Cognitive factory

The Mutual Insurance Company successfully developed its Data Lab a few years ago, enabling its teams to create use cases with real potential. However, the company aimed higher, aspiring to build an AI Factory with a much broader scope than its existing Data Lab. To scale up this initiative, it collaborated with Onepoint’s teams.

Solution

Onepoint defined a target operational model for the Data Lab, designed to transition gradually into an AI Factory through an experimental phase. This approach ensured agility and allowed the effectiveness of the plan to be measured using KPIs. The next step was to deploy this organization by supporting teams in its adoption, with a focus on developing an agile culture and practice, enhancing customer relationship skills, and implementing global best practices in Data Science.

Outcome

This initiative enabled the client to better address its internal growth challenges by allowing its Data Lab to organically evolve toward the intended AI Factory model.

In its first year of operation, our client aimed to provide cultural access services to hundreds of thousands of young people across France.

To ensure high service quality, the company chose Google Cloud Platform for its scalability capabilities. Onepoint contributed to the redesign and implementation of the Data Architecture, leveraging services such as BigQuery, Cloud SQL, Cloud Dataflow, and Cloud Composer. Beyond this, Onepoint also developed a recommendation engine to enhance user experience and broaden cultural discovery opportunities.

Solution

For the recommendation engine, Onepoint’s Data Scientists designed a system based on Machine Learning algorithms using Python and TensorFlow, deployed with Cloud Run. Onepoint played a role not only in development but also in the industrialization and maintenance of the model, implementing A/B Testing methodologies to optimize performance.

Outcome

This recommendation engine enabled our client to offer a distinctive user experience, facilitating the discovery of cultural activities tailored to individual preferences.

In recent years, the client has launched multiple artificial intelligence initiatives. Despite the technical success of its Proof of Concept (POC) models, the impact of the Data Science teams remained limited due to the lack of deployment and maintenance processes for these experiments. This realization led the client to clearly identify its need: the establishment of an industrialization center to transform its Data Science experiments into operational solutions.

Solution

Onepoint structured the approach around three key pillars:

• Pillar 1: Empower the team responsible for Data Science innovations to independently develop AI products that can be directly used by Regional Banks. This reduces unnecessary stakeholders in the model creation and industrialization process.
• Pillar 2: Provide industrial models with a scalable environment to meet availability needs and handle peak loads efficiently.
• Pillar 3: Establish an operational methodological framework to clarify internal validation processes and interactions, streamlining collaboration between teams.

Outcome

Our client recently publicly announced that it now provides ready-to-use advanced analytics services for its retail banking operations, marking the success of the project.