Coty, one of the world's largest beauty companies with a portfolio spanning fragrance, cosmetics and skincare brands, is in the middle of a large-scale data transformation. The company is consolidating SAP, market data and multicloud sources into a unified ‘Coty Data Cloud’. But KPI dashboards alone couldn't answer the most ad hoc questions executives asked. By deploying Genie, Coty now lets leaders query governed, trusted data in plain language, freeing analysts to focus on scaling the next wave of AI use cases.
Dashboards couldn’t answer the why behind leaders’ questions
Coty unleashes every vision of beauty across one of the largest fragrance, cosmetics and skincare portfolios in the world. Delivering this at scale means getting the right data to decision-makers fast, which is why Coty is building a ‘digital brain’ inside its data cloud. This means consolidating SAP, market data and sources across Azure, AWS and GCP into a hub-and-spoke model in which domain teams own and publish their data products. The goal is a single, governed foundation for analytics and AI, with business analytics teams converging on the same platform rather than operating from disconnected BI stacks.
But even as Coty builds that foundation, the way leaders consume data has been slower to change. For years, insight reached leaders almost exclusively through dashboards built and maintained in Power BI. Those dashboards tracked KPIs well, but they were bound to preconfigured views, and each dashboard had to be maintained over time. The moment an executive needed a slice that wasn't pre-built or wanted to understand what was driving a movement in the numbers, the request routed back through the responsible domain team and waited. "Dashboards can't give you that kind of flexibility," said Shishupal Kumar, Senior Solution Architect at Coty. "Leaders had KPIs, but the moment they needed to know why a number was moving, they had to go back to the team that owns it. That takes at least a day."
Coty also knew that standing up a conversational layer outside the core data platform would create new problems. Every additional tool meant more integration plumbing, shadow data copies and a weaker governance model if data left Unity Catalog. The team wanted conversational BI without the overhead.
Genie delivers governed, instant answers beyond static KPI dashboards
Rather than onboarding another tool, Coty deployed Genie directly inside the Databricks Platform, with every Genie space grounded in curated, domain-owned data products governed by Unity Catalog. "Genie lets us stand up a conversational layer on top of our data while leveraging the governance we already have in Unity Catalog," Shishupal said. "No plumbing to build, no data duplication, no separate security model to maintain."
Five Genie spaces are now in production; each tuned to the semantics and metrics of a specific function:
- The Finance Genie. Handles consolidated profit and loss analysis, net revenue by country and division, capex spends and PL line detail by reporting unit.
- The Production Planning Genie. Serves the supply chain team, surfacing production schedules, capacity utilization and inventory levels across manufacturing sites.
- The Product Intelligence Genie. Allows users to query SKU-level supplier coverage, service performance and production volumes across brands.
- The Mixed Market Modeling Genie. Aids the commercial team in analyzing media contribution to sales and ROI by franchise.
- The Customer Service Genie. Handles customer service inquiries and runs in parallel with an existing ServiceNow agent as Coty evaluates consolidating conversational workloads on Databricks.
To keep pace with business demand, Coty accelerated the work behind each Genie space using Databricks Assistant and Lakeflow Designer. The low-code design experience lets analysts who don't write SQL build new datasets into Genie-ready tables, shortening the onboarding cycle that previously relied on scarce engineering time.
Faster decisions today, a supervisor agent tomorrow
Genie is already changing the pace of executive decision-making at Coty. More than five senior leaders now query Genie directly, surfacing product-line revenue, PL detail, supplier gaps and marketing ROI in the moment rather than waiting for a domain team to produce a follow-up view. Leaders are closer to the data at the point of decision, and demand for new Genie spaces is outpacing the rollout rather than the other way around.
The productivity gains extend across the data organization. Lakeflow Designer makes it easier for analysts who are less comfortable with SQL to build and onboard new datasets into Genie-ready tables, reducing dependence on engineering resources for every new use case. As more teams adopt Genie, the data organization can focus more of its time on expanding into new domains, improving data quality and supporting the next wave of AI initiatives.
Coty plans to deploy a supervisor agent on Agent Bricks that coordinates across domain-specific Genie spaces, enabling enterprise-wide questions spanning finance, supply, product and commercial data. “Databricks gave us a platform where the building blocks are already stitched together," Shishupal said. "Our data is centralized, governed and ready for business teams to build on. That is what makes the digital brain real."