Assess a project

Score one grant, end to end.

Paste a title and description — or let the assistant draft the whole assessment from text, a PDF or a Word file — then refine every field. The result updates live, with an uncertainty band.

New here? See the whole model filled in and running.
✦

Assistant — auto-draft the assessment from a description

fills the model & runs it
Paste a project title and description, or upload a PDF / Word / text file. The assistant matches it to an objective and its binding constraints, drafts contribution, plausibility, theory-of-change and enabling conditions, fills the form below, and runs the model. It’s a transparent keyword-based first draft — review every field before relying on it.
Advanced — use an AI model instead (optional; your key stays in your browser)
With a key, the assistant sends the text to your chosen model for a higher-quality read, then validates the result against the taxonomy; if the call fails it falls back to the built-in matcher. The key is used only for direct browser calls and is never stored or sent anywhere else.
▲

Closing the gap — live impact visualization

updates as you fill the form
Select an objective to visualize the gap.
1

Project

Step 1
The award or budget behind the project. Used for the cost-effectiveness read-out (global significance per $ million); it does not change the impact score itself.
2

Ultimate objective & counterfactual

Steps 2–4
5

Binding constraints — what blocks the objective

Steps 5–6
Each binding constraint carries a wedge share: how much closer to the objective we get if it is fully removed. Shares sum to 100% per objective. Tick the constraints this project works on — their shares sum to the project’s impact wedge.
Binding constraintWedgeShare
7

Per constraint — contribution, theory of change, plausibility

Steps 7–10
For every constraint you ticked: how much of it this project removes if successful (contribution %), the theory of change in words, and how credible that pathway is (plausibility 0–100).
How Total Projected Impact is computed (step 10)
For each selected constraint: (contribution% ÷ 100) × wedge share × (plausibility% ÷ 100), summed across constraints. Units: percentage points of the counterfactual-to-objective gap this project is expected to close.
11

Enabling conditions — what success depends on

Steps 11–12
List the conditions outside the project’s control that must hold, each with a % chance. Instead of a crude min/max, the tool treats them as a chain of dependencies and runs a quick simulation — giving a mean and a P10–P90 band. More dependencies lower the expected impact; the mean drives the score.
Uncertainty model: 0%