Mergers and acquisitions leave a familiar mess behind: different BI systems, different technologies, and thousands of reports scattered across tools nobody can fully navigate. In our latest webinar, Metric Insights CEO Marius Moscovici sat down with Victor Gabriel Castillo, Advanced Analytics Lead at Coca-Cola Europacific Partners, to walk through how the world's largest Coca-Cola bottler brought its reporting together into a single, AI-powered analytics portal.
The Problem: Too Many Reports, Nowhere to Look
As the biggest bottler in the world, CCEP grew through decades of mergers, and each one brought its own BI stack. The result was thousands of reports across Power BI, Tableau, and more, with no reliable way to find the right one. Victor cited a McKinsey study estimating that employees spend around two hours a day just searching for information. People coped by hoarding bookmarks, forwarding links over email, and keeping shared "pockets" of report links, none of which scale. Worse, if you don't know a report exists, you can't ask for it, so you rebuild it, and duplication spreads.
From In-House Experiments to a Real Portal
CCEP first tried building a catalog in-house using SharePoint. It technically worked, but a list of links is hard to maintain and doesn't scale for a company of their size. That led them to implement a dedicated analytics catalog, and they are now a year into a rollout that has been a success so far.
What the Portal Delivers
Victor demoed several capabilities from their live (lightly anonymized) system:
Search and find: Users search across rich metadata, not just report titles, including business descriptions and KPIs, then open reports embedded directly in the portal alongside certification status and definitions.
Discover what you can't yet see: A discoverability mode surfaces reports users don't have access to, showing metadata and a contact, plus a direct link into the identity management system (SailPoint) where the right access role is pre-selected.
A conversational AI assistant: An LLM answers questions in natural language, returning KPI definitions in the organization's own context with links to the data catalog, and even handling surrounding BI questions like licensing and how to request a new report.
The right reports for the right audience: Content is curated by business function and country, so a commercial user in Germany lands on the reports most relevant to them from a single shared link.
Lessons Worth Stealing
A few themes stood out from Victor's experience:
Curate deliberately: Rather than dumping everything in, CCEP set a goal that 80% of user activity should be findable in the portal, which meant roughly 300 to 400 well-chosen reports out of tens of thousands.
Start small and control standardization: A trained set of report curators keeps metadata consistent. Expand the group slowly so quality doesn't slip.
Guard against hallucinations: The assistant clearly flags when answers may be AI-generated and cites whether information comes from general knowledge or the organization's own indexed sources.
Roll out market by market: Tailored communications for each business function drove adoption far better than a single company-wide email.
Watch the Full Conversation
If you're planning a BI Portal, or wondering how to layer AI onto one responsibly, Victor's walkthrough is full of practical detail. He and Marius cover metadata strategy, security and access, governance, and how to plan a rollout that actually sticks.
Watch the webinar to see the demos and hear the full story.
Metric Insights uses cookies and similar technologies to enhance site navigation, analyze site usage, personalize content and ads, and as further described in our Privacy Policy. Click “Accept All” to enable all cookies or “Reject All” to reject cookies.
