Score Documentation

Go to Score > Models > New Model. Name your model (e.g. "Graduate Recruitment 2025"), choose a model type — Tender, Recruitment, Grant, or Custom — and define your evaluation criteria. Models can be reused across multiple evaluation rounds.

Add criteria and set a percentage weight for each. All weights must total 100%. For each criterion, write a clear description of what you're looking for — the AI uses your descriptions to evaluate submissions, so specific language produces better scores.

Upload submissions as PDF or Word documents (up to 50MB each) from Score > Submissions > Upload. Drag and drop multiple files at once for bulk upload. Score accepts documents in English by default; contact support for other language support.

The Tender model type includes pre-built criteria categories: Methodology, Demonstrated Experience, Price, and Risk Management. Adjust the weightings to match your procurement policy. Scores are calculated per criterion and combined into a total weighted score for each submission.

The Recruitment model type is designed for screening candidates against a role description. Upload the job description and Score will suggest starting criteria extracted from the document automatically. You can accept, edit, or remove any suggested criterion before running the evaluation.

Grant models include pre-built criteria for Alignment with Objectives, Impact, Feasibility, and Budget Justification. Each criterion includes a confidence indicator — low-confidence scores (below 70%) are flagged automatically for human review before finalisation.

Use a Custom model when none of the pre-built types fit your use case. Start with a blank rubric and define all criteria from scratch. Custom models support up to 20 criteria and any scoring scale (e.g. 0–5, 0–10, or 0–100).

Once submissions are uploaded and your model is configured, click Run Evaluation. Score reads each submission against your rubric and assigns a score per criterion with a plain-language justification. Processing takes 30–90 seconds per submission depending on document length.

Any AI-generated score can be manually overridden. Click the score field to edit it directly and add a note explaining the override. All overrides are recorded in the audit trail with the evaluator's name and timestamp.

Each AI-generated score includes a confidence percentage indicating how clearly the submission addressed that criterion. Scores below 70% confidence are highlighted in amber — these should be reviewed by a human evaluator before being relied upon in a final decision.

The Comparison tab shows all submissions ranked side-by-side by total weighted score. Filter by individual criteria to see where submissions diverge. Tied submissions are shown in the order they were uploaded.

Export to PDF or Excel from Score > Export. PDF exports include the full rubric, all scores, AI justifications for each criterion, manual override notes, and a summary ranking. Excel exports contain raw scores per criterion for further analysis in a spreadsheet.

Generate a shareable read-only link from Score > Share. Recipients can view results and justifications in a read-only web view without a SuiteCorn account. Shareable links expire after 30 days and can be revoked at any time from Score > Share > Manage Links.