Artificial Intelligence in the Marketing Sector

Mar Castaño

4 January 2018

180104 gr rebelthinking tv mar wp 1

In mid-2016 Forrester listed the five emerging technologies that will have the largest impact on the business world in the next five years: the Internet of Things, Artificial Intelligence, augmented reality, intelligent agents and hybrid wireless networks.

mar 1


Today, thanks to Artificial Intelligence (AI), we are able to understand consumer behavior, interpret their needs and make decisions in real time. Its application is already unstoppable in the world of Marketing. The advantages are clear and it is up to you to decide how and where to implement it in each organisation.

Smart Insights classify AI techniques throughout the customer lifecycle that can be implemented by companies of any size – they are not only within the exclusive reach of technological giants- and that contribute to the customer’s conversion funnel.


mar 2


It groups them into three typologies depending on whether they use Machine Learning -learning from past datasets and generating propensity models-, propensity models -which indicate events such as the rating of a lead based on its conversion probability- or Artificial Intelligence applications -which carry out tasks that are normally done by a human operator such as answering customer questions or creating new content.



In this phase AI models are aimed at attracting visitors to the site and providing for those who come from more attractive experiences.


AI Generated content

CrewMachine is a platform capable of identifying gaps in e-commerce content (keywords, product descriptions, category pages, style guides) and automatically suggests content where it is missing, so that it can be prioritised and modified. At the same time, it learns from the previous publications to improve its suggestions for future proposals.

There are certain areas where AI content generation can be very useful. Wordsmith is a platform that allows you to very easily define templates for paragraph and phrase generation, from which it automatically creates content. In 2016 alone, it created one and a half million products.

The home automation company digitalSTROM uses it to inform its customers about their energy consumption and how they can make better decisions to saving money. converts its users’ training data into a personalised and stimulating email, and Orlando Magic sends personalised and individualised offers to season ticket holders.


Content Curation

The curation of AI-enhanced content makes it possible to more effectively attract visitors to the site by showing them relevant content. Suggestions that we see on Amazon, like ‘customers who bought this product also bought’ or ‘products frequently purchased together’, use this type of technique. However, it is also useful in displaying content on blogs or newsletters.

Curata and PublishThis are business solutions with discovery engines and personalised content recommendations.


Voice Search

Today 20% of mobile searches are conducted orally, and it is no longer unusual to turn to Siri, Google Now or Cortana. This technology will change future SEO strategies and companies should consider making these new sources of organic traffic appear when people invoke them. Sherry Bonell provides the keys to this in her article: identify the questions that can be asked about the business, product or service (tools such as, StoryBase, Question Samurai and SemRushque can help this); prioritise long-tail keywords; create content to answer those questions, etc.


Programmatic Media Buying

Programmatic purchasing of media can utilise propensity models to more effectively target adverts to the most relevant users. In addition, AI can help us recognise sites where the advert may have less impact and remove them from the sites listed for advert placement.



Propensity modeling help us make predictions about client behaviour and which of our products or services will best suit their needs along this phase.


Predictive Analytics

Predictive analytics involves the use of data, statistical algorithms and automatic learning techniques to identify the probability of future results based on previous data. The more accurate and precise the data provided to the model, the better it will be. Regression analysis in its various forms is the main tool that companies use for predictive analysis.

The most common models include the prediction of customer behavior and preferences, probability of conversion, propensity to repetition or churn, cross-selling or the next sale. Companies such as Amazon and Netflix are the clear leaders in this field.

Some providers in the predictive analytics domain are IBM Analytics, Optimove, AgilOne, Infogix.


Lead Scoring

The propensity models are also used for:

  • qualifying leads according to different criteria, so that their potential can be determined at the time of acquisition or implementation of an action, and the profitability of dedicating efforts and resources to it;
  • identifying and acquiring leads with similar characteristics to existing customers;
  • automatically dividing leads or clients so they can send personalised messages or follow different communication and customer service strategies.

