Scaling AI at Enterprise Level
AI Düsseldorf - Scaling AI at Enterprise Level
Struggling with navigating large scale AI adoption?
June's session brings together senior AI, data, analytics, and transformation practitioners for a focused evening on scaling AI in enterprise environments.
This session looks beyond isolated use cases and explores the operating model around enterprise AI: discovery, governance, reusable architecture, internal enablement, and the path from experimentation to production.
Featured Talks 🗣️
Florian Roscheck Senior Data Scientist, Data & Analytics Strategy Henkel AG & Co. KGaA
From Discovery to Scale: Building Shared Pathways for Enterprise AI at Henkel
What does it really take to scale AI across a 47,000-person organization, beyond pilots, isolated use cases, and hero teams?
At Henkel, we learned that sustainable AI adoption is less about pushing more projects into production and more about designing the right conditions: technical foundations that reduce friction, governance that’s embedded rather than enforced, and communities that make learning and discovery visible across the organization.
In this talk, Florian Roscheck from Henkel’s Data & Analytics Strategy team shares how Henkel is building a shared environment for enterprise AI, anchored by a central AI Engineering Hub and complemented by initiatives like the Henkel Data Club community and an interactive AI Explorer for non-experts. Together, these elements help teams discover what’s possible, reuse proven patterns, and move from experimentation to production without reinventing infrastructure or becoming governance specialists.
Rather than presenting a rigid, one-size-fits-all framework, the talk shows how Henkel connects discovery, learning, and production through shared pathways. Florian will briefly outline the principles behind the AI Explorer – starting from use cases rather than technology, making constraints visible early, and clearly linking discovery to next steps – before showing how this orientation flows into the AI Engineering Hub as the central place for compliant architecture blueprints, reusable patterns, and embedded governance. The Data Club complements this setup by turning individual learnings into shared knowledge. Attendees will leave with practical ideas they can adapt to support responsible, human-centric AI at scale, starting where they are.
Dr. David Zibriczky Director of AI & Data Science at DSV previously DB Schenker
Architecting Enterprise AI Assistants: Local Ownership, Central Orchestration
Title: Architecting Enterprise AI Assistants: Local Ownership, Central Orchestration
AI-based knowledge assistants, often built on RAG, are especially valuable in logistics, where teams constantly need answers about company policies, SOPs and operational guidelines. They are easy to prototype but hard to scale across an enterprise. Going straight for one large central agent is tempting, yet the running cost climbs quickly and the quality is hard to control.
This talk walks through a more strategic, incremental path. We start with a single assistant of limited scope, owned by the team that knows the domain best, and get it right before going wider. This is where proper evaluation matters, an often underestimated aspect that goes beyond usage metrics. From there, various teams build and own their own specialized assistants, covering the knowledge space in a distributed way rather than as one monolith. A light central orchestration layer then routes each query to the right assistant and stays cheap by design, with small instances, limited traffic and controlled loops. We close with practical observations on monitoring usage and cost, plus real examples and lessons learned.
Discussion Focus
Moving beyond pilots and isolated hero teams
Building reusable pathways for enterprise AI
Connecting discovery, learning, and production
Embedding governance into delivery
Designing AI assistants and platforms that scale across domains
Turning individual project learnings into organizational knowledge
Agenda 📅
18:30 — Opening & Introduction
18:40 — Florian Roscheck, Henkel From Discovery to Scale: Building Shared Pathways for Enterprise AI at Henkel
19:20 — Networking Break
19:30 — Dr. David Zibriczky, DSV Architecting Enterprise AI Assistants: Local Ownership, Central Orchestration
20:10 — Closing Session & Networking
Who Should Attend
This session is especially relevant for:
Heads of AI, Data, Analytics, and Digital Transformation
Directors and senior managers responsible for AI adoption
AI platform, data platform, and enterprise architecture teams
Senior data scientists, ML engineers, and AI product owners
Practitioners working on the transition from PoC to production
AI Düsseldorf is a curated practitioner community. Seats are limited to keep the room relevant and discussion-oriented.
Registration 📝
Please register only if the topic is relevant to your current work in AI, data, analytics, platform engineering, governance, or enterprise transformation.