A full-cycle program officer tool

See what the rule does before applicants do.

Compile eligibility prose into readable tests, run those tests against fictional or user-supplied profiles, project reviewer and applicant burden, and carry the same evidence discipline into proposal diligence.

Mode 1

Call Design

Needs a draft rule

Draft eligibility rule Or upload the draft rule TXT, Markdown, PDF, or DOCX. Maximum 6 MB. Or load a saved RuleSpec or SessionBundle JSON only. Older files are migrated with a visible warning. Files from a newer major schema are refused.

Candidate profiles optional

Use the fictional starter set, upload up to 500 past applicants by CSV, or add profiles manually. Profiles are evaluated locally by deterministic code after the rule is compiled once.

Upload CSV

Funnel and burden inputs

Total grant pool ($)Tool default: $10 million Grant size ($)Tool default: $500,000, producing 20 grants Expected applicationsTool default: 1,000 Minutes per first readTool default: 20 minutes Advanced review hours per admitted applicationTool default: 8 hours Expected admit rateTool default: 20% when no profiles are loaded Loaded staff cost per hour ($)Tool default: $65 Applicant hours per applicationDefault source: published sector estimate, user-editable Loaded applicant labor cost per hour ($)Default source: published sector estimate, user-editable

You keep the continuity. This server does not save the rule, profiles, or results. Download the RuleSpec or session bundle before leaving.

Call impact report

A rule is a filter.

Compile a draft to see each test, language that cannot become a test, who the rule excludes, filing-thin effects, and the burden created by the funnel.

Compiled RuleSpec

Readable evaluator

Compiled clauses

Source, fact, operator, value, and consequence remain visible.

Values and ambiguity

Uncompiled language

Language the tool cannot state as a testable fact stays here.

Candidate test

Exclusion report

Organization	Outcome	Reason code	Deciding clause

Retrospective rule test

Clause frequency

Each clause is tested independently against every supplied profile.

Clause	Would exclude	Missing fact	Source language

Filing-thin comparison

Rule Impact Report

Mandatory clauses only

Self-screen question set

Copy-ready HTML

Visible arithmetic

Funnel and burden projection

Fields to capture

Instrumentation plan

Method

The evaluator is inspectable code.

Open the complete fictional worked example to inspect the compiled rule, twelve-profile exclusion report, impact report, self-screen, and funnel projection without submitting data.

Compilation is constrained

The model may translate testable prose into a RuleSpec. Values language, quality judgments, and instructions addressed to the model remain uncompiled.

Eligibility is deterministic

The same RuleSpec and applicant facts produce the same status, reason code, and clause reference.

Self-screen parity: 100/100, boundary-value suite, run 2026-09-05.

Determinism: 20/20 repeated evaluations, run 2026-09-05.

People retain authority

The tool does not recommend awards, score quality, judge community representation, or predict funding probability.

Current boundary: the tool does not run a live intake workflow or produce cohort reports from live submissions. Those functions require real call data and remain outside this release.

See what the rule does before applicants do.

Call Design

A rule is a filter.

Compiled clauses

Uncompiled language

Exclusion report

Clause frequency

Rule Impact Report

Self-screen question set

Funnel and burden projection

Instrumentation plan

Diligence Memo

Evidence before inference.

Claim versus filing

Financial diligence questions

Fit against stated strategy

Next diligence actions

The evaluator is inspectable code.