Assessment infrastructure for enterprise teams
Scorafy is assessment infrastructure for organisations running evaluation at programme scale. You define the rubric, the AI evaluates each submission against it and cites the evidence for every judgement, a named reviewer can override any score, and nothing reaches the respondent until an assessor releases it. Data is isolated per organisation with row-level security, hosted in the EU (Dublin) with an Australian (Sydney) residency option. Volume beyond the self-serve plans is handled through a conversation, not a checkout.
Free plan - no card required. Trial it on one assessment.
A few questions about volume and your current process, then a real conversation.
The problem this solves
Where enterprise assessment teams commonly lose time and consistency.
- Evaluation volume has outgrown the spreadsheets and shared inboxes holding the process together
- Different assessors reach different conclusions on comparable work, and nobody can show why
- Procurement and legal ask where the data sits, who can see it, and what the audit trail contains
- Results reach people before anyone senior has reviewed them, and there is no gate to stop it
- General-purpose AI tools produce plausible feedback with no rubric behind it and no record of who approved what
How Scorafy fits the workflow
The AI drafts, a qualified human reviews and decides.
Encode the standard once
Build the rubric - up to 10 criteria across 6 performance levels - plus a grading schema and terminology set, so the same standard applies across every programme and every assessor.
Collect work at scale
Respondents complete branded assessments with no account required, optionally timed with the limit enforced on the server. Written answers, uploaded documents, and video or audio answers with transcription are all read by the evaluation.
The AI evaluates and cites its evidence
Each submission is evaluated against your rubric in about a minute, and every score comes back with the evidence for it quoted from the respondent's own answer.
A reviewer decides what gets released
Assessors see the draft first, can override any score with a comment, and release results deliberately. The release is stamped and written to an audit log. Cohort and group reports give the programme-level view.
Why it works for enterprise assessment teams
The specific gains for this use case.
What Scorafy does and does not do
Scorafy is the evaluation layer, not your system of record. It does not manage enrolments, HR records or workforce decisions, and it does not make the assessment decision - a qualified reviewer does, and stays accountable for it. Scorafy holds documented controls; it does not hold SOC 2 or ISO 27001 certification, and single sign-on is not yet available. The security page states plainly what is and is not claimed.
Frequently asked questions
What does enterprise pricing look like?
Self-serve plans run to Pro at $249 USD per month for 2,000 AI evaluations. Above that, the Business band starts from $1,000 USD per month billed annually, and Enterprise agreements are quoted against your volume, residency and support requirements rather than published as a price. Tell us your volume and we will tell you which band fits.
Where does our data live, and who can see it?
The primary region is EU (Dublin), with an Australian (Sydney) region available for organisations that require it. Every table carries row-level security, so data is scoped to your organisation at the database layer rather than only in application code. Data is not used to train AI models. See the security and GDPR pages for the detail.
Do you have SOC 2, ISO 27001 or SSO?
No. Scorafy is an early-stage product and says so. The controls described on the security page are real and in place today; the certifications are not, and single sign-on is not yet available. We answer security questionnaires directly and publish a DPA.
Can we integrate it with our existing systems?
Yes. A REST API v1 with Bearer authentication and webhooks (including a report.released event) is documented at /docs/api. There is no LMS standard integration such as LTI, so integration is via the API.
What stops a result reaching someone before we are ready?
The release gate. Respondents see nothing - no score, no report, no PDF - until an assessor releases the result. The release records who released it and when, and the action is written to an audit log.
Related resources
Other use cases
See it on your own assessment
Start free, build one assessment, and see the AI-drafted feedback for yourself.