AI policy

Where we use AI

• Aspectora (in development): an AI site assistant that suggests layouts and copy. You accept, edit or reject each suggestion; nothing is published without a person's approval.
• Aspectobots (alpha), the bot module: Discord bots you build and own. Bot owners decide which features use AI, if any.
• Custom AI agents we build for clients: drafting, triage and routine operations, with confirmation queues and audit logs.
• Our own work: we use AI tools to help write code, documents and site content. A person reviews and is responsible for everything we publish.

How our AI features behave

• People confirm what matters. Actions that change data, spend money, publish content or contact people wait for a person to approve them.
• Scoped tools. Each agent can use only the tools and data its role allows, per customer.
• Everything is logged. Agent actions go to an audit log that can't be quietly edited.
• Defenses against manipulation. We treat text from web pages, files, emails and users as data, not instructions, and test our features against prompt-injection attempts.
• Labels. Content written mainly by AI is marked where a reader could reasonably be misled.

Your data and AI

• Customer data is used only to provide the service you asked for.
• We don't use customer data to train AI models without your clear consent.
• Where a feature sends data to an AI provider, we use providers whose business terms don't allow them to train on it, and we send only what the task needs.
• The contact form on this site isn't processed by AI to make decisions about you.

Limits

AI can be wrong. Treat AI suggestions as drafts to check, especially for legal, financial, medical or safety questions. If an AI feature of ours gives a harmful, biased or wrong result, tell us at hello@symservices.site and we'll look into it.

Questions

Email hello@symservices.site. Our other policies: Privacy · Cookies · Terms · AI policy · Security.