AI Accountability
Our AI recommends — it doesn't decide.
Every product we build that uses AI is designed around one principle: you are always in control. AI surfaces options, flags risks, and makes recommendations. The decision is yours. We will never design a product where AI forecloses a choice without telling you, or presents a single option as if no alternatives exist.
When Purple Mainspring has a financial relationship with an option we're recommending — a carrier, a merchant, a service provider — that relationship is disclosed before you act. We also show the reasoning. You have what you need to agree, disagree, or ask more. We earn trust by being transparent about how we earn revenue, not by hiding it.
AI-generated content is labeled at the point you're reading it — not buried in a help article or a terms update three clicks away. The same standard applies to human expert input. When a licensed professional, domain advisor, or credentialed specialist contributed to what you're seeing, that's noted, and where possible, who they are.
We tell you when the AI doesn't know. A model that presents uncertainty as confidence is more dangerous than one that says nothing. When our systems lack sufficient information, when a question is genuinely ambiguous, or when the right answer is "speak to a professional," we say that — directly, in the moment you need it, not after you've already acted.
No AI output is final. You can dismiss any recommendation, inspect the reasoning behind it, and choose differently without penalty. The system remembers your preference.