We are industry-agnostic, but not opportunity-agnostic.
Grey Haven does not pick sectors because they sound fashionable. We look for places where expertise can become durable product advantage. Sometimes that is healthcare or education. Sometimes it is oil and gas, logistics, public infrastructure, insurance, agriculture, or a specialized services market most software investors ignore.
The industry label is less important than what becomes hard to copy.
What makes an advantage durable
A defensible AI product usually has more than model access. It has workflow depth, proprietary context, distribution through trusted relationships, evaluation data, user feedback, operational knowledge, and a product surface that gets better as experts use it.
Those advantages compound. They are difficult for a generic tool to copy because they are created through daily use with the people closest to the work.
What does not count
A wrapper around a frontier model is not enough. A dataset no one uses is not enough. A broad market map is not enough. Even a talented engineering team is not enough if the product never touches a workflow that matters.
We want opportunities where the first wedge can lead to compounding context. If use does not deepen the advantage, the project will become a services engagement with better branding.
The question we ask
Before we commit, we ask: if this product works, what becomes harder for the next team to copy?
If the answer is only "we shipped first," we keep looking. If the answer is expert-shaped context, production feedback, customer trust, and a workflow that improves through use, we lean in.
