Claude Enterprise Summit

Moderatorinnen der Veranstaltung:

• Gloria Wagner, Senior Consultant, valantic
• Victoria Sophie Sturm, Consultant, valantic

Lukasz Chlipala, GTM Partnerships DACH, Anthropic

AI is improving so quickly that the gap between early movers and everyone else keeps getting wider. The talk shows how AI changed software engineering in 2025, where agents now take on whole features and projects, and why operations, finance, legal and marketing are next. It includes a live look at Claude Cowork and customer examples such as Novo Nordisk. Starting from the finding that 95% of AI pilots show no P&L impact, it sets out five levers that separate companies getting real value from those stuck in pilots.

Holger von Daniels, Founder & CEO, valantic

Arik Ott, Head of Digital Innovation, Siemens Energy

Running AI strategy for a 30,000-person industrial business means more is happening than any leader can hold in their head. The speaker therefore built a work brain: a Claude-maintained knowledge layer, inspired by Andrej Karpathy’s LLM Wiki, that works as a personal agentic coworker with the context it needs to be useful. The talk shows how this personal problem mirrors the organization’s: knowledge sits in file shares without usable metadata, access rights break when agents consume data, and experience leaves with retiring colleagues. The case it makes: context, not model capability, is the real adoption question.

Dr. Ursula Feindor-Schmidt, Senior Partner, LAUSEN 

Rolling out an AI assistant across the company raises legal questions that need answers early. The session covers what the EU AI Act means for operators, who owns AI-generated results and who is liable when the AI makes mistakes. It also shows what GDPR-compliant use looks like in practice, what to look for in enterprise license agreements, and when to involve the works council and data protection officers.

Yavuz Bogazci, Managing Director & Co-Lead Telco Practice, valantic

Inference prices have fallen sharply, yet many AI bills keep growing. The reason is that volume, model choice, context size and retries drive the invoice, while the price per token barely matters. Yavuz shows where AI budgets leak in practice, from agentic workflows to flagship models used as the default. Expect a practical view of what to measure, how to match the model to the task, and why guardrails in the request path control costs better than budget alerts.

Customer & User Panel

Elvira Wagner, Data Intelligence Platforms Strategy & Enablement Lead, Raiffeisen Bank International

For years, modernization projects were constrained by documentation gaps, technical debt and limited engineering capacity. AI coding agents are changing that equation. The talk explores how organizations can use AI to accelerate migration, improve governance, generate platform knowledge and unlock productivity gains across analytics and data engineering teams, without compromising control and compliance.

Mikkel Astrup Jensen, Engineer, AI and Digital Transformation, NIRAS

At NIRAS, a Danish engineering consultancy, the environment business moved from Copilot via ChatGPT to Claude. The main lesson: adoption works when it stays simple. Licenses go to people who are keen to use AI and can check its output in their own field, supported by a quick-start onboarding and a user group for sharing experience. Expert know-how is packaged as shared skills, for example to produce parts of nature assessments from a single shapefile upload. The talk also covers how usage is tracked and what it takes to bring Claude from pilot to everyday use.

[Speaker and abstract to follow]

[Speaker and abstract to follow]

[Speaker and abstract to follow]

Live-Demo

• Christoph Pschorn, Partner, valantic
• Sebastian Drickl, Partner, valantic