Grant & Proposal Writing is a Claude AI skill — researches funder priorities before drafting a word, so the proposal answers the questions the reviewer actually has.
A program officer at a mid-sized education foundation once described her review process like this: she could sort the applications she was reading into two piles within the first paragraph — not funded and maybe — based on a single signal. Not the quality of the programme being described. Not the strength of the organisation. Whether the applicant had clearly read the foundation's current strategic framework before writing.
She wasn't looking for flattery or name-dropping. She was looking for evidence that the applicant understood what the foundation was trying to accomplish right now, in this funding cycle, with this cohort of grantees. Most applications, she said, described strong work in the abstract. Fewer than a third showed any sign of having thought carefully about why this foundation, at this moment, would be the right funder for that work.
That gap — between describing your work and making the case for why this funder should fund it right now — is where the majority of strong proposals lose.
The Four Tells That Mark a Generic Proposal
These aren't obscure stylistic problems. They're structural patterns that appear in the same form across hundreds of applications in any given review cycle, and experienced reviewers recognise them immediately because they all point to the same root cause: the proposal was written without prior research into the funder.
Tell one: the needs statement describes a general sector problem, not a funder-specific gap. Generic needs statements open with statistics about the field — dropout rates, unmet need figures, national prevalence data — that are technically accurate and completely disconnected from the specific gap this funder has identified in their own strategic documents. The reviewer knows what the sector problem is. What they're reading for is whether you understand which slice of that problem they've committed to addressing.
Tell two: the outcomes don't map to the funder's theory of change. Every funder has a model — explicit or implicit — of how the work they fund produces the change they want to see. A proposal that lists outcomes in a generic logic model format, without referencing the causal pathway the funder believes in, reads as written by someone who hasn't studied how this funder thinks about impact. The outcomes may be real and measurable. They're just not speaking the funder's language.
Generic proposals describe excellent work to the wrong audience — they're written for any funder in the category, which means they're optimised for none of them.
Tell three: the organisational capacity section emphasises the wrong credentials. Capacity sections in generic proposals list everything: years in operation, staff count, programme history, awards received. Funder-aligned capacity sections lead with the specific credential this funder has historically weighted most heavily — whether that's evidence of community relationships, financial management track record, prior grant performance, or subject-matter expertise. The difference is research: knowing what this reviewer is actually assessing in the capacity section before writing it.
Tell four: the budget narrative doesn't reflect the funder's funding philosophy. Some funders are deeply uncomfortable with overhead above a certain percentage. Others actively want to see organisational capacity investment. Some require specific cost categories; others read a prescriptive budget as a lack of organisational maturity. A budget narrative written without knowing the funder's philosophy lands wrong — not because the numbers are off, but because the framing doesn't match what the reviewer expects.
That gap is exactly what the Grant & Proposal Writing skill for Claude was built to close.
Why This Happens Even When Writers Are Trying Hard
The tells above don't appear because people aren't working hard enough on their proposals. They appear because funder research is genuinely time-consuming, and most grant writers are operating under deadlines that compress the preparation phase. The instinct is to start writing — to make visible progress against the deadline — rather than spend the first half of the available time reading the funder's recent 990s, annual reports, current RFP language, and publicly announced awards before drafting a word.
The instinct to start writing early is the single most common reason strong proposals go unfunded — research is the writing.
The problem is compounded when using AI assistance without funder research built in. Vanilla AI produces proposal text that is structurally correct and substantively generic — it writes the needs statement, the outcomes section, the capacity narrative — but it does so without any knowledge of this specific funder's current priorities, recent awards, or theory of change. The output is faster to produce and just as likely to land in the not-funded pile, because the tells are still there. The writing is cleaner; the alignment is still absent.
What Changes When Research Precedes Every Section
The NovaKit Grant & Proposal Writing skill's research stage runs before the first section is drafted and stays active through each subsequent section. This isn't research as a front-loaded phase that gets folded into a brief and then forgotten — it's research as a continuous reference that each section is checked against as it's written.
| Proposal section | Without funder research | With funder research active |
|---|---|---|
| Needs statement | Sector-level statistics, general problem framing | Opens with the specific gap the funder has publicly identified, in their language |
| Outcomes | Generic logic model, measurable but unanchored | Maps to funder's stated theory of change and published success metrics |
| Organisational capacity | Comprehensive list of credentials and history | Leads with the specific credential this funder has weighted in recent awards |
| Budget narrative | Formatted correctly, framed generically | Framed around funder's known philosophy on overhead, capacity, and cost categories |
| Evaluation plan | Standard metrics, self-reporting structure | Reflects the evidence standards and data types this funder has required in past cycles |
The research doesn't add length to the proposal — it changes what goes into each section so that the content is aligned rather than generic. A well-researched proposal is often shorter than a generic one, because generic proposals compensate for weak alignment by adding more words. A funder-aligned proposal doesn't need to.
Who This Matters Most For
Development officers at nonprofits writing multiple applications per quarter who know the tells but don't have time to eliminate them manually on every submission. Researchers applying to programme grants where funder alignment is evaluated as heavily as research quality. Consultants whose proposal win rate determines their revenue.
The people who benefit most from removing these tells are those already writing enough proposals to feel the pattern: the applications that land, in retrospect, are the ones where the research happened first. The applications that don't land have the same structural features as the ones described above — not always, but consistently enough that it's a pattern worth interrupting. The skill makes the research stage the default, not the exception, which is what it needs to be for proposals to perform consistently rather than occasionally.
The program officer who sorts applications into two piles in the first paragraph isn't looking for proof that you understand the problem. She knows you understand it — that's why you applied. She's looking for proof that you understand her. That understanding only comes from research. And research that happens before writing, on every application, is what separates a strong success rate from a strong portfolio of rejections.
The next piece most people tackle from here is a visa cover letter that addresses every required point. If you're working across the full Legal & Biz workflow, the Legal & Biz bundle covers everything in one place.
If you're doing this manually, the grant and proposal writing skill is the faster path.
Put this to work: the Grant & Proposal Writing skill for Claude turns everything above into one guided workflow you run in a normal Claude chat. Not ready to buy? Start with a free Claude skill and see how it works first.
Related reading: The Grant You Didn't Get Wasn't Rejected for Lack of Effort