Sentinel questions

HumanLayer is a platform that enables qualified professionals (lawyers, physicians, analysts, domain experts, compliance leads, etc.) to issue formal human decisions on demand when AI systems reach their limits.

You do not execute tasks.

👉 You exercise your decision authority in a structured, traceable, paid framework.

Your role is to:

• review a case prepared by AI
• render a decision according to your expertise and authorizations
• justify your decision in writing (mandatory justification)

You are:

• identified
• qualified
• responsible for the decision you make

HumanLayer never alters your decision.

No.

You decide:

• for a client organization
• within a clear contractual framework
• with rules defined in advance

AI prepares the case.

👉 You make the decision.

It depends on your profile, but typically:

Level 1 — Compliance validation

• rule checks
• regulatory compliance
• approval according to defined policies

Level 2 — Expert judgment

• arbitration of ambiguous situations
• contextual decisions
• justified exceptions

Level 3 — Authorized authority

• formal sign-offs within your authorization
• regulated decisions
• legally binding acts

Requests are assigned based on your domain and, when specified, the jurisdiction. If a request exceeds your authorizations, escalate it with a justification: it will go to another Sentinel.

💰 Revenue share, per decision

You are paid for each decision rendered (approval or reasoned rejection), not by the hour.

Indicative example:

• compliance decision: €10 to €12
• expert judgment: €50 to €60
• authorized authority: €125 to €150

The amount depends on the decision level (ranges above).

👉 You are paid for accountability, not passive availability.

The client organization pays HumanLayer for the decision.

HumanLayer pays the Sentinel up to 60% of the decision price; the rest funds credential verification, SLA orchestration, traceability and the platform.

The percentage depends on the decision level and never exceeds 60%.

The model is transparent and contractual.

Yes, in most cases.

HumanLayer is designed for:

• occasional interventions
• targeted
• high value

When you apply, you state:

• your domain of expertise
• your jurisdictions
• your capacity (occasional, regular but limited, sustained or priority)

Requests are then assigned automatically, least-loaded Sentinel first; the deadline depends on the level (30 min, 2 h or 4 h). A request you cannot handle can be escalated.

👉 No exclusivity required.

Yes — as in your usual professional activity.

The difference:

• the framework is clear
• the context is documented
• the trace is complete

HumanLayer protects as much as it exposes:

• no informal decisions
• no ambiguity about "who approved what"

HumanLayer provides:

• the case as you saw it
• your decision
• your justification
• the timestamp
• your identity and authorization

👉 You don't depend on fuzzy emails or memory.

Yes, as long as the request is assigned to you.

You can:

• reject it: a decision on the merits, justified and sent to the client
• escalate it with a justification (out of scope, conflict of interest): it then goes to another Sentinel

No penalty for a justified refusal.

Not like a marketplace.

Your application receives an initial score based on what you declared (experience, authorization, qualifications, capacity); among equally loaded Sentinels of a domain, it decides who is assigned first. SLA breaches are timestamped in the log.

👉 The outcome of your decisions never affects this score.

Multi-step process:

• Identity verification
• Qualification verification
• Interview (depending on level)
• Account activation: domain, jurisdictions, mandatory two-factor authentication

The goal is:

• protect your authority
• protect client organizations
• ensure network credibility

No.

HumanLayer's value rests on:

• clear identification
• assumed accountability

You are pseudonymized: the client sees a stable reference (S-XXXXXX) attached to each signed decision; HumanLayer keeps the link to your identity, which can be disclosed to authorized parties within the contractual framework.

Because:

• you don't have to prospect
• cases are prepared
• demand is qualified
• compensation is clear
• the trace is protected

👉 You intervene where your decision has the most impact.

• ❌ micro-work
• ❌ a low-cost marketplace
• ❌ a replacement for professionals

👉 It's an infrastructure for human accountability.