Score and shortlist
What the board shows#
Every startup in the batch with its current status, AI Total Score, Jury Score and key finding. Statuses tell you where the batch actually is without opening anything:
| Status | Meaning |
|---|---|
| Ready | Evaluated, no Jury Score yet |
| In review | A reviewer is working on it |
| Scored | A Jury Score has been submitted |
| Blocked | Something prevents evaluation or scoring |
Four steps from report to ranking#
Review the summary, the AI Total Score, and the findings across your criteria. The AI number is context — it is read-only and it does not rank anyone.
Use the criterion-linked questions during the live session and keep notes beside the evidence they belong to. Notes that live next to the finding survive the meeting; notes in a separate doc do not.
Score each criterion from 0.0 to 10.0. Your score is a separate value from the AI Criterion Score for that dimension — both stay visible, side by side.
Once every startup is scored, generate the ranking. It is built from submitted Jury Scores weighted by your criteria.
Review signals: where to look closer#
The board flags three cases that repay a second read:
| Signal | What to check |
|---|---|
| Judges disagree | Open the split and compare the reasoning before you score |
| A high score has weak evidence | Open the findings and check what actually supports it |
| A lower-ranked startup fits your criteria | Pull up the evidence and give it a closer read |
The third one is the reason a batch view beats a sorted list. A team that lands mid-table on the advisory score can be the one that matches your thesis best, and the only way to notice is to have the criteria and the evidence in the same view.
Comparing before deciding#
Line every startup up against the same criteria and sort by total or by any single dimension. Shortlist from inside the board — moving the decision into a spreadsheet is where the evidence gets left behind.
The decision trail#
What a run leaves behind is as much the point as the ranking:
- The AI Total Score, as context, per team.
- The Jury Score per dimension, as the verdict.
- The findings, quotes and slide references behind both.
- Where judges disagreed, and what was decided about it.
This is what turns "why did that team win?" from an archaeology dig into a five-minute reply. Whether you run open appeals stays your policy; the record is what makes either policy workable.
Independent scores first#
One procedural rule is worth keeping regardless of the tool: every judge submits scores before deliberation, and the chair speaks last. Research on live review panels found scores shifting with the room — with the chair's interventions and even laughter measurably moving numbers. Discussion should resolve flagged conflicts, not manufacture agreement.
Next steps#
- Disagreement and spread — the thresholds behind the conflict flag.
- Brief your jury — how to get agreement before scoring rather than after.
- What EvalLens does not do — the four boundaries to state before the awards night.