What actually happens during Day One™

Everyone says they do something better than the interview loop. Everyone says they do a work sample. Very few will tell you what's in it, how it's scored, or what happens to the people who fail. Here's ours, in full.

The brief

You get a scenario built from the actual role, not a puzzle. For an executive assistant it might be a week that's fallen apart: three conflicting calendar holds, a client escalation sitting unanswered since yesterday, and a travel booking that needs to change. For an operations hire it might be a broken intake process and a spreadsheet that disagrees with the CRM.

The brief is deliberately incomplete. Real work is incomplete. Part of what we're watching is what you do about the gap: whether you assume, ask, or state your assumption and move.

The clock

The window is 90 to 120 minutes, and it's a real boundary. Not because speed is the point, but because unbounded take-homes quietly test who has a free weekend. A time-boxed exercise tests prioritization, which is the skill the job actually needs. Nobody finishes everything. What you choose to finish is the signal.

AI is allowed. Obviously.

Use whatever you use on a normal working day. We ask you to tell us how: which tools, for which parts, what you changed in the output.

We're not testing whether you can work without leverage; nobody on your future team does. We're testing judgment on top of leverage. The failure mode we score down is not "used AI." It's pasted the first draft and shipped it, and the tell is confident output that doesn't survive contact with the brief's constraints. The strongest submissions are almost always heavily AI-assisted and heavily edited, and the candidate can tell you exactly which parts they didn't trust.

The rubric

Written before anyone sees a submission, specific to the role, and consistent across every candidate for it. Broadly it covers four things:

Output quality. Is the deliverable something you'd send to a client without a rewrite?

Judgment. Given an incomplete brief, did the ordering of the work make sense? Did the right thing get done first?

Communication. Are the assumptions stated? Is the handoff note something a busy person can act on in thirty seconds?

Leverage. Was the tooling used well, and can they account for it?

Scoring happens against the submission, not the person. Name and history aren't what's being read.

What happens after

If you pass, you're in the Talent Pool and can be matched to a role. Sometimes there's a short QA chat first, but only when a detail in your work is faster to talk through than to trade emails about, and it's skipped entirely when nothing needs clarifying. When a conversation does happen it starts at a different altitude than an interview. Nobody is asking you to describe your strengths, because your strengths are in the file.

If you don't pass, you're told why. Specifically, against the rubric, in writing. This is the part of the process that costs us the most time and the part we'd defend hardest. A rejection with a reason is the only kind a person can do anything with, and it's the difference between being screened out and being seen.

Why we can't just show you the brief

The one thing we protect is the specific scenario, because a public brief becomes a prepared brief within a week and the signal evaporates. Everything else (the rubric structure, the timing, the AI policy, the scoring criteria) we'll tell anyone who asks. The rubric isn't a secret. The surprise isn't the point. The work is.

The bar

Most candidates don't clear it, and that's the design. Actors audition. Athletes try out. Musicians play before they get signed. In every field where performance is what matters, people show the work first and get the role second. Corporate knowledge work is the strange exception, and Day One™ is just us declining to be exceptional about it.