Frequently Asked Questions (FAQs)
Q: When are you collecting my data? Do you track any of my interactions outside A2?
A: No, we only collect interactions that occur within the A2 site. This includes your typing metadata during conversations with A2, which are analyzed to assess your neurostate (e.g., cognitive load, motor activity, behavioral state). We do not capture everything you type or do on your device — only what you actively enter in the A2 site.
Additionally, we do not store personal data. Your interactions are processed in real-time to help A2 adapt to your needs, but the content of your inputs is not retained beyond the session. Since your prompts are sent to foundational LLMs (e.g., GPT) to generate responses, and these models have their own data policies, we advise avoiding the input of sensitive information.
Q: Why does A2 place so much emphasis on neuroperformance metrics in basic prompts?
A: For demo purposes, we wanted to highlight the adaptive nature of the A2. The system leverages neuroperformance metrics to tailor responses in real-time. In this early release, we’ve intentionally emphasized the impact of these metrics to demonstrate how A2 adapts to cognitive, motor, and behavioral signals. In real-world applications, this adaptation occurs more gradually as the system collects data over multiple sessions. We are actively working on improving the orchestration, so it will feel more natural and less noticeable in basic interactions in future versions.
Q: Why am I not getting a response from A2?
A: There may be occasional delays in A2’s responses due to high demand on the servers or reaching connection limits with the underlying LLM. In this demo, we use shared endpoints and non-dedicated resources, which can cause latency, especially during peak times. If this happens, try logging out and switching the LLM selection, or simply come back and try again later. We're actively working on improvements, including dedicated endpoints, to reduce these delays in future updates.