Belcorp - Delivering Superior Brand Value and High-Tech Consumer Products Through a Data-Driven Culture and AI Democratization
Daniel Gonzaga, VP - R&D Yancarlo Rivas, Corporate Innovation & Chemical Development Director - R&D Helen Andrade Arcuri, Expert in Bioinformatics - R&D Angie Zubieta, Statistics Analyst - R&D Nicolás Giraldo, Statistics Analyst - R&D Team: Laboratory, Chemicals, Quality and Safety - R&D Venkat Gopalan, VP – TI Alejandro Palomino Samaniego, Corporate Analytics COE Director - TI Jose Israel Rico Peña, Corporate Data Architecture Director - TI David Narváez, Data Product Manager - TI Diego Chirinos, SRE – TI
BELCORP is a multinational corporation with more than 50 years of experience in the production and commercialization of beauty and personal care products under the B2B, B2C, and B2B2C schemes, through its sales channels such as Direct Sales, eCommerce, and Retail. With a presence in 12 countries in Latin America (Peru, Colombia, Chile, Mexico, Ecuador, Costa Rica, El Salvador, Guatemala, Panama, Bolivia, the Dominican Republic, and Puerto Rico) and the United States, we have a purpose that transcends economic results – we promote beauty to achieve personal fulfillment, and to inspire each person to give their best to achieve the extraordinary.
Value at Scale
The VP of R&D in recent years has been looking for technologies to carry out more and more digital transformation within its research and product development laboratories. The aim is to innovate, be productive, and be effective by simplifying the innovation process through advanced experimental techniques and by bringing a data-driven culture and AI democratization across the VP.
The new experimental technologies have been generating a large amount of data, and the storage was done in a dispersed way in different formats, and that did not bring the relationship between them. There was difficulty in accessing and locating the data, as well as in having the generated data combined to ensure the generation of new knowledge and added value within our VP of R&D.
The development of an intelligent services platform that combined the best processing technology with artificial intelligence and advanced analytics technologies was essential to close the entire cycle of the digital transformation that the VP of R&D was pursuing. From the platform, we expect to accelerate growth, productivity, and efficiency to deliver superior consumer brand value and bring the knowledge and optimization of its processes to the business and operations, thus enhancing our differential in innovating with high-tech products that are safer and in accordance with the needs of our consumers.
Through collaboration between different areas of knowledge within our VP R&D and VP IT, we worked transversally on complex prediction and simulation models of product behavior, the interaction of products with the skin of the body and face, and the anticipation of possible security risks in our products. We guarantee to have applications 100% focused on solving problems that go from the conception and design of the product, through all the parts of efficiency and safety, to the monitoring of its behavior in the market.
That said, we felt the lack of a partner that would give us the opportunity to work with the best servers in technologies such as neural networks, computer vision, NLP, fuzzy logic, AI, and advanced analytics, providing agility to our data modeling and training processes. It was at that moment that we found in Dataiku the technology differential that we needed to develop the platform.
Through Dataiku, our platform now has a robust and transparent architecture. We scraped and compiled the data using data mining technologies to compile the data needed for our models from 23 international public benchmark external databases versus the data generated internally since 2016.
Dataiku was the key player in developing the Innovation Lab, a platform developed in-house that hosts different AI and ML use cases powered by Dataiku. It's capable of increasing the efficiency of researchers and formulators by 30%, achieving predictive results with an accuracy greater than 95% that was previously achieved, and reducing the time to market by 10-20%.
Business Area: Other - Research and Development Cosmetic Products
Use Case Stage: In Production
The platform has a roadmap of 22 projects distributed in five main pillars until 2024 that aims to bring the following benefits:
Higher quality and better customer experience (consumer ultra-personalization).
Improve performance (10-30%) and claims support into formulation while keeping the costs down (10-20%).
Improved productivity, efficiency, flexibility, and agility in areas of laboratory (until 60%).
Consumer risk assessment reduction and guarantee of the safety of products (80%).
Centralization, traceability, quick access, and security of access to information eliminate redundant processes and optimize the time of researchers (100%).
Reduction in time-to-delivery (10-20%).
Value Brought by Dataiku:
Speed and agility through increased team efficiency, providing new artificial intelligence and advanced analytics capabilities that take the R&D team to the next level of the analytics maturity stage. This hugely enhances our tech stack efficiency.
The projects we work on within the platform require a customization of the solution according to the question to be answered in each area of the business, and with the option to create custom plugins and components of metadata, we have a differential that has brought an increase in efficiency, quality, and granularity within our projects. Furthermore, the ability to compare small to large-scale datasets from various public databases with different formats (e.g., xml. txt, csv, json) and internally generated unstructured data quickly and transparently brought savings in time and increased efficiency in the development of AI models. This increased the range of information and diversity of applications and brought knowledge and anticipation of trends, as well as unique insights to our users.
Furthermore, the image processing capacity is a very important capability that brings us tremendous value. This is because we have some applications that use image capture to analyze efficiency and quality in products that are under development, as well as to be able to verify what our gaps are to offer products that meet the needs of our consumers and gain market share with new ones. The possibility to cross this using NLP through the responses to the questionnaires makes our applications a technology differentiator and a new way of proposing new products in our portfolio, thus delighting our consumers.