Use Cases

IoT

PROBLEM

A land transportation company, had significant losses in the revenues collected at the time of centralization of profits from each of its drivers in its different fleets. Additionally, it did not have optimal control of its fleets.

SOLUTION

A management and monitoring solution was generated for its units. With this, the company has a "real time" management and monitoring where it can view at all times (GPS, unit battery level, distance in km, route tracking, other KPI's), additionally with the use of Some sensors were obtained to record the passengers who get on and off the unit, thanks to this data predictive analysis techniques were incorporated to estimate the amount that each driver has to be reporting.

Scoring

PROBLEM

One lending organization was concerned about the increase in its past-due portfolio, since they were about to expand their service and the loans they granted were made under the supervision and in-depth investigation of each of their clients, so this was very manual and time-consuming.

SOLUTION

A scoring solution was generated which allowed them to have a considerable level of risk acceptance to keep their overdue portfolio in a healthy percentage and at the same time provide an automatic, reliable and fast service for their clients.

MarkDown

PROBLEM

A company was selling its products at a general level, that is, without considering geographical location, socioeconomic level of the population, elasticity of demand...

Therefore, it had an area of opportunity at the time of these sales.

SOLUTION

A MarkDown solution was generated, which generates a proposal for discounts by geographical location of their stores, thus enhancing the depth of the reduction of the articles, favoring the desired business logic, either give a higher sales speed to their products or take care of the profit margin.

Forecast

PROBLEM

A manufacturing company requested its material according to the demand it expected to have throughout each season of the year, this stock estimate was made with tools with low potential for this task, consequently these forecasts had a wide margin of error this incurring in issues of overstock or lost sales.

SOLUTION

A Forecast solution was generated, which allows them to have a much more accurate forecast of their stock, and above all, fast and automated. In this way, they were able to reduce the costs they had due to over-stock issues and generate profit by avoiding lost sales.

Anomaly Detection

CRM

PROBLEM

A retail store had all the purchase transactions that their customers made day by day (Amount of purchase, Date, Sku's acquired, Method of payment...). The problem of this company was the ignorance of the exploitation of the data, since they only had a descriptive use..

SOLUTION

A customized CRM solution was developed, that is to say, a value was given to each group of customers according to their purchasing behavior. In this way, a recommendation model was generated to promote cross-selling, and finally, demographic information was collected from the customer in order to have a clearer profile of each group of customers and at the same time determine the product basket of each one of these groups.

Churn

PROBLEM

A telecommunications company had a big problem in the part of the retention of its clients, since it was increasing considerably and at the moment of doing the retention management they granted enough benefits to avoid the abandonment of the client, causing this major financial loss for the company.

SOLUTION

A Churn solution was developed that helps them to know the probability of abandonment of each of their clients in advance, so they focus only on those who are really willing to leave the company, thus reducing the expenditure of resources allocated to this area. On the other hand, the profile of the customers with the highest probability of abandonment was identified in order to create a product or retention campaign more in line with this type of customer.

Scoring

PROBLEM

SOLUTION

Retail

PROBLEM

A company was selling its products at a general level, that is, without considering geographical location, socioeconomic level of the population, elasticity of demand...

Therefore, it had an area of opportunity at the time of these sales.

SOLUTION

Forecast

PROBLEM

SOLUTION

Anomaly Detection

PROBLEM

A company allowed its branches to carry out a certain number of transactions within a certain time limit with a certain amount of money. In case this was not complied with, alerts were launched to review the transactions because they were not complying with the rules that the company had authorized. This prevented them from having precise supervision.

SOLUTION

An Anomaly Detection solution was generated, with which it was possible to detect the cases that were not complying with the rules, reducing the number of false alerts, since only those that were more likely to be true were left. Additionally, alerts were generated for those transactions that were very close to non-compliance with these rules.

Geostatistics

PROBLEM

A company at the time of opening a store assigned one of its existing stores as a reference store, this so that when you open the new store, have the same stock, staff, based on which the new store was expected to have the same level of sales as the reference store. This was limited to the business knowledge of the person opening the stores.

SOLUTION

A Geostatistics solution was generated, which by means of merely external attributes such as NSE, Climate, Type of Commerce, Type of Population, among others, delivers a recommendation proposal, giving the user 3 options of possible reference stores for the new store, estimating that it will have a similar sales behavior as the recommended stores.

Text Mining

PROBLEM

A customer service company wanted to improve its customer satisfaction, as it was receiving many complaints from its customers through its different means of contact. The great difficulty is that the way this company was mitigating this problem was manual and very unreliable.

SOLUTION

A Text Mining solution was generated, which allows a text analysis of the negative and positive comments of their customers in their different contact channels, then a quadrant was generated to give a level of priority and impact of each of these, also generated an indicator to estimate customer satisfaction.