These techniques require large volumes of sales to properly construct and shape predictive models, and it should be noted that this potentially favours larger companies over smaller or new ones due to the volumes of data available.

Lattice Engine, 6sense, InsideSales, Angoss Predictive Analytics,… are some of the companies that offer lead scoring solutions but also are able to develop their own processes to calculate them.

Marketing Automation platforms usually incorporate the creation of scoring models within their services, but there are also CRM applications such as SalesWings for Salesforce that track site activity, integrate them with e-mailing systems (MailChimp, Autopilot, Gmail, Outlook) and qualify leads or customers according to their behavior.


Ad Targeting

ML algorithms can utilise large amounts of past data to establish which ads perform best, for which people and at what stage of the buying process. They can be used to inform, inspire and guide actions based on customer behavior or other business information.

AI-based tools are more efficient than more traditional A/B Test tools. This is because they allow you to simultaneously test a variety of page elements and variations with less traffic than is normally required for an A/B Test to be statistically significant.

Giants like Facebook and Google have stepped up to make their content evermore relevant to each user. All this is done through algorithms based on ML that make rankings of feeds, content, interactions with publications, searches and of course, advertisements.



This phase is characterised by encouraging leads to become customers.


Dynamic Pricing

Through setting dynamic prices, the price of a product or service changes according to supply and demand in real time and allows you to target special offers only to those leads who are likely to be converted. With ML techniques it is possible to build propensity models for leads that require an offer to convert against those who will convert regardless, or customers willing to pay different prices for the same product/service depending on the time of the situation, place, time of year, etc… This will make it possible to increase sales without greatly reducing profit margins.

Tools such as Competera, Wiser (QUadAnalytics), Market Track and Upstream commerce allow you to monitor the prices of catalogue products, in some cases even comparing them with competitors’ products, and see any changes so you can match or improve them.


Web and App Personalisation

Using a propensity model to predict a customer’s stage in the purchasing process allows us to provide the most relevant content on the site, either in an application or through a call center.

Furthermore, from the point of view of user experience (UX), ML can draw upon data generated by the user in different interactions in order to transform and personalise his/her experience when interacting with the product or brand.



Chatbots are beginning to replace features that until now have supported Apps and the rapid growth of their use. This is clearly attributed to the population’s increased use of messaging applications; they can interpret customer queries, provide information, complete orders.

In the industry we find all sorts of coding tools for Bots since they are designed for people with virtually no coding skills. For example, ManyChat and Chatfuel, or Conversable are SaaS services for designing, building and distributing AI-enhanced voice and messaging experiences across multiple platforms, including Facebook Messenger, Twitter, SMS, Amazon Echo, Google Home and many others.



Just as in ad targeting, you can define models that establish what content is most likely to make users return to a site based on the available data.



In the engage phase, we will create a stronger link between customers and the brand.


Predictive Customer Service

As previously mentioned, it is possible to develop models that characterise clients who are most likely to give into offers, so companies can act quickly and activate alerts that allow them to retain them as soon as they are identified.


Marketing Automation

In an environment where the user receives a multitude of impacts through different channels, it is essential that they are carried out with the right message, at the right time and in the place where the audience expects to find it.

Marketing Automation techniques help to do this by promoting lead generation and customer loyalty and affiliation.

To this end, a series of rules are defined which, in the face of customer interaction or non-interaction, are activated and, thanks to machine learning, for example, decide when it is most effective to interact or send a message and which content is most relevant based on knowledge gained from user behaviour.

Pardot, Marketo, IBM Watson Campaign Automation, HubSpot are some of the best known tools that incorporate AI into their solutions.


Emails with Dynamic Content

Predictive analysis can also be useful to establish customers’ tendency in purchasing certain categories, sizes and colors through their behavior and thus can generate personalised emails or newsletters with the most relevant products or information for each customer. They also have the most up-to-date information of the product stock, offers and prices when opening the email.


Artificial Intelligence represents a new era in consumer knowledge for Marketing in which we approach the customer, provide the best purchasing experience and predict their behavior. This is a reality that companies can’t ignore.