What changed
VA posted a Grants.gov forecast for the Homeless Providers Grant and Per Diem (GPD) program covering FY2026-FY2028 (CFDA 64.024, opp VA-GPD-FORECAST-FY2026-FY2028) — FACT from the cited listing. This signals another multi-year cycle in which community nonprofits will be reimbursed per veteran per bed-night and must continuously document eligibility, occupancy, and services (inference: the per-diem documentation mechanics are the program's standing structure, not stated in the forecast text itself).
Why now
A new grant cycle means both renewing grantees and new awardees must stand up or re-justify their documentation workflow now, and every month of operation generates a reimbursement submission. Unlike a one-shot application, GPD creates RECURRING paperwork — the ideal per-transaction monetization shape for this founder. The forecast status (no close date) gives a 3-6 month runway to build before the NOFO lands, which the founder can now afford.
Converging signals
Three signals meet at one point: (1) a funded federal program with a forecasted new cycle (FACT, Grants.gov 356629); (2) a defined compelled class — GPD grantee nonprofits who cannot get paid without submitting bed-night and eligibility documentation (inference from program structure); (3) a submission path (VA GPD program office vouchers plus HMIS data feeds — inference). Per the system's own scoring rule, mandate + filer class + submission = convergence even though it is unglamorous.
Customer pain
Small transitional-housing nonprofits track bed-nights, veteran eligibility, discharge dates, and service notes across paper logs, spreadsheets, and an HMIS that was built for HUD reporting, then hand-assemble monthly per-diem vouchers for the VA and pray they survive annual inspections and audits. A rejected or under-documented bed-night is lost revenue; a failed audit threatens the whole grant. HYPOTHESIS: no direct complaint evidence was provided — this pain is inferred from the program's reimbursement structure, not demonstrated by PAIN-type evidence.
Who pays
The GPD grantee nonprofit (executive director / program director / billing coordinator). They are a FORCED BUYER in the scoring sense: they cannot collect their per-diem revenue without producing this documentation, and the program's own money (per-diem revenue plus admin allowance) funds the tool. ~600 provider sites and a ~$250M/yr program are stated in the input as inference — treat the exact figures as unverified but the order of magnitude as plausible.
Solved today
HYPOTHESIS: a mix of spreadsheets, paper sign-in sheets, generic HMIS platforms (WellSky/ServicePoint, Bitfocus Clarity, Foothold, CaseWorthy) that handle HUD data standards but not VA per-diem voucher assembly, and manual monthly voucher preparation by program staff. Larger multi-site providers may have internal tools; small single-site providers almost certainly do not.
Why current solutions are bad
HMIS is oriented to HUD CoC reporting, not to generating a VA reimbursement packet; the gap between 'data is in HMIS' and 'voucher the VA will pay without questions' is closed today by staff hours. Occupancy math (leave days, discharge timing, eligibility lapses) is exactly the kind of rule-bound reconciliation software does better than a part-time billing coordinator. Audit trail assembly is retrospective and panicked rather than continuous.
Proposed product
GPD BedBill: a per-site web app where staff log check-ins/check-outs/leave nights in seconds (or import HMIS CSV exports), which continuously computes billable bed-nights under GPD rules, flags eligibility/documentation gaps BEFORE month-end, and generates the monthly per-diem reimbursement packet plus a perpetual inspection-ready audit binder (occupancy ledger, eligibility docs index, service notes). Charge per site per month, or per billed bed-night to align price with their revenue.
MVP version
Single-site web app: roster + nightly census entry, GPD bed-night calculation rules, monthly voucher/packet PDF generator matching what the VA liaison expects, and an exportable audit ledger. No HMIS API integration in v1 — accept CSV import. Build with founder's standard fast-prototyping stack; the domain rules come from the GPD Handbook/regulations (38 CFR Part 61) and 2-3 design-partner providers.
