For the admin

Author a take-home challenge. Send it in a click.

Pick from the library or write your own. Add a grading guide and a short questionnaire, then drop the candidate a tokenised link and watch it land.

1

Set the brief

Challenge

Library challenges are shared across the platform. Org-owned ones stay in your team. Either way, you shape the brief, grading guide, and questionnaire.

Challenge

Acme relaunch campaign brief

Mid · 3h estimated · Written response

Draft a launch plan for Acme's reissued flagship product. Two pages. Name the audience you're betting on, the positioning line you'd lead with, and how you'd know in ninety days if it worked.

Edit brief Grading guide Questionnaire Send to candidate →

2

Watch the activity feed

Timeline

Sent, started, submitted, reviewed — every event lands in the audit log. Reviewers can answer "when did this candidate actually start?" without a Slack thread.

Jane Candidate started the challenge 09:42

Jane Candidate submitted — response locked at 13:11 13:11

Graden AI finished the review 13:13

Link lands in the candidate's inbox…

For the candidate

Open the link, do the work, submit it cleanly.

No account, no countdown timer. They see what's expected, start when they're ready, write or upload their response, and explain the choices they made in their own words.

1

Welcome page sets expectations

Kick-off

No countdown timer. They see what they need to do, get a friendly nudge to write in the portal or upload a PDF, and click Start when they're ready. That's when we record their start time.

Ready to begin, Jane?

Acme has sent you a take-home challenge. No countdown timer, we just note when you click Start.

Read the brief.
Write your response in the portal, or upload a PDF.
Answer a few short questions about the choices you made.

I'm ready, start the challenge →

2

Answer the questions on your choices

Free-text

Short prompts in their own words. A strong signal against AI-drafted submissions. We cross-check the stated rationale against what they actually wrote.

One last step

Answer in your own words so Acme can see the choices behind your response.

Who did you decide to talk to, and why them?

I went for returning lapsed customers rather than net-new buyers. They already trust the brand, and the reissue is the obvious reason to come back…

Which part of your plan are you least sure about?

The ninety-day measurement. I've leaned on first-order revenue as the signal, but I know repeat rate would tell a truer story…

AI review runs in the background…

For the reviewer

Open a review that's already prepared.

By the time the reviewer sits down, Claude has drafted scores and talking points against the grading guide. The reviewer reads the work themselves, agrees or overrules, adds their own notes, and walks into the interview prepared. Scores are inputs to judgement, not verdicts.

1

Scores, summary, talking points

AI review

Calibrated scores against every grading criterion, a short plain-English summary, and a list of talking points tagged by category and severity.

Audience insight

82/100

Positioning

74/100

Measurement

60/100

Confident piece of thinking. The audience bet is specific and well-argued, and the positioning line earns its place. The weakest spot is measurement: the ninety-day plan leans on revenue and skips retention, which is where the reissue thesis actually lives or dies.

Audience segment is named and defended, not hedged. "Returning lapsed customers" is a real pick, not a wish list. response.md:p2 · audience insight

Success metric is first-order revenue only. Worth pushing on how they'd know the reissue landed with the audience they picked. response.md:p6 · measurement plan

2

Read the candidate's own words

Q&A

The questionnaire answers sit alongside the AI review. Where the stated rationale contradicts the work itself, the AI flags it as an integrity talking point.

Who did you decide to talk to, and why them?

I went for returning lapsed customers rather than net-new buyers. They already trust the brand, and the reissue is the obvious reason to come back. Net-new would need a broader story I don't think ninety days is enough to land.

Which part of your plan are you least sure about?

The ninety-day measurement. I leaned on first-order revenue because it's the cleanest signal, but I know repeat rate would tell a truer story about whether the audience I picked actually stuck.

Reviewer walks into the interview prepared.

See Graden in action: a marketing take-home, prepared for the reviewer.