When it comes to artificial intelligence (AI), the way it’s represented in movies and books creates assumptions — and confusion — about what it can do and the practicality of its application in the business marketplace.

Because of this, the hospitality industry tends to see AI as a buzzword among companies that want to seem innovative and cutting edge. In the simplest terms, AI refers to systems or machines that mimic human intelligence to perform tasks and can improve themselves based on the information they collect.

AI works by combining large amounts of data with fast, iterative processing and intelligent algorithms. This allows the software to learn automatically from patterns or features in the data. There are four types of AI:

• Reactive AI  functions the way it was programmed with a predictable output based on the input it receives. Examples include email spam filters and Netflix recommendations.
• Limited memory AI can complete complex classification tasks and uses historical data to make predictions. Think predictive personalization and self-driving cars.
• Theory of mind AI exists when machines acquire decision-making capabilities equal to humans. This type has not yet been fully achieved.
• Self-aware AI exists when machines are not only aware of the emotions and mental states of others, but also their own. Bringing this type to bear depends on expanding robust theory of mind AI.

When it comes to digital marketing and hospitality, limited memory AI is best applied via advanced algorithms to analyze historical data faster and at multiple levels. This intelligent processing is key to identifying and predicting rare events, understanding complex systems and optimizing unique scenarios. This autonomous algorithmic driver comprises the basis of predictive personalization.

Before we investigate the details of predictive personalization, it’s important to understand basic personalization. It allows businesses to target users based on:

• Timing: Date range, days of week, time of day, time zone
• Demand: Stay dates, length of stay, booking value and availability
• Travel party: Number of adults, children and rooms
• Visitor profile: Location – country, state or city
• Source: Google, Instagram, etc.
• Visitor behavior: Members versus non-members, previous interactions
• Custom targeting: Device, URL variables, data layer variables

This type of personalization helps boost conversion by delivering the right message to the right person. But it doesn't understand the user’s purchase intent. Therefore, predictive personalization is a two-step process. First you apply machine learning techniques to understand user behavior. Then you personalize their experience by automatically presenting the best content and offers for that individual.

Predictive personalization builds a score for the user based on their past behavior (before coming to the website), current behavior on the website, any interaction with on-site personalization messaging, and external market data.

For example, there is one AI built for hotel companies that anonymously captures every event of the online user journey from numerous hotel websites across 100 countries. It processes tens to hundreds of millions of data points every single day. By tracking over 150 variables about each user, the algorithm leverages machine learning to find patterns by comparing against previous users. The algorithm then makes a behavior prediction to identify how likely a user is to book or not. AI helps marketers determine the right time and the right booking intent.

The majority of hotel bookings come from a very small percentage of visits (roughly 3%). The other 97% of users that visit a website leave without completing a reservation. Furthermore, of the typical booking engine visitors, the lowest intent visitors (the bottom 30%), generate only 4% of a hotel's total bookings. Inversely, just 10% the highest intent users generate 49% of a hotel's total bookings. Predictive personalization solves the marketing problem to differentiate website visitor intent and truly optimize marketing campaigns.

Once implemented, predictive personalization can address two issues hotel marketing faces. Either they aren’t running value-targeted campaigns or they’re running promotional campaigns that target everyone. A hotel can run value-targeted campaigns and drive more bookings and revenue. The solution will split promotions and uses targeted offers for low-intent users (and drives 20% to 30% more bookings) and targeted upsells to high-intent users (which generate up to 30% more revenue).

AI can optimize campaigns for hotels running promotions that don't differentiate based on intent and save money by splitting the campaigns and reducing incentives for high-intent users. Suppressing promotions to those with high intent will save the hotel up to 80% on unnecessary discounts to these specific bookers.  

When properly implemented, conversion rates for low-intent users increase between 50% to 60%  against a control group. High intent user conversions increase over 150% over the control group.

When it comes to perception, AI probably will remain a buzzword, especially if people see it as synonymous with self-aware AI. However, if you shift the conversation to focus on a single type of implementation as it relates to predictive personalization, AI is not only real, it’s currently being successfully deployed in the marketplace revealing and driving hidden opportunities to maximize conversion, bookings and revenue.