How Inco AI works, what it decides, and your rights.
Effective date: April 23, 2026
IncomeOS runs its own AI, built in-house. We call it Inco. Everything that decides something about you — matching, ranking, pricing, income paths, and your Trust Score — is our own code running on our servers.
One exception, stated plainly: when you chat with Inco, a third-party language model (Google Gemini, accessed through the Lovable AI Gateway) is used to understand your question and put the answer into words. It never decides anything on its own — every number, name, and recommendation it repeats comes from our in-house engines. We send it only what your question needs: your profile basics, your listed skills and assets, earnings totals, your trust connections' public skills, and open marketplace listings. Anything you mark Vault is filtered out before the model sees it and never leaves our platform. Model providers do not train on this data. Every one of these reads is written to your AI access log.
You can turn off Inco's use of your data anytime in Settings → Data & Export.
Inco suggests; you decide. Its suggestions are not financial, legal, or tax advice. Your Trust Score is a helpful signal, not a credit rating.
AI can be wrong, incomplete, or biased. Always double-check before acting on anything that affects money, contracts, or safety. If Inco ever says something harmful, please tell us at ai-review@incomeos.com.
Where the EU AI Act applies, we treat Inco as "limited risk" — meaning we tell you when AI is involved. We do not do any of the banned things like government social scoring, biometric grouping of people, or using AI to read emotions for hiring.