Founding a Unified Data & AI Function in a Multinational
Team scope 14×, headcount 2.5×, stakeholder NPS to 92.
Theme Building & Scaling Data/AI Organizations · Also Strategy
In brief
Situation. A multinational operating across 20+ countries had grown its data, BI, and AI capabilities organically, which is a polite way of saying they were scattered across reporting lines (CTO, CFO, Innovation Lab), with different career paths, compensations, and processes.
Complication. On paper, the organisation said “we are data-driven.” In practice, dashboards contradicted each other, AI initiatives competed with reporting for attention, and the C-suite did not know which numbers to trust.
Resolution. Reporting to the CTO, I consolidated three departments into one Technology, Data & AI function. Using a Goals/Strategy/Tactics framework, I aligned C-level vision, built an 18-month action plan, restructured the team into clear topologies, unified budgets and partnership management, and worked with the People function to fix compensation discrepancies and standardise career paths.
Impact. Three senior heads reported directly to me: Technology and Big Data Infrastructure; Business Intelligence and Analytics; and Data Products and AI. The area moved from being a report-provider to becoming the breeding ground of a strategic competency. Every project in this portfolio’s 2022-2025 section happened because this foundation existed.
The longer story
Imagine someone gives you three boats, each crewed by people who slightly dislike each other, and asks you to win a regatta. Your first instinct is to buy faster sails. That is the wrong instinct. The right instinct is to figure out why the crews dislike each other, because no amount of equipment beats a crew that rows together.
My single biggest leadership conviction is that the central data function in any company should aspire to be like electricity: ubiquitous, invisible, taken for granted, and unbelievably important the moment it goes off. You cannot get to that state by buying a better data warehouse.
You get there by doing the unglamorous work of fixing compensation bands, unifying budgets so people stop fighting over scraps, drawing organisational charts that make competing agendas impossible, and giving teams clear KPIs that they themselves measure against.
The technology was the easiest part. The hardest part, and the part that took 18 months, was making three sets of people who had previously reported to three different bosses believe that they now belonged to the same thing. Once they believed it, every subsequent project in this portfolio became dramatically cheaper to deliver.