How we think about applied AI at Eclipse

Every technology company today is expected to have an AI strategy. At Eclipse, our approach is grounded in a simple belief: AI should work quietly, reliably, and practically. We are not interested in hype or buzzwords.

Applied over abstract

We focus on applied AI — systems that automate concrete tasks, reduce manual effort, and improve decision-making. This means we are more likely to build a system that automatically categorises inventory items than we are to build a general-purpose chatbot.

Our research in AI is guided by real product needs. We ask: what repetitive task can we eliminate? What pattern can we detect? What decision can we accelerate?

Small models, big impact

We are particularly interested in small, efficient models that can run on modest hardware. The trend in AI has been toward larger models, but for most practical business applications, a small, well-trained model that runs locally is more useful than a massive model that requires cloud infrastructure.

This is especially important in markets where internet connectivity is not guaranteed. An AI feature that stops working when the network drops is not a feature — it is a frustration.

Ethics by design

We believe AI systems should be transparent, auditable, and under human control. We do not build black-box decision systems. Every AI feature in our products can be explained, overridden, and improved by the people who use it.

This is not just ethical — it is practical. Trustworthy AI is adopted faster and generates more value over time.