This is the standard everyone at Presaga is held to — employees, contractors, investors, and the AI that helps us build. It lives on this public page for the same reason it lives in our handbook: so anyone can hold us to it.
It is also the standard the product is built to. The rules below are not a mood. They are the reason a Decision Report is shaped the way it is.
This is an operational document, not marketing copy. We version-control it like software — it is never silently rewritten. Every change carries a date, a rationale, and an approval, and it changes only when our understanding of how to better serve people changes — never for style.
We would rather tell you something smaller and true than something larger and impressive. If we can't show you the thing working, we don't say it works — and if the evidence is thin, the report says so instead of sounding sure.
If a feature is still being built, we say it is being built. If we are uncertain, we say we are uncertain. If we discover we were wrong, we correct it publicly. We believe credibility is built by evidence, not adjectives.
Our reports are simulations, not predictions. Our recommendations are designed to inform human judgment, not replace it. We believe our customers own their decisions, their data, and ultimately their future.
Every page on this site is written to these three. So is every report the engine produces.
We model how the people who matter are likely to react. We do not foretell events, and we do not poll real people. The report is built to inform the call. You still make it — and where the evidence is thin, the report says so rather than sounding sure.
A number on this site names where it came from. A finding in a report names the simulated reactions behind it. An example written to show the shape of a report says on its face that it is illustrative. A figure that cannot be traced is removed, not rounded into a sentence that sounds better.
The confidence reading in a report is computed from what the evidence supports. Any reviewer in the process can lower it; nothing can raise it. A contested situation earns a low reading with its reasons printed beside it. That is the system telling you the truth, not failing at its job.
Every automation we build — in the product, and in how we run the company — must reduce human work without reducing human responsibility. An automation that prepares better information for a person has succeeded; one that makes the decision instead of preparing it has failed. We keep a human at every consequential decision, our own included.
Security is not a marketing feature. Privacy is not a sales tactic. Transparency is not something reserved for audits.
When we say we encrypt customer content, we can show you where it happens — not just assert it.
When we describe where and how Presaga runs, we describe what is actually deployed today.
When we describe what PRISM does, we can show it doing it. If we can't demonstrate it, we don't claim it.
We believe admitting uncertainty creates more confidence than pretending certainty. Every customer deserves the same careful reasoning we would expect if the decision affected our own families, communities, or organizations.
Presaga exists to help people think before they act.
Everything else is secondary.
