Agentic development and AI Advisory

We connect humans and software with AI

Deep technical expertise meets AI innovation. We help organisations modernise, build, and scale software that drives real business value using AI agents.

What We Do

Three areas, one approach

End-to-end AI solutions from strategy to production.

By the numbers

umage.ai by the numbers

26 hours From idea to released open-source Windows app
40 models LLM models run and tested locally
12 agents Concurrent coding agents on a single machine
65+ years Combined industry experience

A senior team built for what's next

We combine decades of consulting experience with a conviction that AI fundamentally changes how software should be built.

Led by Allan Thræn, Martin Ottosen, and Morten Aslo-Østergaard.

AI-first

Intelligent automation as the default.

Transparent

Full visibility, honest trade-offs.

Pragmatic

The right tool, not the trendiest.

Your code

No lock-in, no black boxes.

umage

[ˈuˌmæːjə]

Danish for giving a damn.

Most people do the work. We do the umage. There is no English word for it — the closest you get is "taking pains," but umage has none of that suffering. It is the quiet commitment to doing something properly, because the result matters.

Frequently asked

Questions we hear often

What is AI Sovereignty?

AI Sovereignty means running AI on infrastructure you control — using open-weight models and open-source tooling rather than third-party cloud APIs. It lets enterprises put AI to work on sensitive data and business-critical workflows without sending that data outside their environment. umage.ai specialises in designing, deploying, and operating locally-hosted AI.

What is Agentic Development?

Agentic Development is software delivery where AI coding agents do a meaningful share of the work — writing, reviewing, refactoring, and testing code under engineer direction. It compresses idea-to-production cycles dramatically. We run up to 12 concurrent coding agents on a single engineer's machine and have taken production-ready open-source apps from idea to release in 26 hours.

Do you work with open-weight or locally-hosted large language models?

Yes. Running LLMs locally is a core part of what we do. We have run and tested 40+ open-weight models and routinely help clients evaluate which model fits a given workload, how to deploy it on-premise or in a sovereign cloud, and how to integrate it with existing systems through RAG and agent architectures.

Where is umage.ai based, and who do you work with?

umage.ai is based in Virum, Denmark (Teknikerbyen 5, 2830 Virum). We primarily serve Danish and other Nordic enterprises, working on-site and remotely. Reach us at hello@umage.ai or +45 7071 3333.

Who leads umage.ai?

umage.ai was founded by three partners: Allan Thræn (Optimizely MVP with 25+ years in software architecture), Martin Ottosen (former Director of Product Management at Optimizely and Product Architect at Sitecore), and Morten Aslo-Østergaard (.NET architect with deep AI-sovereignty experience and a DTU engineering degree).

Ready to build with AI?

Tell us about your project

We'll show you how AI-driven development can work for your organisation.