About ekkodale
ekkodale is a Wiesbaden-based consultancy and software development firm that helps companies in the construction industry navigate digital transformation. The company supports clients across the full project lifecycle, helping them work more efficiently through better processes, data, and custom software.
AI sits at the core of how ekkodale works, both internally and in the products it builds. The team relies on AI daily for code generation, workflow automation, and model evaluation. Beyond internal use, ekkodale has developed hugo: a model-agnostic AI knowledge management platform that connects to any AI provider and is deployed on-premise at each customer site. Because hugo is provider-agnostic, customers need a reliable, GDPR-compliant way to manage their own AI model access. Keeping hugo flexible while also managing ekkodale's own growing AI usage made one thing clear: direct provider accounts were not enough. ekkodale needed a proper AI gateway at the centre of its stack, and one it could confidently recommend to hugo customers as well.
The Challenge
ekkodale had moved quickly on AI adoption, integrating models into internal workflows and building hugo as a product on top of AI infrastructure. But operating across multiple providers, and serving customers with their own compliance requirements, exposed a set of compounding problems:
- Fragmented billing. Every provider (OpenAI, Anthropic, Google, and others) came with its own invoice, dashboard, and credit system. Tracking total AI spend across the team required manual consolidation with no single source of truth.
- GDPR compliance gaps. As a German company handling client data and operating a product deployed at customer sites, knowing exactly where each model was hosted was non-negotiable. With multiple direct provider contracts, documenting data residency was difficult and error-prone.
- No clean path for hugo customers. hugo customers need their own AI model access, and ekkodale needed a solution it could confidently point them to. One that is easy to set up, GDPR-compliant, and works out of the box with hugo's model-agnostic architecture.
- No ability to optimise on value. AI pricing shifts rapidly across providers. Without access to multiple models through a single integration, switching to a provider with a better price-to-performance ratio meant new contracts, new credentials, and new overhead.
- Limited model access. Enterprise platforms like Azure AI Foundry gatekeep the latest models behind approval processes that small companies struggle to navigate, putting cutting-edge capabilities out of reach.
“As a small team running both our own workflows and a customer-facing AI product, we were juggling multiple provider accounts, separate invoices, and no clear picture of where our customer's data was going. The compliance burden alone was a distraction from the actual work.”