Agent by An
Privacy & responsible AI

Plain-language rules for safe workflow scoping.

Agent by An is built around supervised AI systems: anonymized samples first, clear review gates, logs, cost awareness, and fallback paths.

Do not submit secrets

Do not submit credentials, API keys, tokens, passwords, private customer data, financial records, or sensitive raw company documents through the initial quote flow.

Use anonymized samples first

Initial review should use examples that preserve structure without exposing private data. If real data is required later, access boundaries must be scoped first.

Quote submissions

Quote submissions are used to assess project fit, risk, scope, recommended next step, and whether the workflow should be automated at all.

Human review

Important workflows should keep human approval, logs, and a manual fallback path. Blind autonomous decisions are not the default.

Sensitive workflows

Legal, medical, financial, HR, customer support, and destructive data workflows require extra scoping before implementation.

Retention and access

Early intake should stay minimal. Larger projects can define separate data retention, access, logging, and deletion rules in the proposal.

Have a sensitive workflow?

Start with a high-level description and anonymized structure. Do not send raw sensitive data in the form.

Request a workflow quote →