Civic & Government /Decisioning interactive demo field guide · 8 min

Grant Eligibility Checker

Removes ~40% of out-of-scope applications before they hit the assessor queue.

theater/demos/civic-gov_grant-eligibility-checker.html · sandbox · read-only

Open

FIG. 1

The live demo, running on fabricated data. Open it to step through the full flow — every output is shown for a person to approve before anything happens.

How it would work

Reads each grant application against the program's published eligibility rules, flags every criterion that isn't met with the evidence behind it, and hands the decision to a grants officer to approve before anything reaches the applicant.

Input 01

The application + the rules

A submitted grant application — incorporation details, location, insurance certificate, requested amount, project dates, prior-grant history — whether digital PDF or scanned, plus the program's eligibility guidelines.

Agent 02

Extracts, checks, flags

Pulls the relevant fields off the form, checks each against a program criterion — incorporation, location, not-for-profit status, insurance cover, amount cap, acquittals, project timing — and marks each met or not met.

Output 03

A checked application, working shown

Every criterion with its result and the document it was read from, a drafted request-for-information letter for the gaps, and a Recommend / Request info / Decline choice that a grants officer makes — the tool never decides funding.

Where it works well

It runs the full eligibility checklist on every application, the same way, with the evidence attached.

Done by hand it is fifteen to thirty minutes per application, and the boring criteria — insurance minimum, acquittal status — are the first ones skimmed at volume.
Best for a council or foundation running open rounds: dozens to hundreds of applications screened against one published ruleset before a panel sees them.
Clearing the out-of-scope applications early gives the officer their time back for the judgement calls — merit, fit, community benefit — not the checklist.

The slow, invisible cost of a grant round is the screening — re-reading every application against the same dozen criteria, chasing the one insurance figure or overdue acquittal that makes it ineligible, while the genuinely fundable ones wait in the same queue.

Where it works badly

It is confidently wrong when the rules it's checking against are stale, or the document it read is ambiguous.

Weak on anything that isn't a hard, checkable fact — whether a project genuinely serves the community, whether an auspice arrangement is legitimate. It should flag for a human, not rule.
Scanned and handwritten forms are where extraction errors hide; a misread figure becomes a wrong eligibility result that looks just as confident as a right one.
If your program's "eligibility" is mostly discretionary merit rather than checkable criteria, the tool screens the easy 20% and leaves you the part that was always the work.

The honest test

If you cannot say, today, which version of the guidelines this is checking against and that the figures were read from a legible source — the tool makes your wrong rejection faster, not fairer.

Point it at last year's guidelines after the insurance minimum changed, or a scanned certificate where the cover amount is smudged, and it produces a clean, authoritative "not met" — or "met" — against the wrong number. The output looks more certain than the evidence behind it.

What it doesn't do — and shouldn't

It flags. A grants officer decides. That boundary is deliberate.

WHAT IT DOES

Marks each criterion met or not met against the published rule

Shows the document and field each result was read from

Drafts a request-for-information letter for the unmet items

WHAT IT WON’T

Approve, decline or rank any application

Decide whether a project merits funding

Notify the applicant on its own say-so

A grant decision by a Victorian council is a public, reviewable act — subject to the Freedom of Information Act 1982 and the council's governance obligations, and a refused applicant can ask why. An eligibility call wrongly made by a tool, with no person who stands behind it, is exactly the decision that doesn't survive that scrutiny. The accountable officer stays on the decision because the consequence lands on the council, not the software.

What your data has to look like

Eligibility criteria written as checkable rules, and application fields the tool can actually read.

Typical readiness

across orgs we see, before the first job

Program guidelines as discrete, checkable criteria

Needs shaping

The current round's rules, and a way to know they're current

Usual weak point

Application fields in structured, legible form

Usual weak point

Prior-grant and acquittal history, linked to the applicant

Needs shaping

Supporting evidence attached, not just asserted

Usual weak point

The real first job

Most programs hold their eligibility rules as prose in a PDF and take applications as uploaded scans. Turning the guidelines into discrete checkable criteria, and capturing applications as structured fields with evidence attached, is usually the real first job — larger and more valuable than the checking layer on top. Once the rules and the intake are clean, every application after that is screened faster and the same way.

Right fit if…

You run open grant rounds with dozens to hundreds of applications against a published ruleset

Most of your eligibility is hard, checkable criteria — incorporation, insurance, amount, acquittals

You can point to the current round's guidelines as the version in force today

Applications arrive as structured online forms with evidence attached

Walk away if…

Eligibility is mostly discretionary merit, not checkable criteria

Your guidelines live as last year's PDF that nobody has reconciled to this round

Applications come in as scans or handwritten forms with no structured fields

You want a tool that approves or declines applicants for you

Open questions

The worried-buyer questions, answered straight

It can mark a criterion wrongly — which is exactly why it never rejects or approves anyone. For each application it shows every criterion as met or not met, the rule it checked against, and the document the figure was read from, then drafts a request for the missing information. A grants officer reviews that and makes the call. The tool surfaces the evidence; a person at the council stands behind the decision.

That is exactly where it is weakest, and where it should be honest about it. The tool can read a scanned certificate, but a smudged or handwritten insurance figure can be misread, and a misread figure becomes a confident wrong result. It works best on typed, structured online-form fields. Moving intake to a structured form with attached evidence is usually the first piece of work, and the piece that pays off across every application after.

No. It removes the repetitive screening — checking the same dozen criteria across every application — so the officer spends their time on the judgement the role exists for: whether a project genuinely serves the community, whether the funding is the right fit, how to weigh competing requests. The decision stays theirs. The capacity it frees goes back into the assessment that actually needs a person.

Current to the round you are assessing. Council and foundation grant rules change between rounds — an insurance minimum, an amount cap, an acquittal requirement — and the tool checks against whatever version it is given. Point it at last year’s guidelines and it produces clean, authoritative results against the wrong rule. The honest test: do you know, today, which version of the guidelines this round is being checked against?

Grant applications hold personal and organisational information that, for a Victorian council, is covered by the Privacy and Data Protection Act 2014 and its Information Privacy Principles. Any deployment runs against your own systems and data handling, not a shared pool — we scope where the data sits and who can access it as part of the build. The demo here runs entirely on fabricated data; Broadmeadows Community Hub Inc. and its applicants are invented.

Yes — that is the point of showing the working. For every application the tool records each criterion, its result, and the document the result was read from, so an officer can explain precisely which requirement was unmet and on what evidence. That trace matters because a council’s grant decisions are reviewable under the Freedom of Information Act 1982 and a refused applicant can ask for reasons.

What it takes to build

3–4 weeks · 4 phases

Reused from template~70%

Bespoke to this skin~30%

stack · Claude · audit log · review UI

What it would cost

Fixed scope, fixed price, fixed dates.

Bite-sized first piece

One contained change, low risk

Pilot build

Most builds land here

Embedded support

Scale on proof

Considering this for your grants program?

The honest place to start is a bite-sized first piece — one round, one ruleset, low risk. Tell us where the screening hurts; we'll play it back, scope it, and show you what's possible.

Book a call How we work