Bank Postal: Data is a team game

Data and AI are central components of La Bank Postal’s strategic planning, with backlogs for more than 200 uses. To align data projects and strategic plans, companies set up distributed governance systems.

Matthew Olivier, CDO of Postal Bank

In March 2021, La Bank Postal presented its new 2030 strategic plan. Its ambition: to become the bank of choice for the French, whether they are customers, citizens or employees.

The effectiveness of this plan is based on 14 major projects, including data and artificial intelligence. To run this data component, the company has set up a data office of about thirty people and hired a chief data officer, Matthew Olivier.

Federate a data community of 500 people

The goal of this data division is to federate developments around four axes. The first is data evaluation. For this axis, the data office has a central staff of about ten data scientists

These skills are complemented by hundreds of other data scientists distributed across the various business lines of La Bank Postal. The Data Office is also responsible for managing data governance, creating and maintaining a technical foundation.

For the CDO, Matthew Olivier, his missions are “to support strategic planning, to animate the entire data community representing approximately 500 people, and to prepare for the future in terms of both technical base, business model IA and control”, he said at Hubday Data & IA for Business. Explained at the conference.

Its roadmap is already complete. The chief data officer reported backlogs of 200 to 300 uses of data and artificial intelligence. However, the ambition is not just to increase the number of Data & IA products. It is, in fact, a question of “aligning these projects on the basis of strategic planning so that each project is able to create value.”

These values ​​can be of different types: income, efficiency, customer improvement and employee satisfaction … Projects must be “part of the service of corporate citizenship action, which fulfills the public service mission.”

The distributed rule is being implemented

Matthew Olivier, like his CDO counterpart, wants to involve all of the company’s businesses and subsidiaries in other organizations so that they can “create value.”

Administration is one of the levers implemented by the data office to lead this data component of the transformation of La Bank Postal. “We are in the process of developing a new regime, a framework for all projects, so that from the outset a project carries its scalability and becomes part of a vision of value generation,” the CDO declares.

This regime involves collaboration between the entire company and the project teams, which is multidimensional by nature. “To run a data project, you need to bring together data scientists, data engineers, businesses, developers, all around the same table. Data is a team game, “Matthew Olivier emphasized.

La Bank Postal thus seeks to integrate a data community of 500 people within the organization, of course with “quite different levels of maturity” between business and entity. For this, the company will rely on “Federal Governance” or “Distributed Organization” – as supported by the data mesh system.

In addition to the central team, there is a Data Enhancement Manager in every business of the bank The role of the data office will be in this regime “a framework, setting the playing field and setting the boundaries and rules of the game” for interaction and internal cooperation.

150 employees meet twice a month to talk about data

This legitimacy, the data office develops it specifically through activities aimed at “communication, federating, execution”. In this context, various events are organized, such as the meeting titled “Let’s talk about data”.

For more than 6 months, they gathered 150 people in one hybrid every 15 days. Purpose: To discuss a data project in the presence of project leaders. And these initiatives are therefore multiple. The projects relate to various business and processes of Bank Postal: marketing, anti-fraud, compliance, AML, etc.

“The challenge is to ensure that each project anchors the strategic plan and generates value,” the chief data officer noted. Before any launch, a value is measured and a value axis is defined (NBI increase, cost reduction, improvement of customer satisfaction, etc.).

Complications, underlining Matthew Olivier, have to be reunited several times. “The challenge is to be able to predict the future, to build the present, to be very effective, but also to keep track of the changes.”

The data office is setting up a monitoring center to address these issues. His job will be to help the company prepare for the transition to data and AI. For the same purpose, it already has a research center on the interpretation of algorithms aimed at applying AI law.

The application for this “GDPR BIS”, as the CDO calls it, must be prepared in advance. “Each model must be reproducible, audible and shareable in a simple and transparent way,” he expects. For this reason, but also as part of its public service mission, La Bank equips all data scientists with postal. Purpose: Design an explanatory and ethical AI.

Leave a Comment