or administrative positions, Y 260 000 sales positions and related posts. Many companies seek to increase the number of human workers with technologies including artificial intelligence. But make human work side by side with robots remains a challenge. Our goal CloudReputation is provide tools so you can deal to offer a better quality of service, more human interaction and a good deal to customers who visit your restaurant.

About the Author

Product Manager at IBM Watson specializes in Artificial Intelligence and founder of CloudReputation, Spanish startup based in Silicon Valley that offers online reputation for restaurants


Leave a comment