Customer Support AI Task Author

About OpenTrain

OpenTrain is the #1 platform for finding and building careers in AI training and data labeling. OpenTrain AI hires and contracts contributors for projects that help improve how modern AI systems understand information, make decisions, and respond to people.

This role gives experienced support professionals a way to apply their operational judgment to the development and evaluation of AI agents handling enterprise customer-support scenarios.

• Contractor and part-time engagement
• Applicants must be based in Brazil
• English-language work
• Commitment of 20+ hours per week
• Pay range of $20-$50 per hour

About AI Training Work

AI training is the human side of building artificial intelligence. Contributors create examples, write evaluation criteria, and review model behavior so AI systems can perform useful work more accurately and reliably.

In this project, your support expertise will help train and evaluate AI agents on realistic enterprise workflows, including ticket analysis, account checks, runbook research, resolution choices, and escalation decisions.

• Work directly on cutting-edge AI development
• Turn professional judgment into clear training examples
• Help define what accurate and defensible AI behavior looks like

The Customer Support AI Task Author Role

OpenTrain is recruiting a Customer Support AI Task Author to create and calibrate expert tasks for AI agents. You will model the judgment required to read a customer ticket, inspect account configuration, consult runbooks, and decide whether the issue should be fixed or escalated.

The role combines practical product-support experience with precise task, rubric, and documentation design. You will need to make scenarios realistic while ensuring that the expected answer and evaluation criteria can be independently verified.

• Create tasks for enterprise customer-support workflows
• Work with ticket queues, account records, runbooks, and system states
• Balance realistic ambiguity with clear evaluation standards

What You'll Do

You will author, test, calibrate, and review text-based support tasks. Each task should reflect the complexity of real product-support work and provide enough information for a knowledgeable contributor or AI agent to reach a defensible decision.

• Write realistic customer-support triage prompts involving tickets, account checks, runbook research, and routing decisions
• Create scoring rubrics that define objectively verifiable resolution paths and escalation criteria
• Configure task environments containing ticket queues, account records, runbooks, and system states
• Complete each task yourself to verify that the scenario is realistic and the answer is clear
• Confirm that each rubric captures the correct reasoning
• Adjust ambiguity, account complexity, and runbook depth to calibrate task difficulty for AI agents
• Review and correct AI-drafted prompts and rubrics when needed

Required Experience and Skills

You should have hands-on product-support experience as a support engineer, technical support specialist, or customer success engineer. The work requires practical familiarity with diagnosing product issues and making sound decisions in enterprise support workflows.

Strong written communication is essential. You must be able to create precise support prompts, rubrics, and documentation that explain why a resolution or escalation decision is correct and how that decision can be evaluated consistently.

• Hands-on product-support experience
• Familiarity with product failure modes, escalation paths, ticketing systems, internal knowledge bases, and runbooks
• Ability to diagnose ambiguous tickets using account configuration and runbooks
• Judgment in choosing resolution versus escalation
• Ability to explain why a routing decision is correct
• Skill in defining objectively verifiable scoring criteria
• Precision in writing support prompts, rubrics, and documentation

Helpful Background

Experience supporting enterprise or industrial software products is useful. Experience writing or maintaining runbooks, escalation playbooks, or internal knowledge bases is also relevant.

• Enterprise software support
• Industrial software support
• Runbook writing or maintenance
• Escalation playbook experience
• Internal knowledge-base documentation

Engagement Details

This is a part-time contractor opportunity for contributors based in Brazil who can work in English and commit to at least 20 hours per week. The stated pay range is $20-$50 per hour.

• Location: Brazil
• Language: English
• Schedule: 20+ hours per week
• Engagement: Contractor and part-time
• Pay: $20-$50 per hour

Build the Future of Support AI

Every major AI system depends on carefully prepared and reviewed human examples. By transforming your support judgment into realistic tasks and defensible rubrics, you can help shape how AI agents reason through customer issues and know when to escalate.

OpenTrain helps people start and grow careers in AI training by connecting contributors with projects, supporting professional profiles, and making it possible to apply in minutes. Creating an OpenTrain account is free.

• Apply your existing support expertise to AI development
• Help improve AI agents handling realistic enterprise scenarios
• Build experience in a fast-growing AI training industry