Skip to content
IdeaScore
Evaluation

Scoring Your Committee Can Defend

Evaluators form their own view before they see anything IdeaScore AI wrote. Judge each round its own way, build a panel in minutes, and know whether scoring will finish in time.

Evaluators form their own view first. Judge each round its own way, build a panel in minutes, and know when scoring will finish.

ways to judge a round
5
choices for what evaluators see of IdeaScore AI
3
of scoring before a forecast appears
3days
window Reminder Effect watches after a first reminder
48h

How Evaluation Works

Fair scoring

Fair by default, and you can show why

IdeaScore AI's notes stay hidden from evaluators unless you choose otherwise. Normalise evens out harsh and lenient scorers, and shows who moves and why.

  • Hidden, shown after they submit, or always shown
  • Blind rounds never show IdeaScore AI's notes
  • Raw and normalised order side by side, with the reason a team moves
Seed Fund 2026 · Cohort III

Fair Scoring

Rank, raw to normalised

  • 1Halcyon Gridwas 1=
  • 2Quorum Biowas 31

    Rises 1: both its evaluators score harsher than the panel

  • 3Farmlightwas 21
  • 4Tessel Labswas 51
  • 5Orrin Healthwas 41

How each evaluator leans

  • Dr. S. Banerjee1.8 harsher
  • Prof. R. Bose1.2 more lenient
  • M. Rao, investor0.3 harsher
IdeaScore AI's notes for evaluators:Hidden (default)
Judging modes

Judge each round the way it should be judged

Rubric scores, ranking, head-to-head, live pitch or public vote, chosen per round. Rubrics carry weights and descriptors so a panel scores alike.

  • Head-to-head is fair once every entry has been compared five times
  • Rank by dragging, with arrow buttons or from the keyboard
  • Public vote, available once your organisation turns it on
Seed Fund 2026 · Cohort III

How this round is judged

Edit round · Semi-final

Fixed once rankings, comparisons or votes exist.

IdeaScore AI

A panel for the whole round in minutes

IdeaScore AI proposes evaluators for each team with the reason for every match, after conflicts and full loads are ruled out. Accept, swap or skip, then save.

  • Each person's load before and after, with a warning before anyone goes over
  • Accept all without a conflict, in one click
  • Names are swapped for labels before anything reaches IdeaScore AI
Choose the teams, and IdeaScore AI proposes evaluators for each one, with the reason. Review, swap or skip, then save.
Assignments

A plan ready the moment you open it

The plan is ready when you open it: who would score what, and why, with every declared conflict kept apart. Check, apply, and undo in one step.

  • A locked cell for each declared conflict, with who declared it and why
  • Submitted scores never move
  • Open declarations, each with a one-click reassign
The Conflict-Aware Assignment card on Meridian Seed Fund 2026 with its plan drawn: evaluators on the left joined by lines to teams such as CraftEnergy and TrueSense on the right, a lock on a conflict kept apart, 130 assignments with 73 by expertise and 57 by lightest load, open declarations below and an Apply 130 assignments button.
Conflict-aware assignment has its plan ready when the page opens: who would score what, and why, with conflicts kept apart.
Blind review

Blind review that holds

In blind rounds of scholarships, fellowships and awards, reviewers see case numbers, and written answers reach them with names removed.

  • Case numbers in place of names on every reviewer's list
  • Names removed from written answers before a reviewer reads them
  • An entry stays blind for the reviewers who judged it blind
A reviewer's Merit review list on Meridian Merit-cum-Means Scholarship: applicants shown as Case #0103, Case #0104, Case #0107 and more, each with the level of study, Due in 12 days, Pending and a Score button.
On a scholarship's blind Merit review, a reviewer sees each applicant by case number.
Deck masking

Decks without names, contact details or photos

Evaluators open decks with names, contact details and photos taken out of the file. Preview a real slide first; staff and teams keep the original.

  • Text removed, so it cannot be copied or searched for
  • Decks with words inside pictures held back until you choose
  • Pages in a script IdeaScore can't check held back from blind reviewers
The teamTidewell

Who is building Tidewell

At our Hosur pilot, March 2026

Kabir Sethi

Co-founder and CEO

kabir@tidewellwater.in

Nandini Joshi

Co-founder and CTO

linkedin.com/in/nandinijoshi

Advised by Dr Farah Siddiqui and Arvind Menon2

As uploaded

1 / 3

Programme settings · Cohort III

Mask decks for evaluators

Off. Evaluators see every deck as uploaded.

On this slide: 4 names, 2 contact details, 1 photo

The original deck is kept. You, your staff and the team always see it as uploaded.

Switch masking on and page through a deck as an evaluator sees it.

Rounds

Promote with the line in plain sight

Promote by top N, a cut-off or by hand, with ties at the line pointed out first. Committee mode runs the selection meeting, one team at a time.

  • Ask a round's teams for a pitch deck, by a date
  • Share criterion averages with each team after its outcome
  • In committee, the chair confirms each outcome and dissent is recorded
The Rounds page for Meridian Seed Fund 2026: Screening, Semi-final and Final with their teams, and the Screening standings below.
Every round side by side, with standings and one click to promote.
Review Forecast

Will scoring finish before the panel meets?

A finish date from each judge's own pace, who is holding it up, and Remind or Reassign on the card. Reminder Effect shows whether reminders work.

  • Appears after three days of scoring, so it is never a guess
  • Past the deadline, the line turns amber and says Finishes late
  • Reminder Effect shows counts only, never an evaluator's name
The Review Progress Forecast card: at this pace scoring finishes on Wednesday 18 November, with 52 of 72 reviews scored, a burn-down chart and the four judges holding it up.
Will scoring finish before the panel meets? A forecast from each judge's own pace, with who is holding it up.
IdeaScore AI

Better feedback, and disagreements explained

Turn a judge's own notes into kind, specific feedback, and see why two judges disagree, with sources and one neutral question. It never says who is right.

