An Evidence Phase is a short, structured piece of work that tests the assumptions a build decision rests on, and ends in a decision-ready recommendation. It is the engagement through which a team experiences the Before You Build Method. Before You Build is the methodology. An Evidence Phase is how it is delivered.
It moves the five uncertainties to the front, while they are still cheap to answer. Will the people it is built for use it. Does it fit real work, and who acts on the output. What should happen once the system knows something. Who benefits, and are they the ones who pay. Is it a medical device, and what must it be built towards.
Most Evidence Phases can run on synthetic, test, public or appropriately de-identified information, which keeps the data and governance burden proportionate while the proposition, workflow and decision logic are tested. Where representative or real-world data is genuinely needed to answer the decision, that gets said rather than assumed away.
A stop recommendation is a real possible outcome, not a theoretical one. Sacher AI does not sell a technology platform and does not depend on winning the build that follows, which removes the commercial incentive to recommend building regardless of what the evidence says.
There is a quieter benefit that teams tend to notice second. An Evidence Phase gives people something to stand behind when the honest answer is that they do not know yet, which is a much easier thing to say in front of a board when it comes with evidence attached.
Where this is explored in depth