Companies seek work signal beyond the résumé.

Before the final decision, a candidate works with the team on a paid, bounded real project for two to five days.

A trained interviewer from another department evaluates values, communication and async fit near the end; a “no” is binding.

Candidates do paid real work to generate hiring signal.

A paid project is run with candidates before a full-time offer.

The first period combines explicit expectations, intensive context, early feedback and progressively larger responsibility.

Application, video interview, assessment, panel, offer, closing, and onboarding are published as separate stages.

Cal.com uses a 30-day contract-to-hire trial before most full-time hires, then makes a full-time decision at its end.

The stated engineering hiring flow includes a past-project deep dive, feature design with an engineer and an onsite Super Day with work-like problems; other roles may follow different steps.

Elastic’s guide dated 16 July 2026 gives candidates an explicit list: AI is fine for polishing a résumé, structuring thoughts and practising likely questions; inventing experience and using AI in live interviews is not. Candidates who prepared with AI must be able to explain their approach and reasoning themselves.

Zapier sets a minimum AI fluency bar for every new hire and measures it with a rubric. The second version expects candidates to show AI embedded in their core work, repeatable systems instead of one-off prompts, and visible impact on quality or efficiency; four components are assessed: mindset, strategy, building and accountability. It is measured at four points: the application, an initial screen, a skills test where candidates are observed using AI live, and an executive interview.