Hardest Part

Parsing dense, unstructured government and foundation RFPs to deterministically evaluate strict eligibility constraints against a user's specific technical capabilities and entity structure.

Min Viable Scope

Restrict v1 exclusively to matching deep-tech startups with active US federal SBIR and STTR solicitations based on uploaded pitch decks. Leave out academic grants, foundation RFPs, and all automated proposal drafting features.

Cold Start Problem

Users need a comprehensive, up-to-date database of active solicitations on day one to trust the matching engine. Break this by narrowly scraping one agency portal and seeding the matching engine with public data from previously funded companies in that exact vertical.

Time To First Value

Under 1 hour to index a user's technical documents and return a ranked list of high-probability active solicitations

Data Moat Available

true

Technical Difficulty

Moderate

Wedge

Start with climate and environmental non-profits applying for federal EPA and DOE grants. This niche faces a massive influx of new federal funding but deals with uniquely complex compliance and eligibility requirements. Once established, expand horizontally to health and scientific research grants, and finally to general philanthropic foundation grants.

Timing

Large language models now possess the context windows and reasoning capabilities necessary to ingest 50-page organizational histories and map them against complex, jargon-heavy federal grant RFPs with high accuracy.

Why This ICP

Mid-sized non-profits with $1M to $10M operating budgets rely entirely on grants for survival but lack the budget for full-time, dedicated grant prospecting teams. This makes them highly motivated buyers for automated labor leverage.

Size Of Prize

There are roughly 1.5 million registered non-profits and research institutions in the US, with an estimated 150,000 actively hunting for recurring grants. At an average annual spend of $5,000 per entity on outsourced grant writer time for discovery and legacy database subscriptions, the addressable market is $750 million.

Gap Narrative

Non-profits and research labs spend hundreds of hours manually parsing fragmented federal, state, and foundation grant databases to find eligible funding. They lack a system that automatically reads their organization's historical work and instantly cross-references it against live grant requirements to qualify and prioritize matches.

Defensibility

Defensibility compounds through a proprietary dataset of successful versus rejected grant narratives and nuanced eligibility edge-cases. As the system ingests more feedback on which matched grants actually win funding across the network, the matching algorithm becomes a specialized model that generic search tools cannot replicate.

Why This Thesis

A Service-as-Software approach fits perfectly because the required output is a qualified list of matches and drafted narratives. This replaces the raw labor of a junior grant writer rather than just providing another search interface for the executive director to manage.

SAM

~$250M-1B US mid-market nonprofits actively seeking foundation grants

SOM

~$10M-25M

TAM

~300k active US nonprofits × ~$5k-10k/yr software spend ≈ ~$1.5B-3B

Growth Rate

~12-18%/yr, driven by tightening philanthropic giving forcing nonprofits to increase grant application volume without adding headcount

Paid Comparable Spend

~$1.5k-3k/yr for legacy grant databases like Foundation Directory, plus ~$50k-90k/yr for dedicated grant writing staff

Kill Thresholds

• Less than 20% of active users submit an AI-drafted grant within their first 30 days
• Average human editing time exceeds 4 hours per proposal during the trial phase
• CAC exceeds $1,000 against a $3,000 ACV target in the first 90 days
• Month 3 gross revenue retention falls below 80%

Leading Metrics

• Time from account creation to first completed proposal draft
• Percentage of AI-generated text retained in the final exported submission
• Number of qualified grant matches moved to the active application pipeline weekly
• Historic grant data ingestion and parse success rate

What Proves Right

Nonprofits connect their organizational history and successfully submit at least two AI-drafted grant proposals within their first 14 days. The platform secures $3,000 annual contracts by directly displacing legacy database subscriptions like Foundation Directory. Cohort retention exceeds 70% at month six as organizations integrate the prospector into their weekly fundraising cadence.

What Proves Wrong

Development directors reject the generated drafts for lacking organizational voice and revert to writing proposals from scratch. Nonprofits refuse to upload past successful grants due to board-level data privacy mandates, breaking the personalization model. The human editing time required per proposal equals or exceeds manual drafting time, resulting in immediate churn after the initial trial.