AI intake and staffing layer that helps regional home health agencies accept high-friction referrals they currently turn away.

1. AI has moved from note-taking into the full intake-to-QA stack, making it credible to automate the exact workflow that blocks referral acceptance today.
2. This is no longer a theory because one operator has already delivered more than 100,000 visits across 500-plus referring organizations using the AI-native model.
3. Payer-agnostic traction matters because agencies can now pursue commercial and Medicaid-adjacent volume that legacy processes often made too painful to serve.
4. The clearest new wedge is accepting patients competitors reject, since source reporting ties automation directly to serving referrals that coordination costs used to kill.
5. Legacy agencies face pressure now because the advantage appears structural, not a point-solution feature they can bolt onto their existing stack later.

Catalyst. Adaptive Innovations' scale and funding show that AI-native back-office compression is already good enough to turn coordination-heavy home health referrals into real visits, creating urgency for legacy agencies that are losing volume to faster operators.

The product sits between referral intake and start of care. It pulls structured and unstructured data from faxed packets, portal exports, or EHR-generated discharge summaries; checks payer eligibility and branch rules; recommends the right clinician and visit window; and assembles a draft start-of-care work queue with missing items highlighted before staff touch the case. For agencies, the first value is simple: accept more profitable referrals without adding intake headcount or burning out clinicians with incomplete charting. Over time, the system learns which cases are likely to fail QA, get delayed, or be rejected by a branch, letting operators intervene earlier and protect hospital relationships. The same workflow can later standardize acquisitions by giving newly purchased agencies a common intake and QA engine on day one.

What's different. Most home health software vendors sell record-keeping, billing, or documentation tools after an agency has already decided to take a patient. This company would own the decision window before acceptance, where the economic value is highest because every recovered referral immediately adds revenue and deepens hospital trust. That creates a better moat than another generic healthcare AI assistant: the system accumulates branch-level acceptance logic, payer nuances, staffing constraints, and QA outcomes that improve routing and pricing over time. If it later powers acquisition integrations, the company becomes part operating system and part network spine for fragmented regional agencies.

Jobs to be done

flowchart LR
  Buyer[Agency COO] --> Pain[Rejected or delayed referrals]
  Pain --> Product[Referral conversion engine]
  Product --> Outcome[More accepted starts of care]

Executive takeaways

• The pain is conversion, not demand: home health referral rejection rates hit 76% while patient acuity rose versus 2019, so agencies lose census because intake, staffing fit, and documentation readiness break before care starts [13][15][16].
• CMS is effectively validating the category by requiring acceptance-to-service policies tied to staffing, caseload, skills, and patient needs; intake discipline is becoming a compliance issue, not just an ops preference [4][5][9][10].
• Incumbent home health suites own documentation and billing after an agency decides to take a patient, but the sharpest economic value sits earlier in the workflow when a messy referral is either converted or lost [30][31][34][36][38].
• Texas is a credible beachhead because Adaptive already scaled across all major metros, 500-plus referring organizations, and 100,000-plus visits, proving referral density and urgency in-market [1][2].

Market definition

This market is the pre-acceptance operating layer for home health: ingesting referral packets, checking coverage and service-area fit, matching staffing capacity, surfacing missing documents, and teeing up a compliant start of care before the agency commits. It overlaps post-acute referral management and home health EHRs, but the most defensible wedge is the yes/no decision window where agencies currently rely on manual intake coordinators, branch rules, and phone calls [6][31][32][34].

Customer and buyer

The economic buyer is usually the COO, VP of Operations, or regional ops leader at an independent or PE-backed home health platform that owns intake throughput, branch staffing, and start-of-care performance. Daily users include intake coordinators, branch directors, and QA/compliance teams; hospital discharge teams matter as referral influencers because they value responsiveness on hard-to-place patients [13][14][18].

