Reviewer’s Mind by ProPil
The five criteria a real evaluator applies — used on your proposal before you submit it.
This is not a general AI reading your draft and telling you what it thinks. It is the evaluation method Paulius Yamin uses as a funding evaluator and proposal specialist, run on your proposal line by line, with every finding anchored to your own words.
One full review — $19
One review is one complete five-criterion Red Team analysis of one proposal. The review is generated by AI using this method; no human expert reads your proposal.
Whose method this is
Paulius Yamin
The review encodes how he reads a proposal from the evaluator’s side of the table — the questions a panel asks, and the answers a proposal has to already contain.
- Behavioral scientist
- 15 years writing, coordinating and evaluating funding proposals across academia, international development and consulting
- $18M+ secured in funding
- Evaluator for European Research Council (ERC) Consolidator and Synergy Grants
- Member of the Assessment Panel, ESRC Large Grants for Ambitious Research in the Social Sciences (UK)
- Worked through institutions including LSE, the University of Pennsylvania, the Harvard Kennedy School and the Paris Institute for Advanced Study
- Consulted with organizations including the World Bank, USAID, OECD, ILO and UN agencies
Why not just ask any AI?
Because a generic read and an expert method are not the same thing.
Five fixed criteria, not a fresh opinion each time
A general AI gives you a fluent, generic read that changes every time you ask. This review runs the same five criteria an evaluator applies, in the same order, every time.
Every finding anchored to a verbatim quote
No objection is allowed to float. Each one carries the exact sentence from your proposal that caused it, so you can judge the finding rather than trust it.
Deterministic checks, not vibes
Contradictions between sections and claims with no supporting evidence are caught mechanically, not noticed by chance.
It refuses to invent problems
Before a criticism is shown, it is re-checked against the full text. If your proposal already answers it, the criticism is dropped. A general AI invents problems as readily as it finds real ones.
It finds what an expert reviewer would flag — and refuses to invent what isn’t there.
The five criteria
Goal, method and impact coherence
Whether the aim you state, the method you choose and the impact you promise actually support each other — or quietly ask the reader to accept a jump between them.
Delivery realism
Whether the plan, timeline, team and resources make the work achievable exactly as described, not just plausible in outline.
Relevance beyond the discipline
Whether the significance still holds for a panel member who does not share your specialism, which is who usually decides.
Researcher and host fit
Whether you and the host environment are the credible place for this specific work, and whether the proposal shows it rather than asserts it.
Call and instrument fit
Whether the proposal answers this call and this funding instrument, or reads like a good project sent to the nearest available scheme.
Sample review — illustrative example on a fictional proposal
The proposal excerpt and the findings below are invented for illustration. They are not a real client, a real proposal or a testimonial.
Fictional proposal excerpt
“The project will demonstrate that participatory budgeting increases citizens’ trust in local government. Working with four municipalities, we will run participatory budgeting cycles over 18 months and survey participants at the end of the process. The approach has been shown to reduce administrative processing times by up to 40%, and the consortium will disseminate the results widely to policy audiences across the region.”
“The project will demonstrate that participatory budgeting increases citizens’ trust in local government.”
Why it doesn’t hold: The only trust measure in the proposal is a post-intervention self-report survey in participating municipalities. A perceptual measure collected once, from people who took part, can show that participants report more trust — it cannot demonstrate that the mechanism caused it, and it cannot separate the effect from the attention of being included in a pilot.
You must answer: What comparison and what timing would let this design distinguish a real change in trust from a favourable response by participants who were selected and consulted?
“The approach has been shown to reduce administrative processing times by up to 40%.”
Why it doesn’t hold: The figure is presented as a property of the approach, but nothing in the proposal generates it: the cited source is a national reform programme of far larger scale, with a different baseline and a different administrative system, and the proposal never states what the processing time is in the target municipalities today.
You must answer: What is the current baseline processing time in the target sites, and what magnitude of change does this project itself commit to achieving and measuring?
Fictional example. A full review covers all five criteria across the whole proposal.
What the review gives you
An assessment against the call's own criteria
Your proposal is read against the five fixed criteria above, adapted to the funding programme and instrument you are applying to.
Findings anchored to your own words
Every objection quotes the passage that caused it, so you can see exactly what an evaluator would question and check it yourself.
An overall assessment
A clear verdict on how the proposal currently stands, with the issues ordered by how much they would weigh on a funding decision.
Concrete improvements
Specific, actionable changes for each issue raised, so the feedback turns into a revision you can actually make.
How it works
Add your proposal
Upload the draft you intend to submit.
Add the call
Upload the call documents and name the funder and programme, so the review judges your proposal against the right standards.
Run the review
ProPil reads the full proposal, checks each criterion, and verifies every finding against what your text actually says.
Read the report and revise
Work through the findings and the suggested improvements. When you have revised, you can review the new version again.
No review can promise you the grant.
Competitive funding depends on strong competitors, limited budgets, evaluation criteria and human judgment. Reviewer’s Mind does one thing: it shows you what needs improving in your proposal while you can still change it.
See the weaknesses before the evaluator does.
One full review — $19
Reviewer’s Mind by ProPil · powered by the Funding Proposal Secrets Method