Solv Research
Better evidence for better engineering.
The world of work is accelerating. Our ability to notice, govern, and learn from change has to accelerate with it—without surrendering human judgment.
These are research briefs, not peer-reviewed studies. We distinguish published evidence, our interpretation, and the hypotheses Solv is built to test.
Current briefs
The retro is late by design.
Retrospectives are valuable conversations held on a weak evidence schedule. What changes when capture happens in the moment and reflection remains human?
AI should not get the final word.
Explanations can make people accept an AI recommendation whether or not it is correct. Governance must be architectural, not ceremonial.
Accepted is not improved.
A proposal, an approval, and a completed task are different kinds of evidence. Organizational learning begins when the outcome can answer back.
Editorial standard
Primary evidence first.
We prefer original papers, standards, and research programs over summaries of summaries.
Limits stay visible.
A finding from clinical recall research can inform a design principle; it cannot prove a software-team outcome.
Claims become tests.
Where evidence ends, we state a hypothesis and name the measure that could prove it wrong.
Bring us a hard question
What would make improvement more truthful?
We want the failure cases, governance questions, and field evidence that sharpen this work.
Start a conversation