Retail Colombia

Retail solution


What problems does artificial intelligence solve?

Artificial intelligence allows to make comparisons and eliminate biases in the behavior of each user at a level that human analysis does not reach, finding differences between each person or process that can result in opportunities to know the ideal moment for the right offer in retail Colombia.

AI applications in industry

AI applications in industry

One of the applications of artificial intelligence in the retail industry is behavioral segmentation on digital platforms. Through the recognition of unique users, whose fingerprint comes from a device that can be accessed by multiple users. This allows:

  • Understand the purchasing behavior of users from the recognition of the categories in which it interacts.
  • Predict a user's next steps from grouping by similar behavior.
  • Recommend products and / or content according to the behavior of other users in the same segment.
  • Detect high value segments.
artificial intelligence Business impact

Business impact

  • 45% of optimized interactions on platforms with a high volume of visits, allowing debugging those events that are not relevant to the monitoring of users and highlighting others that increase their stay on the site. This maximizes the probability of high value interactions such as registrations, purchases, etc.
  • Detection of unique users on the platform based on behavior, taking into account the business hypotheses, which helps to discriminate the contents of each user separately.
Suggested AI and analytics roadmap

Suggested AI and analytics roadmap

  • Guide and align all business stakeholders about the importance and benefits of applying AI.
  • Identify and prioritize the needs of the vertical and the opportunities that generate real value for the company.
  • Recommended: Unify different data sources such as analytics platforms, CRMs, CDPs, transactional, data warehouses. (Creation of Data Lake).
  • Identify and categorize the interactions of users or clients with the channels offered by the company.
  • Pre-process and transform the data.
  • Segment using different machine learning models.
  • Generate descriptive reports on the segments found.
  • Generate specific models of recommendation or prediction for each of the segments.
  • Industrialize the model and integrate it into the business.
Suggested data strategy

Suggested data strategy

A company that performs this segmentation will have an additional view based on data that, together with business knowledge, will allow a better understanding of the company's customers. Additionally, the resources destined to campaigns or recommendations will have a greater use and organizational impact.

Solution: Identification of user behavior by their interactions

Using traces of user interactions on digital channels, the construction of the behavior graph for groups of people is carried out.

A “behavior graph” is called a map of the path of the actions carried out, which makes each graph personalized, for example:

Andrés and Marta bought the same product in the same e-commerce, however, Andrés viewed the product several times before checking out, while Marta arrived directly at checkout through a guideline, for this reason, both have different graphs despite buying the same product.

solutions artificial intelligence retail sector construction graphs

Adequacy of behavioral data

Through an embedding process, graphs are transformed into numerical representations.

This allows the model to put all the graphs, one on top of the other, to compare them. It is about analyzing all the journeys of each user, and identifying how much they resemble each other to group them.

artificial intelligence solutions retail sector construction embedding

Identification of segments based on behavior

Using the unsupervised segmentation model, all the maps or graphs of behavior that have been compared are analyzed and divided into similar groups from retail Colombia.

From these groups, a characterization is obtained based on the forms of interaction.

artificial intelligence solutions retail sector segmentation

Analysis of results and validation of the model

Based on the previous characterization, a validation is carried out with the business on the hypotheses initially proposed and new interest groups are identified.

artificial intelligence solutions retail sector analysis results

Discovery of new opportunities

From the experiences of individuals belonging to the same segment, new insights are identified with the information provided by the data. For example, an analysis of distributions of the populations associated with the segments to discover those with the highest value.

artificial intelligence solutions retail sector discoveries opportunities