Buying triggers

• A visible jump in declined or delayed referrals that threatens census or hospital trust pushes agencies to revisit intake workflow and staffing logic. [13][14][15]
• New acceptance-to-service policy requirements make undocumented tribal rules harder to defend during compliance and survey reviews. [4][5][9][10]
• Higher-acuity discharges and more care shifting home make slow, manual acceptance decisions more costly. [16][18][25][27]

Willingness to pay

Buyers have a direct revenue-protection ROI case: the Medicare-standardized 30-day payment rate is about $2,057, while rejection rates remain historically high. A tool that converts even a small number of previously declined referrals or trims intake labor can pay for itself quickly in preserved starts of care and lower rework [4][13][15]. [4][13][15]

Category dynamics

Growth signal 7.4% CAGR

Validation signals

• Adaptive reports more than 100,000 visits, 500-plus referring organizations, and AI across intake through QA, validating that admin compression can support real operating scale.
• WellSky cites a customer processing 38% more referrals weekly with Enterprise Referral Manager, indicating measurable demand for intake automation.
• Referral rejection rates at 76% are severe enough that buyers do not need to be convinced the pain exists.
• More than half of home-based care providers have invested in AI or plan to, with staffing pressure cited as the main driver.

Regulatory & technical constraints

• Acceptance-to-service decisions must align with documented agency capacity, staffing levels, competencies, and patient needs under CMS rules.
• Any workflow affecting timeliness, start-of-care, or transfer-of-information processes intersects with HH QRP and OASIS reporting expectations.
• Value-based purchasing and quality benchmarking increase the need to connect intake choices to downstream outcomes rather than treating them as isolated admin tasks.
• Hospitals and agencies increasingly need secure, interoperable flows between referral networks, EHRs, and home-based monitoring tools.

The field splits between broad home health operating suites (HCHB, MatrixCare, AlayaCare, Axxess) and referral-network/intake platforms (WellSky/CarePort). The suites are strongest once a patient is already accepted; WellSky is closest to the proposed wedge because it centralizes inbound referrals and adds AI extraction. The opening for a startup is a vendor-neutral system focused on branch-specific acceptance logic, pre-acceptance staffing fit, and clinician-ready work queues before documentation and billing systems take over [30][31][32][34][36][38][39].

Why incumbents do not win by default

• Home health EHR suites. These vendors win the system of record after admission, but their core UX and ROI story still centers on documentation, billing, and compliance rather than converting borderline referrals before a branch says yes.
• Referral network platforms. WellSky/CarePort owns upstream traffic and referral routing, but its broad cross-continuum positioning leaves room for a branch-level decision engine tuned to home health acceptance economics.
• Hospital-at-home programs. These programs increase the volume and complexity of patients moving home, but they do not solve the agency-side intake and staffing logic needed to accept those referrals.
• Manual in-house ops teams. Manual coordinators remain the default substitute because agencies trust local knowledge, yet the evidence shows capacity constraints, rising acuity, and rework are overwhelming purely human workflows.

This company should start as a vendor-neutral referral conversion layer for regional home health agencies that lose census on high-friction referrals before care even starts. The best first customer is a Texas-based independent or PE-backed agency with 5-20 branches, 100-500 clinicians, multiple hospital referral partners, and a visible backlog of evening, weekend, or mixed-payer referrals that staff cannot clear fast enough. The product wedge is intentionally narrow: ingest discharge packets, verify coverage and service-area fit, encode branch acceptance rules, recommend staffing, and generate a clinician-ready start-of-care work queue before the agency accepts the patient. That scope matches the buying trigger because operators act when declined referrals, branch expansion, or a new hospital relationship expose lost revenue and response-time risk. The strongest strategic advantage is not generic OCR; it is accumulating branch-level accept-decline logic, payer nuances, missing-document patterns, and downstream QA outcomes that incumbents and manual teams do not centralize today. Product sequencing should stay disciplined by proving an overlay deployment on fax, email, and referral-portal exports before deeper EHR integrations, broader staffing optimization, or payer-performance products. Public market sizing and pricing data are still modeled rather than transaction-backed, and the biggest open question is whether agencies will buy a standalone overlay fast enough to support repeatable ACVs. The first 12 months therefore need customer-owned proof on acceptance-rate lift, referral-to-decision time, pilot-to-production conversion, and implementation burden.

Problem

• Agencies lose profitable starts of care because complex referrals require manual coverage checks, branch-specific rule lookups, staffing calls, and chart-prep rework before anyone can say yes.
• Hospitals and discharge teams remember slow or inconsistent response times, so agencies risk both immediate census loss and long-term referral share when after-hours or mixed-payer cases stall.

