Optimising Enterprise AI with Sovereign Routing and Model Selection

Enterprises deploying artificial intelligence agents often run into a severe and unexpected financial bottleneck. While software engineering and operations teams rush to put automated agentic workflows into production, most routine requests are automatically routed to the most expensive flagship models available. Because complex multi-step agent loops routinely retry commands, call external tools, and parse structured output, a single business task can consume up to fifteen times the token volume of a standard chat conversation. Dubai-based technology startup AVELIN AI was built to solve this cost imbalance by providing organisations with total governance over model selection, routing logic, and underlying infrastructure.

The company was founded by Yury Akinin, a doctorally trained entrepreneur who previously co-founded the life sciences technology firm Quantori. Having observed how large organisations consistently overspent on top-tier language models for straightforward administrative queries, Yury Akinin initially set up the business as a specialised research and development lab in 2022. Industry analyses indicate that over half of enterprise requests are routinely sent to systems far more capable and expensive than the underlying task demands. By building an automated decision mechanism into the software stack, the founding team set out to ensure that every request receives precisely the processing tier it requires.

At the heart of the product offering is Decision AI, a trained routing model that evaluates each incoming application prompt or agent task in real time. Simple requests such as rewording notices or basic text edits are automatically directed to lightweight, low-cost models, whereas intricate multi-step reasoning tasks are passed to premium model providers. The software seamlessly integrates with an enterprise's existing model providers and can operate in public cloud environments, on private graphics processing units, or across hybrid infrastructure setups. By intelligently matching workload difficulty to model capability, the architecture can reduce overall token expenditure per task by up to ninety percent.

In addition to intelligent request routing, the business provides a comprehensive sovereign platform specifically designed for highly regulated sectors including financial services, healthcare, and public administration. The platform incorporates nine proprietary models alongside proprietary cross-model fusion technology, enabling strict data ownership and direct governance over critical digital assets. Enterprise finance and technology leaders gain granular operational visibility through dashboard tracking tools that log every decision alongside its underlying justification. This system breaks down computing costs by team, API key, and model, while enforcing strict budget caps to prevent unexpected financial overruns.

To fuel its rapid growth, the business recently secured 3.7 million dollars in pre-seed funding from angel investors to scale its computing power and international footprint. The startup has built strategic technology partnerships with Nvidia Inception, AWS Activate, and the Dubai Future Foundation as it expands across the Middle East, Europe, and North America. As the engineering team prepares to demonstrate its sovereign platform at Web Summit 2026, the company continues to refine its underlying models and platform architecture. This focused approach gives modern enterprises the structural tools needed to run sophisticated, cost-effective artificial intelligence applications without compromising on data sovereignty.

Find out more at avelin.ai