Most of the tools recruiting teams already use are built to do one part of the job well — score code, filter keywords, schedule calls. Gradogram is built to be the layer that runs the whole assessment step: writing the rubric, grading against it with AI, keeping a human in the loop, and holding a compliant record of what happened. The table below is a straight comparison of what that gets you versus what a typical standalone screening or assessment tool covers. If a specific tool you already use happens to do one of these things well, that's genuinely fine — this is about what's typical of the category, not a takedown of any one product.
| Capability | Gradogram | Most other platforms |
|---|---|---|
| AI grades every answer against a written rubric, and a person confirms the result before it counts | Auto-score or auto-pass/fail, no confirm step | |
| The reasoning behind every AI-drafted score is visible, not just a number | Score or pass/fail only, no visible reasoning | |
| Grading rubric is frozen per role — every candidate is scored on the same criteria | Criteria can drift reviewer to reviewer | |
| One system covers the whole hiring funnel — job, candidate, assessment, grading, decision | Usually one step; the rest gets stitched together | |
| Assessments run under a timed, consent-logged session | Timers are common; logged consent less so | |
| Copy-paste and tab-switch signals go to a reviewer — never used to auto-flag or auto-reject | Where integrity signals exist, review step varies | |
| Recruitment automation can act on its own for routine steps, or hold for a human — by rule | Automation, where it exists, tends to be all-or-nothing | |
| Candidate records support built-in GDPR anonymization ("right to be forgotten"), not just manual deletion | Usually export-then-delete, no native anonymize | |
| Every material action is recorded in an immutable audit trail — who, what, when, why, before/after | If logging exists, rarely spans the full hiring decision | |
| Integrations run on signed webhooks with a full delivery audit trail | Usually CSV export, not event-driven delivery |
Why a checklist instead of a takedown
We're not going to pretend every recruiting team's current setup is broken — a lot of it works fine for a lot of roles. This list exists because "we use AI to grade" gets asked to mean a lot of different things, and the honest answer is in the specifics: whether a person still signs off, whether the criteria are the same for every candidate, whether there's a record of any of it six months later. That's what's actually in the table above. For the deeper, single-comparison version of this argument, see AI-graded assessments vs. manual resume screening.
What this list doesn't cover
A few real things are missing here on purpose, not by omission. Gradogram doesn't watch a candidate's webcam or track biometrics — integrity here means a timer and a copy-paste guard, not surveillance. And multi-tenant data isolation and the full compliance posture live on their own page, Trust & Compliance, because that's a deeper claim than a single checklist row can carry honestly.