Solution

• Deploy an intake overlay that parses faxed or exported referral packets, verifies payer and geography fit, flags missing documentation, and recommends whether a branch can accept the case.
• Turn accepted cases into clinician-ready work queues with staffing recommendations, audit logs, and QA-risk flags so agencies can start care faster without adding proportional intake headcount.

Why we win

• The company owns the highest-value decision window before admission, where one recovered referral creates revenue immediately and where incumbent suites are less specialized.
• Each deployment compounds proprietary data on branch rules, payer exceptions, decline reasons, response times, and downstream QA outcomes, creating a workflow moat that point AI features are unlikely to match quickly.

Milestones

flowchart LR
  Wedge[Referral conversion wedge] --> MVP[Overlay intake MVP]
  MVP --> Proof[Acceptance and response-time proof]
  Proof --> Expansion[Multi-branch benchmarks and portfolio expansion]

Founding team

Experiment roadmap

Risk assessment

What must be true

• The first 5 target agencies must show a material share of declined referrals are caused by intake and documentation friction rather than clinician scarcity alone.
• A thin overlay must go live on existing fax, email, or portal-export workflows without requiring a multi-quarter EHR integration project.
• Paid pilots must raise hard-referral acceptance or reduce referral-to-decision time enough for operators to fund an annual contract from operating budget.
• Buyers must prefer a vendor-neutral pre-acceptance layer over waiting for incumbent EHR suites to add similar features.
• Branch-rule, payer, and QA outcome data must compound into better decisioning and expansion, or the product will collapse into a lower-value automation tool.

Open diligence questions

• What percentage of current declines at target agencies are driven by documentation, payer, and coordination issues versus pure clinician capacity?
• Which branch-specific acceptance rules are still tribal knowledge rather than codified in incumbent systems?
• How many weeks of implementation work will buyers tolerate before demanding native EHR integration?
• What proof threshold converts a pilot from innovation budget to recurring operating budget?
• How exposed is the wedge if WellSky or major EHR vendors add basic AI intake automation in the next 12 months?

Model sanity

• Revenue engine. Base-case revenue is driven by reaching 12 live agencies by Q4Y2 and 40 by Q4Y3 while expanding older accounts from pilot pricing into roughly $72K-plus blended annual revenue.
• Must go right. The company must keep deployment overlay-first and convert pilots to production within about 4-5 months or the lean headcount plan stops covering burn.
• Model breaks if. If sales cycles stretch to 6 months or gross margin falls below 67%, cash drops toward the downside case even without missing the full customer target.
• Next-round proof. The next financing is justified once the company can show 10-12 production agencies, repeatable sub-8-week onboarding, and one portfolio-style expansion path from the first PE-platform channel.

Scenarios

Sensitivity

flowchart LR
  Leads[Qualified agency targets] --> Pilots[Paid pilots]
  Pilots --> Customers[Production agencies]
  Customers --> Revenue[Branch subscription + usage revenue]
  Revenue --> GrossProfit[70% gross profit]
  GrossProfit --> Cash[Cash runway]

Flags: Revenue per exit FTE remains below typical mature SaaS benchmarks because the base case still carries meaningful onboarding and support work. · The model assumes 40 live agencies by Q4Y3, so underperformance in pilot-to-production conversion would likely require a tighter hiring plan or earlier fundraising. · Gross margin is capped at 70%; if buyers demand deeper write-back integrations sooner, services intensity could delay the path to break-even.

• Workflow integration drag. Agencies may resist adopting a new intake layer if it requires too much change management across branches and existing software. Mitigation: Start with a thin intake-overlay product that works from fax, email, and exported referral packets before deeper system integrations.
• Outcome ownership ambiguity. Hospitals and agencies may blame each other for missed starts of care, making ROI attribution harder during early sales cycles. Mitigation: Instrument acceptance time, rejection reason, and start-of-care completion metrics from day one so pilots prove the revenue and service impact clearly.
• Incumbent feature catch-up. Existing home health software vendors could add lightweight AI intake features once the category becomes visible. Mitigation: Focus on the pre-acceptance workflow and branch-level operating intelligence that incumbents lack, then deepen the moat through multi-branch benchmarks and acquisition integration playbooks.