How can we help?

AI Employees are designed to assist with business operations, but their outputs should be evaluated according to the importance and risk of the task. An unexpected result does not always indicate a technical platform failure. It may arise from incomplete instructions, insufficient context, inaccurate source information, integration errors, or limitations of AI-generated reasoning. How to investigate 1. Review the original request. Check whether the instruction clearly describes the expected outcome. Ambiguous requests may produce results that differ from the user's intentions. 2. Check the available business context. Determine whether the AI Employee had access to the information required to complete the task. 3. Review the AI Employee configuration. Check its responsibilities, instructions, available capabilities, and relevant operational boundaries. 4. Verify source information. If the output depends on external records, confirm that those records are accurate and current. 5. Review execution details. Where available, inspect the recorded steps, collaboration results, or external tool outcomes. 6. Identify the type of error. Distinguish between an incorrect explanation, an incomplete response, a failed external action, and an unauthorized or unexpected business operation. 7. Test with a controlled example. Use a clearly defined task with a known expected result. Avoid testing changes directly against sensitive customer operations where a safer test is possible. Should I trust every AI-generated answer? No. AI-generated information can be inaccurate or incomplete. Important financial, legal, security-sensitive, or otherwise high-impact decisions should receive appropriate human review. What if the AI Employee performed an unexpected action? Review the relevant execution history and the affected business system. If necessary, pause the relevant automation or restrict further execution while the issue is investigated. Do not repeatedly run the same task until its effects are understood. When to contact Support Provide the AI Employee name, expected behavior, actual behavior, relevant execution reference, and a sanitized example. Avoid submitting unnecessary confidential customer data.