Overcoming dashboard limitations to deliver timely insights
AkzoNobel is a paints and coatings company operating in more than 150 countries and markets with over 30,000 employees. Its coatings show up everywhere from the Sydney Harbour Bridge to the Mars rover.
Built through mergers and acquisitions, the company ran more than 50 ERP systems with separate analytics layered on top and no standard processes or KPI definitions, all held together with Excel. An enterprise-wide answer could take months or even half a year, and by then leaders had often found another way or made a guess. The response was more dashboards, until AkzoNobel had more dashboards than employees and still struggled to answer business questions.
Centralizing finance data for instant access with Genie Agents
AkzoNobel fixed the foundation first. It centralized IT, cut more than 50 ERP systems to four, consolidated finance data into a single instance with a harmonized general ledger and moved to a lakehouse on the Azure Databricks Data + AI Platform, chosen for its openness and the way it brings AI and analytics together.
Building a solid foundation for finance data
Over two and a half years, the team brought its SAP-based finance data into Databricks and built models for the balance sheet, P&L and margin management.
Introducing Genie Agents for natural language queries
Genie Agents now sit on top, allowing anyone, from the CIO to a business analyst, to ask a question in natural language and get an answer.
Driving Genie adoption through targeted use cases
The business didn’t see the need at first. Early demos elicited the same reaction: "Why use Genie when I have a dashboard?" So the team worked with business analysts to gather the hard finance questions the CIO asks and built a demo that answered them in seconds.
Then they went off-script and asked the CIO for his toughest question. Genie answered it within seconds. Today the CIO is AkzoNobel’s top Genie user and uses it every day, even from his phone on his commute.
The same approach won over a skeptical sales director within a few weeks, after Genie surfaced churning customers he immediately recognized. Teams can now ask strategic questions and the small ones no analyst would be asked to chase, like how much the company spends on laptop bags.
Ensuring trust and reliability in Genie Agents
To make Genie as trusted as the dashboards people already rely on, AkzoNobel set clear standards for every Genie Agent.
- Every Genie Agent has an owner, an SLA and, ideally, a certification stamp.
- Business analysts are brought into Genie rather than replaced, applying their knowledge of calculations and business requirements.
When Genie returned an EBITDA figure that didn’t match the Power BI data, the team provided a screenshot of the dashboard and asked about the discrepancy. Genie explained it, finance confirmed the right number and the Genie Agent was corrected within five minutes.
Next, the team is exploring proactive, contextual and agentic experiences.