Modern data leadership: From static reports to intelligent action

Data leadership is often mistaken for data platform management, but it is in no way all that a modern data leader does. Modern data leadership aims at shortening the time between a critical business question and a concrete decision. This is approached by shifting the focus from data collection to data utilization.

The main idea is to stop building dashboards that only tell you what happened (e.g., "revenue is down 20%") and start building data products that explain why and what to do next.

In many organizations, the reality is a polarization between two extremes. On one side, you have the "Excel Hell"—business units exporting data from ERPs to build manual reports that are weeks late. On the other, you have massive "IT Platform" projects that take years to complete. Neither solves the immediate problem: "Where is the money?"

The cycle of mistrust in data

Building trust is hard, but essential for a data-driven culture. When you’re trying to lead with data, you’re essentially fighting against a vicious cycle. If the data provided by systems is poor or hard to access, business users create shadow processes.

These shadow processes lead to data not being updated in the core systems. Consequently, the data quality degrades further, and trust evaporates. The result is an organization that operates on intuition rather than facts, despite having expensive data infrastructure. To fix this, you cannot just "fix the data" or "fix the process" in isolation; you must address them simultaneously.

Moving beyond the dashboard with Knowledge Graphs

Data has its limitations when it sits in silos. A typical approach is to join tables in a BI tool, but this often hits a wall when dealing with complex, many-to-many relationships across CRM, ERP, and project management systems.

This is where modern architecture, such as Knowledge Graphs and AI Agents, comes into play. Instead of flat tables, a Knowledge Graph maps the relationships between entities—people, projects, invoices, and time entries. This allows AI agents to answer complex natural language questions like "Who has worked with Client X on Project Y?" instantly.

This shift—illustrated by companies like Klarna moving towards an "AI Tier"—is about productizing information. It means delivering the context, not just the raw data, to the decision-maker or the AI agent.

Basic principles of modern data leadership

1. Business owns the information needs: The questions must come from the P&L owners, not the data architects.
2. IT productizes the information: The goal is to build reusable assets (Data Products), not just ad-hoc reports.
3. Focus on the operational frontline: The biggest impact comes from optimizing daily work, not just executive reporting.

Action steps for data utilization

1. Identify the single most critical business question (e.g., "Why is margin dropping?").
2. Map the end-to-end process and identify the data gap.
3. Implement a targeted data product (using AI/Graphs if necessary) to solve that specific gap.
4. Measure the impact on the business process.
5. Scale the logic to other areas.

Getting the priorities right

You can usually tell if a data strategy is working by looking at where the investment goes. Traditionally, 80% of the effort goes into collecting and refining data (the platform), and only 20% into utilization.

How about flipping the script and starting with the utilization?

Based on ⟨ a critical business question ⟩, we believe that ⟨ productizing this specific data set ⟩ will enable ⟨ faster decision making ⟩. We will measure this by ⟨ the time saved or revenue recovered ⟩.

For example, your starting point could look like this:

Based on ⟨ the need to understand project profitability in real-time ⟩, we believe that ⟨ unifying time-tracking and invoicing data via a Knowledge Graph ⟩ will allow ⟨ project managers to react to budget overruns immediately ⟩. We will measure this by ⟨ the reduction in non-billable hours ⟩.