The AI Reviewer

What it scores

For each draft (LOI or full proposal), the Reviewer evaluates seven criteria that mirror how external reviewers actually score applications:

• Alignment: does this match the funder's stated priorities?
• Need: is the problem clear, urgent, and well-evidenced?
• Approach: is the methodology specific and credible?
• Capacity: can the org realistically deliver?
• Outcomes: is the evaluation plan measurable?
• Budget: are costs reasonable and clearly justified?
• Sustainability: what happens after the grant ends?

Each criterion gets a sub-score; the overall 0 to 100 is a weighted average tuned to the funder type (federal funders weight outcomes and budget heavily; foundations weight alignment and need).

Reading the score

Use these rough bands as a gut check:

• 85 to 100: ready to submit. Final read for typos and send.
• 70 to 84: solid, with one or two specific sections worth a revision pass.
• 50 to 69: not ready. The Reviewer's suggestions list will show what's weak.
• Under 50: structural problems. Often easier to regenerate the draft with a sharper prompt than to keep editing.

The score is calibrated against thousands of historical applications, not a hard cutoff. A 68 from a funder that usually scores you in the 60s is still worth submitting; a 72 from a funder where you usually land at 85 is worth another pass.

Line-level suggestions

Click any sub-score to see the specific suggestions for that criterion. Each suggestion is anchored to a paragraph in the draft and shows:

• What's weak.
• Why it's weak (which reviewer expectation it misses).
• A proposed rewrite you can accept, edit, or ignore.

Accepting a suggestion replaces the highlighted paragraph in the draft and re-scores the section. Most users accept 30 to 50% of suggestions and dismiss the rest. The goal isn't a perfect score; it's catching the things you'd miss on your own.

Disqualifier flags

Above the score, the Reviewer surfaces hard disqualifiers it caught in the draft against the NOFO or guidelines. Common ones:

• Eligibility mismatch (you said “501c3” but the NOFO requires “public agency”).
• Geography mismatch (your service area falls outside the funder's defined region).
• Use-of-funds mismatch (you described capital costs in a program-only grant).
• Required-elements missing (logic model, sustainability plan, indirect-cost cap).

A disqualifier doesn't automatically lower the score, but you should fix it before submitting. Most applications that get screened out before review fail on a disqualifier, not on writing quality.

How to use the Reviewer well

• Run it early. Run a first review after the drafter generates and you've done one editorial pass. Catches structural problems while they're still cheap to fix.
• Run it again after the budget. Budget changes ripple into the budget-narrative score; a re-review catches inconsistencies you introduced.
• Run it one last time after final edits. The final score plus the disqualifier check is the submission gate.

Limits

The Reviewer scores against the funder's published priorities. If a program officer told you in a phone call that their unwritten priority this cycle is rural projects, the Reviewer doesn't know that. Capture call notes in the pipeline (Pipeline & deadlines) and weight them yourself.