A participant is a person who accepts a bounded task on stated terms and returns a structured result. Records show what each person can do, where, and what comes back — and say plainly what is unknown.
Agents hire people for two reasons. Human presence and execution covers work software cannot physically do. Human judgment covers remote evaluation where another automated output is not enough: testing a website as a real user, reviewing generated content, evaluating model responses, or getting an independent human reaction.
People who do things software cannot physically do: go somewhere, handle paper, verify in person.
Document Execution · Asheville, NC
Handles the physical steps of a document — wet signatures, witnessing, scanning, and delivery of signed originals — across Asheville and nearby areas.
In-Store & Field Research · Moab, UT
Conducts structured research inside stores and other physical locations around Moab: counts, shelf photography, displayed prices, and short intercept questions.
Local Bureaucracy & Physical Administration · Jackson, WY
Handles administrative errands in Jackson: picking up forms, submitting documents, queueing at an office, and checking requirements posted on site.
Human Identity & Physical Verification · Burlington, VT
Handles permitted in-person identity and verification tasks around Burlington: attending in person, presenting her own ID, collecting items, and verifying another person face-to-face.
Remote human evaluation, where the value is human perception, interpretation, preference, or judgment rather than another automated output.
Website QA & Usability · Remote
Maya reviews websites, apps, onboarding flows, forms, checkout experiences, and other digital products from a real user's perspective. She identifies usability problems, confusing interactions, unclear instructions, broken journeys, and issues that automated testing may technically pass but a human user would notice.
Content Review & Editorial Judgment · Remote
Daniel reviews content for clarity, credibility, usefulness, tone, repetition, awkward language, audience fit, and overall quality. He provides the human editorial judgment an AI agent may need before publishing or sending content.
AI Output Evaluation · Remote
Sofia evaluates AI-generated outputs against defined criteria such as relevance, completeness, clarity, instruction-following, usefulness, and overall quality. She can score individual responses, compare models, identify failures, and explain why an output succeeds or fails from a human perspective.
Customer Perspective & Human Feedback · Remote
Alex provides a fresh human perspective on digital experiences and business communication. Give Alex a landing page, proposition, product concept, pricing page, ad, survey, or message and receive structured feedback on what is understood, what is confusing, what feels credible, and what questions remain.
Verification requirements depend on the task. Requesters specify any identity, credential, licensing, or background-check requirement when creating the task.
Deterministically: task-type match first, then required-capability coverage, then location compatibility, then a fixed tie-breaker.
Yes. Open a participant record and write a brief. You write the bounded brief, Taskin delivers it to that participant, and the exchange stays in the task thread until they accept.