How it works (from your side)

For candidates · How it works

Five stages. Each one

explained from your side.

What you do, what we do, how long it takes, what to expect when you do not get the job, and what you keep regardless. Read it before you apply. Bookmark it for after.

Find a role →Read about your data

Stage 01

Tell us who you are.

You land on a Picked-branded application for the role. Not a CV upload form. A short conversational interface that captures the things the role brief actually asks for. About 5 minutes.

●You can pause and resume up to 72 hours.
●You see a time estimate before you commit.
●You choose the language; we support 25-plus.

APPLY · SAMPLE

Role: Senior Backend Engineer●

Right to work: confirmed·

Notice period: 4 weeks·

Sample. Shipping with V1.

Stage 02

Have a 15-minute conversation.

Within minutes of applying, the AI invites you to a screen. By voice or chat, your choice, your language. Motivation, must-haves, deal-breakers, your own questions. About 15 minutes.

●The AI follows up. It is not a script.
●You can ask the AI questions about the role and the company.
●You can pause and come back. The session waits.

Senior Backend Engineer · screen

09:14 / 25:00

Picked·09:02

Hi Amara, thanks for applying. I'm going to ask a few questions, then send you a short systems exercise. About 25 minutes total. You can pause anytime.

Amara·09:03

Sounds good. Ready when you are.

Picked·09:04

You mentioned a sharding migration at Flutterwave last year. Walk me through the trade-off you actually agonised over, not the one that looked easy in hindsight.

Amara·09:06

The hardest call was choosing between hash-based and directory-based sharding. Hash was simpler, but our access patterns were strongly customer-skewed and we were going to end up with hot shards...

Picked·09:08

Following up

Interesting. What happened the first time you had a hot shard? And, separate question, when did you tell your manager you were going to have a hot shard?

Amara·09:11

Honestly, after. I should have flagged the risk earlier. We caught it in the load tests, not in production, but it shouldn't have been a surprise to anyone...

Type a reply or press Space to speak...

Voice on

Stage 03

Run the assessment that fits the role.

If your screen suggests you fit the role, the AI selects a role-appropriate assessment from the Picked library. Cognitive, behavioural, situational judgement, or a role-specific task (a coding exercise, a writing sample, a sales role-play). About 25 minutes on average.

●The assessment is chosen for the role, not for catching you out.
●Accommodation on request: extra time, dyslexia mode, screen-reader compatibility, slower voice.
●We do not punish you for asking.

ASSESSMENT · SAMPLE

Cognitive module: numerical reasoning●

Situational judgement: 8 items·

Role-specific: short coding task·

Sample. Shipping with V1.

Stage 04

Have a 30-minute behavioural interview.

If your assessment suggests you fit the role, the AI invites you to a 30-minute first-round interview, by voice. It asks behavioural questions tied to the role's competencies, with adaptive follow-ups. Or, if you prefer, you can ask for a human interviewer at the same point in the process. No penalty.

●Voice only at V1. Video lands with V2 where the law allows it.
●The right to a human interviewer: one click, same questions, same rubric.
●Recording is consented per session. You can pause, stop, or redo.

INTERVIEW · SAMPLE

Opening (2 min)●

Behavioural anchors (3 to 4)·

Scenario (1)·

Your questions (3 to 5 min)·

Confidence calibration·

Sample. Shipping with V1.

Stage 05

Get a real decision, with feedback.

After the interview, the AI ranks all candidates for the role and recommends a top three to the hiring manager. Within 48 hours of the role posting, the hiring manager makes the final call. You see the result, the reasons, and the date and time. No ghosting. Ever.

●Advanced? The hiring manager invites you to their final round.
●Not advanced? You get a personalised note: where you scored, what would strengthen a future application, which adjacent roles might fit you better.
●Either way: the data is yours. Export it, delete it, re-use it elsewhere.

DECISION · SAMPLE

Status: advanced to final round●

Date: Tuesday, 06:42·

Decided by: Hiring manager (named)·

Sample. Shipping with V1.

~75 minutes

End-to-end candidate time across all four stages, on average.

72 hours

Pause-and-resume window at every stage.

One tap

To request extra time, dyslexia mode, screen-reader compatibility, or a human alternative. No flag against you.

Your right

You can always ask for a human.

Every stage, every decision, every score: you can ask for a human to review it or to conduct it instead. We open a ticket; a Picked-trained reviewer takes the case with full file access; you get an answer within five working days. The original AI decision is paused while the review is open.

●Human-conducted screen: same questions, scheduled at your convenience.
●Human-conducted interview: same questions, same rubric, scheduled at your convenience.
●Human review of a decision: a Picked reviewer reads your file end-to-end. They can over-ride the AI.
●All free. None of these costs you the application.

After the decision

Four things you keep.

The decision in writing.

Every decision, with the date, time, and the human hiring manager who signed off.

Your transcripts and scores.

All of them. Exportable as JSON or PDF. Yours to keep.

A portable profile.

Reusable on the next Picked role. We do not make you re-do the assessment within a sensible window.

Personalised feedback (if rejected).

Where you scored, what would strengthen a future application, which adjacent roles might fit.