Requesty
Case Study|Enterprise AI Gateway

How LOUPZ Centralized Multi-Model AI Access for Its Entire Team with European Data Residency

LOUPZ GmbH & Co. KG · Construction & Real Estate · Digital Transformation

1 gateway
A single access point for the whole team
~10 apps
A growing AI product landscape
5-15 models
Flexible multi-model usage
+200%
Faster rollout of new services

With Requesty we were able to significantly reduce our administrative effort and accelerate the rollout of new services by 200%.

Marcel Karas
Marcel Karas
IT Officer (AI Team), LOUPZ

About LOUPZ

LOUPZ is a pioneer of digital transformation in the construction and real estate industry. Through a holistic delivery cycle spanning six service areas, from (general) planning, BIM & project management and construction management to consulting, architecture & design, and asset information services, its team of around 50 people accompanies building projects from the very first idea to final realization. From its headquarters in Essen, LOUPZ is currently expanding to further locations across Germany, with the majority of its team working remotely from anywhere.

As a digitally leading company, LOUPZ uses AI systematically rather than selectively. All employees use internal AI chats with connected data sources via the Model Context Protocol (MCP). On top of this, a growing number of in-house applications draw on the strengths of different models depending on the task, along with increasingly multi-layered agents that support concrete domain processes, such as generating building sketches, creating and editing project steps, and supporting individual steps in project execution.

The Challenge

As AI adoption grew rapidly across the company, the underlying infrastructure became more complex faster than it could be governed centrally:

  • Manual model integration. Every new AI model had to be connected individually and by hand, slowing down the development of new applications.
  • Scattered credit balances. Separate credit balances had to be set up and continuously monitored for each provider, a significant administrative burden.
  • Fragmented billing. Multiple separate invoicing paths made clean, centralized cost control difficult.
  • Data protection & data flows. Especially for sensitive project data, including for public-sector clients, it was essential to keep data in Europe and avoid sending it to the US unnecessarily.

Before Requesty, every model meant another integration, another credit balance, and another invoice to watch. Bringing all of that under one roof was exactly what we needed.

Marcel Karas
Marcel Karas
IT Officer (AI Team), LOUPZ

Why Requesty

For LOUPZ, one criterion stood out above all, European data residency, combined with true centralization.

EU routing via a dedicated endpoint

Requesty routes traffic through a European endpoint, so no data goes to the US unless it is strictly required.

One access point for many models

Through a single interface, LOUPZ accesses all common models, from Anthropic (Claude/Opus) to Chinese models.

Central billing instead of many balances

Many separate provider credit balances are replaced by a single, managed billing setup.

Unified cost tracking

A central view of AI spend across all applications and the entire team for the first time.

The Results

1 gateway · A single access point for the whole team

All employees access AI models through one EU-based gateway, no more manual, one-by-one integrations.

~10 apps · A growing AI product landscape

Around ten in-house applications run in production through the gateway, with the number steadily rising.

5-15 models · Flexible multi-model usage

Depending on the task, LOUPZ chooses from 5 to 15 active models, without provider lock-in.

+200% · Faster rollout of new services

New services can be rolled out significantly faster and administrative effort has dropped sharply. Tangible cost savings are currently being captured systematically via KPIs.

In Their Words

Requesty is the central backbone of our AI usage: European, model-independent, and centrally manageable.

Marcel Karas
Marcel Karas
IT Officer (AI Team), LOUPZ