Our Methodology

We use an automated tool as part of assessing applications to our expert network. Every applicant completes a short structured interview. The audio is transcribed and scored against a domain rubric, and that score can determine the outcome without a person reviewing it: a sufficiently strong score approves the application automatically, and a clearly failing score declines it. Scores in between are held for review by our team.

What it assesses: the substance of your spoken answers about your own field, measured against the rubric for that domain, together with the credentials you supply in the application. It does not assess accent, speech fluency, tone of voice, appearance, or any demographic characteristic, and we do not collect demographic information as part of the application.

Language verification is separate, and optional. Approved experts may choose to verify a language they speak through a short spoken conversation in it: we ask a question aloud, you answer, and the next question is generated from what you said. That check is an assessment of language ability — it is the one place where how you use a language is the thing being measured — which is why it is a separate, opt-in step with its own consent, and never part of the application interview above. It is assessed from the transcript, so it measures vocabulary, grammar and coherence; it cannot hear and does not assess accent or pronunciation. The result is a CEFR level used to staff projects that require that language. Unlike the application interview, these recordings are kept, so a member of our team can listen where a decision needs a human ear; we delete them whenever you ask, and when your account is erased. Declining the check, or receiving a low result, has no effect on your approved status or on work you already do with us.

Your rights: you may request that a person review any automated decision, and we will re-review it. You may ask us what data we hold about your application, and have your interview recording deleted at any time by emailing support@pasifloraai.com. Otherwise we keep the recording for up to 90 days and use it only to review your application and to refer you to projects that fit your expertise. We never sell it, never use it to train AI models, and never hand it to clients. Full detail is in our Privacy Policy.

Alternative format: if a spoken interview is not accessible to you — for any reason, including a disability, a speech difference, or not having a reliable microphone — email support@pasifloraai.com and we will arrange an alternative. Asking will not count against your application.

To exercise any of the above, or to ask how the tool works in more detail, contact support@pasifloraai.com.