OnTask.AI is the operating layer behind GigAnywhere's proof-of-work platform. It runs sessions, summarizes activity, flags exceptions, and produces the record that makes remote engagements trustable.
No heavy desktop agents. Sessions run in the browser, with a visible indicator and worker-controlled start/stop.
Plain-language session summaries that surface what got done, what got skipped, and what to review.
Optional. For Verified engagements, OnTask confirms the worker is the person who was hired — without recording private content.
For ongoing work, set expected schedule windows. OnTask records adherence without policing breaks.
Idle work, off-process behavior, schedule drift, missed checklist steps — all surfaced with context, not raw data.
PDF + CSV exports per engagement, suitable for sharing with internal stakeholders or your own clients.
Clients configure OnTask rules in the workspace. Workers see those rules before applying. AI runs the work session. Humans approve the work.
Rules are part of the workspace. They're visible to the worker before applying.
| What OnTask records | Delivery Only | Managed Work | Verified Work |
|---|---|---|---|
| Job & workspace definition | Yes | Yes | Yes |
| Messages & files | Yes | Yes | Yes |
| Milestones & approvals | Yes | Yes | Yes |
| Session start/stop | No | Yes | Yes |
| Active task per session | No | Yes | Yes |
| AI session summary | No | Yes | Yes |
| Schedule adherence | No | Optional | Yes |
| Identity verification | No | Optional | Yes |
| Structured activity evidence | No | Limited | Yes |
| Personal time / paused sessions | No | No | No |
| Audio | No | No | No |
| Raw keystrokes | No | No | No |
| Off-platform activity | No | No | No |
Specific Verified Work configurations vary by category. Workers see the exact configuration on the job posting before applying.
Configure AI supervisor rules per workspace, save them as templates, and apply them to every new project of the same type.
Document review, order processing, data entry, CRM cleanup.
Schedule adherence, ticket QA, escalation review.
Per-record completeness checks, source citation, QA sampling.
Identity verification, audit-grade records, document collection.
Decision logs, sample audits, calibration sessions.
Client-facing work records, scaled fulfillment, white-label reporting.
OnTask was designed around consent, proportionality, and human review. It produces a record both sides can use — not a tool to police people.