30-day build
Pull the current GPD grantee list (VA publishes awardees), verify the actual voucher/submission format by interviewing 5-10 program directors (offer free setup), and confirm the real artifact the VA per-diem payment process requires — this kills or confirms the inference-heavy parts of this brief. Build the census + bed-night engine in parallel.
60-day build
Working MVP live with 2-3 design partners running a real month-end reimbursement cycle. Iterate the packet format against what their VA network homeless coordinator actually accepts. Draft the audit-binder feature from a real inspection checklist.
90-day revenue plan
Convert design partners to paid (~$149-299/site/month), then direct outreach to the grantee list — every one is publicly identifiable, and the pitch is 'recover one denied bed-night per month and it pays for itself.' 10 paying sites = ~$2-3k MRR within the 180-day window; NOFO publication (new cycle applicants) is a second sales moment.
Distribution path
Direct and cheap: the buyer list is a public federal grantee roster of ~600 sites; sell via email/phone demo showing a real voucher generated in minutes. Secondary channels: NCHV (National Coalition for Homeless Veterans) conference/newsletter, and the consultants who help nonprofits win GPD grants — give them a referral cut. No ad spend, no marketplace, no procurement office: the nonprofit signs up like any SaaS.
Pricing hypothesis
$149-299 per site per month flat, or $0.25-0.50 per billed bed-night (a ~70-bed-night/day site at ~$70+/night per-diem grosses ~$150k/month from VA — a $250/month tool is noise against one recovered denial). Annual prepay discount aligned to their grant year.
Technical difficulty
Low-moderate: CRUD app + a rules engine encoding 38 CFR Part 61 / GPD Handbook bed-night billability + PDF generation. The hard part is domain accuracy, not code — which is exactly the founder's ELDT-proven skill (read the mandate, encode the filing rules). HMIS API integration is a later, harder step; CSV import defers it.
Legal / regulatory risk
Low-moderate. The tool touches veteran PII including potentially health-adjacent info (eligibility, discharge status) — needs solid security posture and a BAA-style data agreement with providers, but the founder is not himself a covered entity or licensee (HYPOTHESIS: HIPAA likely does not directly attach to a housing-occupancy billing tool, verify). No license required to operate; compliance content is the moat, not a burden on him.
Platform dependency
None in the deplatforming sense — the 'platform' is a federal reimbursement process that cannot revoke API access he doesn't depend on. Real dependency risk: if the VA modernizes GPD billing into a mandatory portal/system of its own, the packet-generation layer must adapt — but that historically creates MORE demand for submission tooling (his ELDT business is exactly this), not less.
Founder fit
Near-maximal on the stated thesis: public money flows per-transaction to a defined recipient class that must document to get paid; the software layer on that paperwork is precisely the wedge, and it rhymes exactly with his shipped FMCSA ELDT per-upload product. Bonus: fire-service/public-sector credibility plays well with veteran-services nonprofits. Nonprofit buyers are modest spenders but reachable without relationship sales — demonstrated value (a generated voucher) closes.
Breakout potential
Strong horizontal expansion: same bed-night/per-diem documentation engine extends to SSVF and HUD-VASH adjacencies, state homeless per-diem programs, emergency shelter ESG reporting, and generally any 'reimbursed per unit of service, must document each unit' program — a large family of forced-filer niches with 50-state replication on the state side.
Final recommendation
PURSUE-VALIDATE: this is squarely the founder's proven shape (recurring per-transaction government paperwork, defined compelled class, public buyer list, per-filing monetization) and he now has runway for the 30-day validation sprint it needs. Gate the build on one week of grantee interviews confirming (a) the voucher format is real, painful, and manual, and (b) no HMIS module already owns it. If confirmed, build; the recurring monthly reimbursement cycle makes this better than one-shot grant-application tools.
Next action
Download the public list of current GPD grantees from va.gov, read the GPD payment provisions in 38 CFR Part 61 and the GPD Grantee Handbook to document the exact monthly reimbursement artifact, and book 5 program-director calls this week asking one question: 'Walk me through how you got paid for last month's bed-nights.'