AI Usage Policy
Last updated: February 17, 2026
Models we use
We route to the best model for each task: OpenAI GPT-5.x, Anthropic Claude 4.x, Google Gemini 3.x — via the Emergent Universal LLM proxy.
We never train our own models on your data. Provider configurations are set to NOT use customer data for training.
Transparency
Every AI output is clearly labeled as AI-generated.
When a number is an estimate, the model marks it as such.
We do not generate medical, legal, or licensed financial advice.
Safety & abuse prevention
We monitor for misuse (spam, fraud, harassment) and may suspend accounts violating our Terms.
We do not allow the Service to be used to generate disinformation, harassment content, or illegal material.
Bias & fairness
LLMs reflect biases in their training data. We test for problematic outputs but cannot eliminate all bias. If you encounter biased output, please report it via Contact.
Your control
You can opt out of using personalized data (your Finance Digital Twin) at any time by deleting it from Dashboard.