Existing footprint + field-agent hiring (the driver)
State rollout — NEW addressable jails (beyond the 58), by state in priority order
Facility economics — per jail (V1-grounded)
Multi-product upside — same rails, higher-value products
Costs & cash — the P&L side
3-year trajectory — paying lives & jails
Month-by-month
| Month | Agents | New jails /mo | Jails live | Cap % | New paying | Total paying | HC rev /mo | OEP bonus (1-time) | Equifax /mo | Cost /mo | Net /mo (operating) | Cash | IC apps | IC svc | IC close | IC leads | IC heads |
|---|
InsureCred staffing · scales with total APPS serviced — enrollment is one-time, servicing is forever
Field agents acquire jails; the 1099 piece-rate pays to sign a member. But the agency's real workload is applications, not paying lives — your V1 funnel ran ~6 apps per paying member (33k apps → ~5.5k paying), and the team works the whole funnel (pre-check → pending → qualified) plus the active book. Headcount below is driven by total apps. The AI agent goes live mid-2027 — that's the inflection that lifts apps-per-agent and bends the curve.
| Phase | Book | Apps /mo | Servicing | Closers | Leads | Mgmt | Total | Agency $/mo | $/member |
|---|
How it works · the strategy · grounded vs modeled
Rollout sequence. Field agents work each state's jails in priority order — GA → AL → FL, then TX, then TN → SC → MS, then KS → WY — moving to the next once a state's targets are onboarded.
Working-day calendar. Both new-jail onboarding and the existing-58 rail migration are paced by actual Mon–Fri working days that month, net of the 11 US federal holidays — you can't stand up a jail on a weekend or holiday. Holiday-heavy months (Jan/Feb ≈ 19 days vs 22–23 in lighter months) produce fewer go-lives, and quarters carrying those months are harder to clear on the cap goal.
Commission timing (cash vs lives). The lives columns show enrollment; the revenue columns show cash, which lags. ACA carriers pay monthly in arrears, so new-enrollment commission lands ~1–2 months after effectuation (the commission-lag lever). AOR captures pay faster (~1 month — the policy is already effectuated and paying). The existing book pays now (no lag). So early-month HC revenue trails the lives curve — and since Equifax's cap is a % of recognized net, they collect nothing against cash you haven't received yet, which deepens the early deferral. This is the honest cash picture.
Costs & P&L (cost layer). Monthly cost = fixed payroll/G&A (leadership + staff) + field agents × loaded cost + tech/software fixed (cloud, AI & tooling) + a per-life variable that scales with the book + InsureCred enrollment-agent piece-rate (~$10/paying life one-time — $1 pending + $1.50 PQ + $5 qualified; InsureCred is wholly owned by HealthCred, so its commission share stays in-house) + a leadership growth override per net-new life. CAC ~$10 vs LTV $300+ at 85% retention. Net /mo = HC revenue (recognized, lagged) − Equifax − all costs. Cash = starting capital + cumulative net — watch the trough. Net-margin-per-life is the unit-economics number a buyer underwrites.
Three layers. (1) Existing ~58 jails migrate onto Equifax rails up front — one-time activation of $2,500 each (matched to a market JMS benchmark) + per-event on the flow. (2) Field agents onboard new jails down the ladder, each railed on go-live (activation + per-event from day one). New-jail activation is tiered by size: $5,000 small/medium (ADP <1,000), $10,000 large/mega (ADP ≥1,000); existing jails $2,500. (3) The existing book sits underneath, excluded from Equifax's fee. Performance cap (quarterly-gated): the ceiling climbs in six 2-point steps — 5 → 7 → 9 → 11 → 13 → 15% of new-flow net — reaching the 15% max around Dec 2027. Each quarter Equifax must hit a jail-onboarding goal to unlock the next step for the following quarter — miss it and the step holds until they catch up. Once earned, a step is locked in. So Equifax has to perform across the entire ramp to earn the full rate, not just the first two quarters. The "Cap %" column shows the tier in force each month.
Exclusivity (term). The exclusivity runs Equifax → HealthCred: Equifax provides its correctional release-data rails for healthcare enrollment exclusively to HealthCred and won't enable a competitor — that's the moat. HealthCred is not sole-sourced to Equifax (it retains the right to multi-home with other JMS partners). In exchange, Equifax holds preferred-primary status and its fee economics — both tied to hitting the quarterly jail goals. Miss them and HealthCred leans on other JMS partners; underperformance costs Equifax the privileged position, not just cap upside.
Multi-product upside. Once a jail is wired, higher-value products ride the same rails at near-zero added onboarding — Medicare/dual (~$400 PMPM, ≈10–20× ACA), Medicaid redetermination, benefits. The model's "blended PMPM" reflects a Medicare attach; even a modest attach roughly doubles revenue per jail and is the lever that turns the ARR into a data-platform-multiple, billion-dollar story. Equifax's per-event fee is unchanged (still per data event), but the richer net raises their cap headroom — keeping them aligned to push every product through the rails.
Hiring is the throttle. Agents grow 2 → 6 (Jan ’27) → 12 (end ’27) → ~24 (’29). Realized new go-lives run ~1–2 jails/agent/week (each new jail is a sheriff + JMS + data-auth cycle), so onboarding the full ladder paces out over time rather than finishing in a few months. Your agent count and the onboarding cycle time set the pace.
Jail counts. Each state's slider is capped at its number of counties — the county-jail ceiling (municipal/regional jails would be additional upside). FL/GA/AL are shown net of the existing 58 already live there, to avoid double-counting; that per-state split is an estimate — set it to your real live counts when you have them.
Avg ADP is the swing variable (default 350 blended; big named jails dwarf the county tail). It moves the totals more than anything — ground it against the real county mix per state.
Capture = worked share × success rate. The 3% in the V1 pull was effectuation across all raw release events (broken-pathway, manual reach). Forward capture is built honestly as (eligible/worked share of releases) × (success rate on those worked): rails lift the worked share (real-time release data vs old ~6% manual reach); live GA Access lifts the success rate to 40–60%. The "effective capture" readout multiplies them — so the headline 40–60% is defensible as a success rate, not a claim that half of every release enrolls.
grounded in production data go-live burst ≈ 10–12% of ADP; real per-jail ADPs (large jails ~747 policies each); strong footprint in FL & AL with Georgia barely tapped = upside; ~85% renewal; ~$30 gross PMPM. forward / tunable worked share + 40–60% Georgia-Access success rate, agent ramp, state jail counts, new-jail ADP.