AI Applyd

Are scope, spend, and authority enforced while the agent runs?

The vendor publishes a specific, numeric run-time ceiling rather than a general assurance: a hard daily submission cap per account (quoted as 15 a day on one plan and 50 on another) plus pacing of submissions per hiring system, and OAuth access is granted through a consent screen that names the requesting client. Not scored 3 because the agent is explicitly unsupervised at the action level, it fills every field and presses submit with no per-application approval, so authority over the individual consequential act is not gated, and the caps are self-attested with no third-party verification.

Counts Autonomy level, Runtime governance and Permission scopes, each cited below.

Is there a tamper-evident record of what it actually did?

A record exists and is customer-visible: the vendor states it keeps its own submission records, the submit its browser watched go through and a screenshot at the end of the run, reviewable from the customer dashboard. Credit is due for the unusually honest framing, the vendor labels these as its own records describing what it did rather than what the employer received. Scored 1 and not higher because the record is self-reported with no tamper-evidence, no hash-chaining and no immutability (audit_trail_immutable is no_public_information), and the vendor itself disclaims that it reflects employer receipt.

Counts Audit trail, each cited below. Also named Tamper-evident log, which our record now shows as undisclosed; that part of the support no longer holds. What a score rests on

Are compliance obligations attached per engagement?

No public evidence found. The capture names no certification, no auditor, no report and no regulatory framework. compliance_certs and regulatory_alignment are both no_public_information against the page checked.

Rests on Compliance certifications and Regulatory alignment, recorded as undisclosed and cited below. For a 0 that absence is the finding. What a score rests on

Does payment depend on a verified result?

Payment tracks work performed rather than seats or tokens: billing is per finished application at a published unit cost, and a job already dead before a browser session opens is not charged, with the reservation returned and only genuinely used compute billed. That is above the rubric 0 condition of paying per seat regardless of result. It is not 2 or 3 because a filed application is an output, not a verified outcome, nothing releases payment against an interview or a hire, and settlement_mechanism and outcome_based_pricing are both no_public_information on the capture.

Counts Price point and Pricing model, each cited below. Also named Outcome-based pricing and Settlement mechanism, which our record now shows as undisclosed; that part of the support no longer holds. What a score rests on

Can the deployment be insured, and is the customer indemnified?

No public evidence found. No insurance, no named carrier, no published limit, no customer indemnity and no liability cap appear on the capture. All six assurance fields are no_public_information against the page checked.

Rests on Own liability cover, Coverage limits, Indemnification, Insurance available, Insurance carriers and Liability cap, recorded as undisclosed and cited below. For a 0 that absence is the finding. What a score rests on