Canadian Owned
The principles we follow when we design, build and operate AI voice agents, chatbots and automations for clients.
We design AI systems that identify themselves as AI when they interact with the public. We do not build systems intended to pose as a human or to deceive.
Every deployment defines when a person takes over. Actions with financial, legal, safety or significant customer impact require human approval or are excluded from automation.
We collect and process only the data a task needs, restrict access, and handle data as described in our Privacy Policy. We configure recording and retention per deployment.
Each system has documented topics, allowed actions, blocked actions and fallback behaviour. We do not use AI for legal, medical or financial advice.
We test with real and adversarial inputs before launch, ground answers in approved content where possible and review live outputs. AI can still make mistakes, so handoff and review routes always exist.
Clients remain responsible for the consent, telemarketing, recording, employment and sector rules that apply to their use. We do not build systems for unlawful surveillance, harassment, discrimination or deception.
Our systems may use third-party AI and communications providers. We select providers deliberately and describe relevant data flows to clients during scoping.
If you believe an AI system built by Zrafted behaved unsafely or inappropriately, contact contact@zrafted.com. We review reports and fix issues.
Last reviewed November 2026. This policy describes our practices and is not legal advice. Privacy Policy · Terms & Conditions · Zrafted Inc., 36 Toronto St, Suite 850, Toronto, ON M5C 2C5.
