Governed context
Agents work from trusted data, shared business definitions, and the same permission model as the product.
Independent agent-native product lab
We build products where agents do real work.
Flid creates software for a world where people and AI agents operate through the same trusted systems, data, and permissions.
01 / Thesis
Agents should not be bolted on.
Most software was designed exclusively for humans, with an AI interface added afterward. We build from a different premise: agents need structured context, explicit capabilities, permissions, limits, and verifiable results.
Native tools
Actions are deliberate product capabilities with clear contracts—not unrestricted access to the underlying system.
Bounded agency
Limits, evidence, and auditability are part of the architecture, so useful autonomy does not mean uncontrolled autonomy.
02 / Flagship product
LeapView
The agent-native BI platform.
Build dashboards as code, keep analytics in version control, and let people and agents work from the same governed semantic system.
03 / Field work
We build products. Selected field work keeps them honest.
Alongside our products, we work closely with a small number of companies redesigning core systems for agent-native operation.
Making an entire data platform agent-native.
Helping a company redesign how context, interfaces, permissions, and operational workflows fit together when agents become first-class users of the platform.
We do not sell capacity or generic AI implementation. We collaborate where the work advances a reusable product thesis.
04 / The lab
Independent by design.
Flid is an agent-native product lab founded by Jacob Østergaard in Denmark.
We build ambitious, durable products at the intersection of data systems, software, and AI agents—then test those ideas against real operational work.
Discuss a design partnership