What the work involves

micro1 is staffing this for a customer's systems alignment project. You'll be handed realistic (often deliberately messy) slices of a SaaS revenue stack — CRM objects, subscription and billing records, support tickets, automation rules — and asked to find where the operational logic contradicts itself. Duplicate accounts with divergent entitlements. A renewal automation that fires off a stale close date. Permission sets that let a CSM edit a field the billing system treats as source of truth. Your job is to surface these, decide what the correct end state is, and write it down.

The deliverable is rarely a config change. It's documentation: a precise statement of the conflict, the reasoning behind your proposed resolution, the cascading effects across adjacent systems, and — importantly — what you chose to leave alone and why. Tasks often ask you to draft operational prompts describing how an ambiguous scenario should be handled, in language a model can learn from. Expect grey areas around mid-term entitlement changes, co-term renewals, partial refunds, and downgrade logic, where written policy runs out and judgment starts.

What the screening looks for

  • Hands on the keys. The platform probes whether you personally built validation rules, workflows, and permission models — or signed off while an admin did. Follow-ups get specific fast: object names, field types, why you chose a formula over a rollup.
  • Multi-system fluency. One CRM isn't enough. You should be able to trace a record from quote through subscription through invoice through support ticket and explain where each system claims authority.
  • Reasoning under conflict. Scenarios are built so that two defensible answers exist. Evaluators score the transparency of your trade-off, not the answer you land on.
  • Written precision in English. Nuance carries the value here; vague documentation is the main failure mode.

Logistics

Contractor engagement, fully remote, asynchronous. Work is delivered as discrete tasks rather than scheduled shifts, so you set your own hours — though most contributors find blocks of two to three uninterrupted hours produce better documentation than scattered fifteen-minute sessions. Volume varies with project phase. The $45–90/hr band reflects rates observed on similar micro1 engagements and is not guaranteed; placement within it typically tracks depth of system ownership and domain complexity. No prior AI or ML experience is expected or required.