Background
EWR is an energy and water provider running critical infrastructure for an entire region. That changes the rules of any AI initiative: data must stay under the operator's control, models cannot route through foreign jurisdictions, operation must be predictable, and every use must be auditable end-to-end. The simple "let's try ChatGPT" path doesn't survive those constraints.
At the same time, the upside is concrete and measurable — multilingual customer service, internal documentation processing, specific workflows around grid operations, billing and supply. EWR was looking for long-term advisory that could bridge both worlds: bringing innovation in without giving up control or resilience.
Solution
schnaq GmbH acts as long-term AI advisory partner to EWR. The engagement combines three layers we treat as one in every advisory mandate:
- Strategic clarity. We assess where AI delivers real value in the existing landscape — and where it would just be an expensive feature. Use cases are prioritised by economics, risk and maturity. We say plainly what's worth doing and what isn't.
- Technical depth. Model selection, hosting model, data flows, security and sovereignty are clarified at the level of detail critical infrastructure demands. Sovereign cloud and sovereign AI are not buzzwords here — they are measurable requirements: physical location, data control, operator liability, key management.
- Day-to-day adoption. We enable advisors, case workers and IT teams at EWR to actually use AI in their daily work — for example through multilingual assistance in customer consultations or document workflows that still run by hand today.
Outcome
The advisory has produced concrete applications that are in production today. AI is used 30+ times per day to overcome language barriers in customer consultations — customers are reliably understood regardless of their native language, with not a single data point leaving EWR's own data centre.
More important than the number is what it stands for: EWR has shown that modern AI is viable under the conditions of critical infrastructure — when architecture, operations and governance are designed alongside the technology. That is where advisory pays off when it can move in both worlds: deep in AI engineering, and sober about the realities of a regulated utility.
Why this fits us
Classic AI consultants tend to bring slides. Classic IT vendors tend to bring tickets. Neither is enough for a utility. We've been building production AI systems for years — model selection, infrastructure, data protection, audit. We are the people you call when the question between "technically clean" and "regulatorily sound" actually needs to be decided. And we stay until it lands in operations.