Skip to content
VELA ASSIST

Draft — requires review by a licensed attorney before publication.

AI Usage & Transparency Policy

Last updated November 20, 2026

Plain-English summary

Our agents tell people they're AI, answer from your approved knowledge, and hand off to humans when unsure. AI can still make mistakes, so we build in review and controls.

Who we are

This policy is issued by Vela Assist, a business name operated by its owner, a sole proprietor, 12413 Shenandoah Ln, Kissimmee, FL 34744, United States ("Vela Assist," "we," "us").

1. Disclosure

Customer-facing agents identify themselves as AI assistants by default. Customers can ask for a human at any time.

2. How agents work

Agents use large language models from third-party providers listed on our Subprocessors page, combined with retrieval from your approved knowledge. We configure confidence thresholds, topic restrictions, and handoff rules.

3. Model training

Where providers offer it, we use settings that prevent your data from being used to train their public models. We do not use one client's data to train agents for another client.

4. Limitations

AI can produce errors, outdated information, or unexpected outputs. Agents are not designed for emergencies or to give professional medical, legal, or financial advice.

5. Human oversight

Email automation defaults to human approval. All clients can review conversations, adjust knowledge, and turn agents off at any time.

6. Demonstrations

Website demos may use scripted sample responses and are labeled "Demo mode — sample responses."