  • Feedback reads only the judge's own scores, notes and the rubric
  • Judges are sent as Judge A and Judge B; IdeaScore puts names back
  • Why do judges disagree? is for admins; evaluators never see it
HydroLoop's dossier on the Criteria tab, with the Why do judges disagree? helper above the score breakdown and the line Free, reads judges' scores and notes on this team.
Why do judges disagree? sits beside the scores, and says in plain words that it reads judges' scores and notes on this team.
For evaluators

Evaluators open on the work

Up Next shows what to score first and the next deadline. Scores are kept safe on a weak connection, and contact details stay hidden by default.

  • Declare a conflict from the list or the scoring page
  • A submitted score is read-only until Edit my score
  • Flag a criterion for the committee to discuss
An evaluator's dashboard: Up Next lists SwiftAI and NeoWorks to score, At a Glance shows 46 of 48 scored and the next deadline, with two programmes below.
Evaluators open on what is waiting for them, with the next deadline and their programmes below.
Fair scoring

Fair by default, and you can show why

IdeaScore AI's notes stay hidden from evaluators unless you choose otherwise. Normalise evens out harsh and lenient scorers, and shows who moves and why.

  • Hidden, shown after they submit, or always shown
  • Blind rounds never show IdeaScore AI's notes
Seed Fund 2026 · Cohort III

Fair Scoring

Rank, raw to normalised

  • 1Halcyon Gridwas 1=
  • 2Quorum Biowas 31

    Rises 1: both its evaluators score harsher than the panel

  • 3Farmlightwas 21
  • 4Tessel Labswas 51
  • 5Orrin Healthwas 41

How each evaluator leans

  • Dr. S. Banerjee1.8 harsher
  • Prof. R. Bose1.2 more lenient
  • M. Rao, investor0.3 harsher
IdeaScore AI's notes for evaluators:Hidden (default)
Judging modes

Judge each round the way it should be judged

Rubric scores, ranking, head-to-head, live pitch or public vote, chosen per round. Rubrics carry weights and descriptors so a panel scores alike.

  • Head-to-head is fair once every entry has been compared five times
  • Rank by dragging, with arrow buttons or from the keyboard
Seed Fund 2026 · Cohort III

How this round is judged

Edit round · Semi-final

Fixed once rankings, comparisons or votes exist.

IdeaScore AI

A panel for the whole round in minutes

IdeaScore AI proposes evaluators for each team with the reason for every match, after conflicts and full loads are ruled out. Accept, swap or skip, then save.

  • Each person's load before and after, with a warning before anyone goes over
  • Accept all without a conflict, in one click
Choose the teams, and IdeaScore AI proposes evaluators for each one, with the reason. Review, swap or skip, then save.

More on evaluation

7 more

Everything in Evaluation

  • Rubrics

    Criteria with weights and descriptors for a low and a high score; one rubric per round.

  • Blended leaderboard

    Human, IdeaScore AI or a blend at the weight you set; the Human view never includes IdeaScore AI.

  • Normalise

    Each evaluator's scores rescaled by their own average and spread, so harsh and lenient count the same.

  • Compare rounds

    With more than one round, compare two leaderboards to see who moved.

  • Committee mode

    Members vote Accept, Waitlist or Decline from a laptop or phone; the report marks dissent.

  • Evaluator honoraria

    Set rates, record payments in a ledger and issue statements.

  • Remind evaluators

    Remind about overdue reviews from the planner or the forecast, once in any 24 hours.

  • Capacity

    Set how many teams each evaluator can take, and filter for anyone over it.

  • Blind rounds

    Case numbers in place of names, and names removed from written answers, in scholarships, fellowships and awards.

  • Evaluator privacy

    Team contact details and the leaderboard hidden from evaluators unless you show them.

  • Weak connections

    Scores are kept on the evaluator's device and sent as soon as the connection allows.

  • Feedback to teams

    Criterion averages and reviewed comments shared after a team hears its outcome, never another team's.

  • Earlier decks restored

    A replaced pitch deck is kept for 30 days and can be restored.

  • Leaderboard export

    The ranking as a CSV file or an Excel workbook.

Questions Committees Ask

Never. Its notes are hidden from evaluators by default; you can show them after each evaluator submits their own score, or from the start. Blind rounds always hide them, and notes that state a score, rating or verdict are left out even when notes are shown.
Each evaluator's scores are rescaled by their own average and spread, a z-score, so a harsh and a lenient evaluator count the same. The Fair Scoring card shows each team's rank before and after, and why it moved.
Blended mixes IdeaScore AI's score with the panel's at the IdeaScore AI weight you set. The Human view never includes it, and AI ranks by IdeaScore AI's score alone.
No. Suggest evaluators proposes a panel with the reason for each match, after IdeaScore has ruled out conflicts and full loads. Nothing is assigned until you save, and one Undo removes the whole change.
The names of the people in your programme, email addresses, phone numbers and other contact details, and every photograph. The text is taken out of the file, not covered over. Decks with words inside pictures are held back from evaluators until you choose, a page in a script IdeaScore can't check for names is held back from blind reviewers, and staff and the team always see the original.
In scholarships, fellowships and awards, reviewers in a blind round see each applicant by case number. Written answers reach them with names removed, IdeaScore AI's notes stay hidden, and an entry stays blind for every reviewer who judged it blind.
Yes, until the round closes: Edit my score reopens it as a draft, and it counts again once submitted. After the round closes, scores and comments are read-only.

See IdeaScore Run a Call with Your Own Rubric

A 30-minute walkthrough with a product specialist, using your programme's form and criteria.

IdeaScore