Can you build?
Every candidate faces the same interactive challenge and solves it by directing an AI — exactly how the work is done now. You see how someone actually thinks and works, and you judge the work, not the wording.
AI structures the evidence and surfaces the signal. People make the call — Can You Build? never accepts, rejects, or ranks a candidate on its own.
The CV and cover letter just stopped meaning anything.
When every application is written by the same handful of AI tools, the document tells you about the tool — not the person. Can You Build? measures what's left that's hard to fake: how someone actually works when you hand them a real problem.
The cover letter is dead — and AI killed it.
“De sollicitatiebrief is dood, en AI heeft hem vermoord”
Keep your pipeline. Replace the part that's guesswork.
Your existing tools, your sourcing, your final interviews all stay exactly as they are. It drops in right after the CV screen: every shortlisted candidate runs the same interactive challenge, and you get back scored reports backed by evidence — not another stack of résumés to re-read.
The two blue steps happen in Can You Build?. Everything before and after stays exactly where it works today — candidates stay anonymous, and the link-to-person mapping never leaves your side.
What those two steps actually look like.
Anchored every option in mid-market ops pain points, not generic benefit claims.
Cut the jargon; the subhead got tighter without losing the claim.
Chose option B with a defensible why, backed by evidence from customer calls.
Built for everyone in the room.
No names. No emails. No PII to govern.
Every candidate is an opaque link token. Your organisation keeps the link-to-person mapping on its own side; Can You Build? holds the assessment, not the identity — which turns a long security review into a short one.
AI assists. People decide. Every call is on the record.
Evaluating candidates with AI is high-risk under the EU AI Act — regulated, not banned. Can You Build? is built around the safeguards it expects: a named reviewer makes every decision, and the rubric, transcript, and notes form an exportable trail. It never decides on its own — and because assessments are anonymous and text-only, there's no emotion inference, biometrics, or social scoring in the picture. Candidates get the same transparency: before the timer starts they see the scoring criteria, the time commitment, and exactly what is recorded while they work.
Built around the Act's oversight, transparency, and auditability safeguards. Informational, not legal advice; full compliance depends on how you deploy it.
- ✕ Auto-accepts or rejects candidates
- ✕ Infers emotion or uses biometrics
- ✕ Hides the scoring criteria from your team
Can You Build? is a new experiment in hiring for the age of AI. It's early — no customer logos, no case studies. Judge the work, not the marketing.
See a sample scored report ↓It's an experiment. Come try it on one real candidate.
We're opening it to a handful of teams first. Tell us the roles you hire for and we'll set you up. No sales call, no seats, no subscriptions — when it's ready to bill, it's prepaid credits run close to cost, and you only top up when you need to.
Request access →- ✓ Prepaid credits — pay as you go
- ✓ No subscription, no per-seat fees
- ✓ Scenario drafting & AI scoring included
- ✓ Test it on one real candidate first
Questions we get asked first.
What stops a candidate from cheating, or having someone else do it?
The assessment is one evidence point, not a lockdown exam — it isn't hardened anti-cheat. It flags behavioural signals (tab switches, copy/paste actions) as risk hints for a human to review, never as an automatic verdict — and every candidate works under the same constraints, including a fixed per-stage budget of AI replies shown live. Identity gets verified where it actually matters: in the interview, when the person walks you through the same problem they just solved.
What candidate data do you hold, and for how long?
None that identifies anyone. Candidates are opaque link tokens — your org keeps the token-to-person mapping on its own side. Unused invite links expire after 14 days, and deleting a campaign or candidate cascades to its session, transcript, and evaluations.
Which AI model do you use — does it train on our data?
Anthropic's API powers scenario generation and scoring. Anthropic's API terms state they don't train on API data by default, and we don't train anything on your content either. See our Privacy Policy for the full list of sub-processors.
Which roles and languages does it support?
We're starting with engineering — where directing an AI to build something real is easiest to see — but scenarios are authored fresh per campaign, not picked from a fixed library, so any role you can brief works. The dashboard is in English today; since the candidate's AI partner is a general-purpose model, most candidates can also work with it in their own language.
What happens if the candidate's time runs out?
The stage auto-submits at zero and scores whatever they'd written so far — nothing is lost, and there's no penalty beyond the incomplete work itself.
Imagined by builders. Built for hiring.
Can You Build? is an experiment by paddoq, a small group of engineers exploring how work — and how we hire for it — changes now that everyone builds with AI. Every assessment is evidence: what someone can actually do, judged on the work.
Questions? Say hello.
Tell us about the roles you hire for and what you want to measure. We read every message and reply personally — usually within a day.