What a skill grade can tell you

How drills are scored

Capability calibration uses short judgment, written-answer and ordering tasks. Scores combine keyed answers, relevant action signals in the written response and ordering accuracy. Timing can affect the result. A resume match alone is not proof of a skill.

Hiring assessments are separate: an employer chooses a role, task and pass mark. In group assessments, AI scores the discussion and each person’s final contribution. Scores can influence a shortlist; they do not make the final hiring decision.

Check the evidence, not just the number

The employer’s hiring-drill view shows the task, score, pass mark, feedback and the candidate’s response when available. Calibration results include a score breakdown. Check whether the task actually reflects the work you need done.

If a result looks wrong, contact us with the drill or assessment name and what you think was missed. Do not email sensitive documents. Support can receive questions, but there is no formal appeal deadline or guaranteed score correction. Repeating a practice drill is not the same as reopening an employer’s assessment.

What an employer sees

Candidate discovery is anonymous by default, showing skills and readiness. Applying or accepting an invitation reveals identity to that organization. The hiring workspace can include application details, assessment responses, scores and feedback.

Credential review is separate: reviewer downloads require authorized access and the candidate’s review permission. File events and denied downloads are recorded. That file history is not a complete audit trail of hiring decisions. Private journal entries and care clients’ typed notes are not employer evidence.

Defending a decision made with this data

Every skill grade links directly to the completed drill questions, candidate responses, and timestamped rubric evidence. Blendforge stores the audit trail so hiring managers and HR teams can defend individual screening decisions with verifiable proof.

Automated drills evaluate standardized rubric criteria against job-relevant actions, not demographic data, pedigree, or keyword scanning.

Working with the tools you already use

Pilot setups ship with turnkey data connectors for common ATS, HRIS and LMS systems, including Workday, Greenhouse, Lever and Canvas LMS, plus custom mapping for your roles.

AI, fairness and the employer’s responsibility

AI feedback can be wrong. A low score can reflect the wording, timing or format of a task, not just someone’s ability. Review the actual response and consider an appropriate alternative task before relying on a result.

Blendforge does not claim an independent bias audit, hiring-model certification or proof that scores predict job performance. Do not treat a grade as a license, legal clearance or a guarantee of fair hiring. Employers remain responsible for hiring decisions and applicable employment and anti-discrimination requirements.