Unilever’s finance organization manages large volumes of data across SAP systems, customer transactions and financial reporting processes. While the team invested heavily in building a centralized finance data lake, enabling scalable and self-service analysis across business teams remained a challenge. Many analytical workflows still required manual data preparation and spreadsheet-based analysis before insights could be generated. By adopting Databricks AI/BI and Genie, Unilever is enabling business users to explore financial data conversationally, accelerating decision-making and unlocking analyses that were previously not possible at scale.
Manual workflows limited the value of a modern data foundation
Unilever spent several years building a centralized finance data lake, integrating more than 30 SAP data sources and multiple legacy systems into a unified platform on Databricks. This created a strong foundation for analytics, but the way teams accessed and used that data had not kept pace.
Finance teams still relied on analytical workflows that required combining data from multiple sources, validating information and preparing datasets before deeper analysis could begin. Each new question required additional data preparation and refinement, which extended analysis cycles and limited how quickly teams could explore new scenarios.
This gap became clear in business performance meetings. Teams would arrive with a fixed set of prepared insights, but discussions quickly expanded beyond what had been analyzed. As Cyro Souza, Data & Analytics Sr Manager - Latin America Lead at Unilever, explained, “People would come to meetings with a prepared set of analyses, but discussions would quickly expand into many new questions and scenarios. Exploring those additional perspectives often required further data preparation and iterative analysis.”
At the same time, growing data volumes made analysis even harder. Some use cases — such as customer-level profitability across multiple channels, credit risk exposure by customer segment or detailed drill downs into cost and revenue drivers — required working with datasets whose scale and complexity increasingly challenged traditional desktop-based analytical workflows, evaluate customer risk or analyze financial performance at the level the business needed.
Databricks Genie brings conversational analytics directly to finance teams
To close this gap, Unilever implemented Databricks AI/BI and introduced Genie as a natural language interface for finance users. Business teams access Genie, where curated Genie spaces are built on top of governed data using Unity Catalog.
Each Genie space is designed to reflect business context, with clearly defined datasets, standardized metrics and synonyms aligned to how finance teams think about their data. This allows non-technical users to ask questions in plain language and receive accurate, contextual answers without relying on analysts.
Instead of preparing reports in advance, teams can now explore data interactively, even when working with granular, large-scale datasets that previously exceeded the limits of traditional desktop tools. They can analyze customer behavior, evaluate profitability by customer or product, and investigate financial drivers in real time, adjusting their questions as new insights emerge.
The quality of this experience depends on strong data definitions and governance. As Cyro noted, “Configuration is the heart of Genie. You need clear definitions and structure, but once that is in place, people can just ask questions and move forward.”
This approach simplifies the analytics workflow. What previously required multiple tools, manual preparation and coordination across teams now happens in a single environment, directly on top of trusted data.
Faster decisions and new insights drive measurable business impact
With Genie, Unilever has reduced the time required to answer finance questions from days to minutes. More importantly, teams can explore multiple hypotheses in a single session, which changes how decisions are made.
In the past, testing a single hypothesis could take several days of analysis. If that hypothesis proved incorrect, the process had to start again. Now, teams can quickly iterate and refine their approach. As Cyro explained, “Validating a single hypothesis could require significant analytical effort and multiple iterations. If the initial assumption proved incorrect, teams often needed to restart parts of the process. With Genie, you can test multiple hypotheses in minutes.”
This speed also enables entirely new use cases. Finance teams can now analyze datasets that were previously too large or complex to handle, unlocking deeper insights into customer behavior, credit risk and financial performance. For example, teams can move from periodic credit reviews to continuous analysis, identifying risks earlier and making more informed decisions about customer exposure.
The impact is already visible. A growing group of finance users is actively using Genie today, with plans to expand to a broader set of users in the near term and eventually scale to roughly 1,200 users across the organization.
While some benefits are still emerging, early indicators point to meaningful business value. Improved customer analysis and risk management are expected to drive approximately €3 million in annual cost avoidance. “Many things with Genie are difficult to measure because they were not possible before,” Cyro said. “Now, we can analyze risks and opportunities at a level we simply couldn’t reach.”
By combining a unified data foundation with conversational analytics, Unilever is moving finance from manual reporting toward continuous, data-driven decision-making at scale.