Data & AI strategy scoping: target team, use case portfolio and 12-month roadmap

The digital department wanted to create a cross-functional Data & AI team serving all of its products. The ambition of this team had to be defined for the next 12 months, while the digital organization was undergoing transformation.

Before recruiting and launching the first projects, three questions needed answers.

• Ambition. Define what the team must bring to the group's products.
• Organization. Size the team and define its composition.
• Starting point. Choose the first use cases among numerous needs scattered across the business units.

The need: a shared strategy, anchored in the actual expectations of the business units, and directly actionable.

We built the Data & AI strategy starting from the needs expressed by the business units. Three principles guided the work.

• Start from the field. The use case portfolio comes from interviews with the product owners and product managers of the partner business units, not from a generic list.
• Build buy-in. In an organization undergoing transformation, each business unit finds its needs in the roadmap.
• Plan for reuse from the start. A common data model makes it possible to capitalize from one use case to the next.

• Vision. Definition of the role of the Data & AI team and of its contribution to the digital department's products.
• Use case portfolio. Inventory of existing data use cases through interviews with the product owners and product managers of the partner business units.
• Target team. Definition of the size and composition of the team, matching profiles to the use cases to serve.
• Prioritization. Ranking of the collected use cases.
• 12-month roadmap. Sequencing of the prioritized use cases over the coming year.
• Final presentation. Presentation of the results to the digital leadership and the Data leadership.

A team of two experts, a strategy consultant and a data science and data engineering expert, over a 45-day engagement. Tech stack: AWS.

Leadership has a complete strategy report: the vision, the target composition of the team, the needs, the existing assets and the use cases prioritized on a 12-month roadmap.

Of the 38 use cases identified across six business units, 12 were prioritized on the roadmap. Six months after the final presentation, four were launched, two of them in production, carried by a target team sized during the engagement, whose first hires were already in place.

Two effects extend the engagement.

• A common data model. The move to the Haystack data model, an open standard, brings more uniformity and reuse from one use case to the next.
• Better-targeted hiring. Team profiles are matched to the use cases, which guides recruitment and training.

• A strategy that comes from the business. Use cases come from the product owners and product managers, which eases their buy-in.
• Organization designed with the strategy. The target team is sized to execute the roadmap, not the other way around.
• Profiles aligned with use cases. Recruitment and training start from actual needs.
• Reuse as a principle. A standard data model avoids starting from scratch on every project.

This approach addresses four situations.

• You are creating or restructuring a Data & AI team and need to define its ambition.
• Business requests are numerous and you need to choose where to start.
• You want a target organization before recruiting.
• You want to capitalize on your data from one project to the next.

We start from your business units and your existing data to build a roadmap your teams can execute the very next day.