Evaluated by
First-pass match rate against payer 271 ground truth, weekly.
The ASP-RCM AI Suite
The ASP-RCM AI Suite. Every vendor says “AI.” We’ll show you the model card.
AI in revenue cycle is now table stakes. Trustworthy AI is the moat. Six capability clusters, eight production models, one credentialed human gate. Every model has a published evaluation cadence behind it. No black boxes. No “agentic” theater.
Position
AI runs volume.
People run judgment.
Evaluators run trust.
People run judgment.
Evaluators run trust.
A different bar than the rest of the market
Most AI-RCM vendors stop at the first sentence. We finish the paragraph.
WATCH IT WORK · 3 MIN
See ASP Genie in action.
Three minutes of the actual product working on real RCM tasks. No animation, no roadmap slides, no fake screens. Live walkthrough.
Want a walkthrough on your own data? Bring a sample claim file or denial report and a senior partner will run Genie against it live.
Book a live Genie demo →
AI capability hub
32 pages of AI for revenue cycle, without the hype.
Below: every AI capability we operate, every vendor-evaluation framework we use, every specialty-specific AI playbook, and the trust questions we recommend asking any AI vendor. No vendor names in the evaluation pages. Buyer-perspective only.
AI Capabilities 11 pages
Vendor Evaluation Frameworks 8 pages · no vendor names
AI by Specialty 5 pages
Glossary, ROI, and Trust Framework 8 pages
Plus the live tools we maintain.
The ABA Payer Matrix covers 244 cells across 52 jurisdictions and 8 payers, every cell linked to source, refreshed monthly. The 50-state ABA RCM cluster covers state-by-state Medicaid programs, dominant Blues, MCO contracting, and BCBA credentialing for every U.S. market.
The position
The market sells AI like it’s the answer. ASP-RCM sells the governance around it.
In 2024, “does it have AI” was a buying question. In 2026, it isn’t. Every RCM vendor on your shortlist now claims AI on every screen of their pitch deck. The interesting question moved.
The question that matters now: can you trust the output enough to bill against it? That answer doesn’t live in a model name or a parameter count. It lives in the human gate, the eval cadence, and the audit trail.
That’s the page you’re reading. This isn't a feature list. It's a governance brief for the AI we run on your revenue.
Diligence framework
Three questions every CFO should ask any AI-RCM vendor.
If a vendor can’t answer all three with a straight face, they’re selling you a demo, not a system. Here’s how the market typically scores. And here’s how we score.
01
Is there a credentialed human in the loop, on every chart that matters?
Pure-AI vendor
No. The pitch is “autonomous.” The audit is sampling.
Bolt-on AI
Sometimes. Coder reviews suggestions, but at low volume.
ASP-RCM
CPC-credentialed coder reviews every complex chart. Random + risk-based QA on the rest. Independent quarterly audit.
02
Does it retrain on your 835s, or a generic corpus?
Pure-AI vendor
Generic foundation model. Same denial logic for a Florida ER and a Texas ABA group.
Bolt-on AI
Industry-tuned, but rarely tenant-tuned. Updates on vendor cadence, not yours.
ASP-RCM
Payer-specific learners trained on your 835 outcomes. Denial-prediction model retrains weekly on your remits.
03
Can you see the model’s reasoning on a denied claim?
Pure-AI vendor
Black box. “Trust the model.” Audit trail is a status field.
Bolt-on AI
Decision logged. Reasoning often isn’t. Hard to defend at a payer appeal.
ASP-RCM
Per-claim reasoning trace, scored confidence, and human-attributable action log. Compliance-grade on every account.
The Suite
Eight tools, organized by what the AI actually does.
The eight tools aren’t a feature list. They’re a portfolio of six capability clusters. Each cluster solves a distinct AI problem (sequence modeling, payer rule learning, time-series surveillance, etc.). Click any model for the full deep-dive.
01
Front-end intelligence
Real-time payer reasoning & coverage discovery
Two models running pre-visit and on every aged self-pay account. Real-time 270/271 against national payers, plus a multi-payer sweep that finds active coverage hidden in registration data. Stops eligibility-driven denials before the claim is ever built.
02
Autonomous coding
Multi-specialty coding with a CPC gate
Sequence-to-sequence model trained on documentation patterns across hospital E/M, surgical, OB, ortho, cardiology, behavioral health, ABA, and HCC risk-adjustment. Confidence-scored. Below threshold routes to a CPC-credentialed coder. Every complex chart, not random sampling.
03
Predictive yield
Pre-submission scoring & upstream attribution
Two complementary models. The first scores every claim against payer-specific denial patterns and holds the high-risk ones. The second maps denials that did happen back to their upstream owner (registration, coding, eligibility, auth) so the same denial doesn’t come back next month.
04
Continuous surveillance
Always-on claim status monitoring
Time-series surveillance over every submitted claim. 276/277 polling on payer-specific cadence, IVR/portal scraping fallback for payers that don’t support 276, payer-behavior baselining so anomalies surface fast. Specialists work cases, not hold lines.
05
Workflow orchestration
Dollar-weighted routing & auditable action
Optimizer that prioritizes the AR queue by recovery probability × dollar value, not by age. Routes by specialist skill, payer, and complexity. Logs every action against every claim, compliance-grade. The work-list that finishes the work.
06
Financial close AI
Daily three-way reconciliation
Match-and-resolve engine that reconciles bank deposits, lockbox files, ERAs, and posted transactions every business day. Auto-posts 835s with adjustments. Flags variances inside 24 hours, not at month-end. Compresses the close.
Model cards
Eight production models. One spec sheet each.
For every model in the Suite: what it does, where the human gate sits, how it’s evaluated, and the cadence of retraining. Hover any card and tap through for the full workflow + capabilities.
M-01 / FRONT-END
Eligibility Verification AI
live · 270/271
Real-time eligibility against national payers (commercial, Medicaid, Medicare, MA). Surfaces inactive policies, deductible/copay, prior-auth requirements, and COB order before the patient walks in.
Human gateFront-desk re-verifies on any anomaly: terminated policy, wrong primary, missing auth. AI never blocks a check-in unilaterally.
Full model card
M-02 / CODING
Autonomous Coding AI
live · multi-specialty
Full coding (CPT, ICD-10, HCPCS, modifiers, sequencing) across hospital E/M, surgical, OB, ortho, cardiology, behavioral health, ABA, and HCC risk-adjustment. Confidence-scored per chart.
Human gateCPC-credentialed coder reviews every chart below confidence threshold. Risk-based QA sample on the rest. Not random sampling at the rate level. Chart-level.
Full model card
M-03 / YIELD
Denial Prediction AI
live · payer-specific
Pre-submission scoring of every claim against payer-specific denial patterns. Surfaces missing modifiers, NDCs, units, auth gaps, and medical-necessity risk. Each one ships with a fix suggestion, not just a risk score.
Human gateAbove threshold, claim holds for biller review with an actionable fix list. AI never auto-releases or auto-rejects.
Full model card
M-04 / FRONT-END
Insurance Discovery AI
live · multi-payer sweep
Finds active coverage hidden in self-pay accounts (Medicaid retro-eligibility, COBRA, secondary, tertiary) via demographic + SSN multi-payer sweep. Daily re-sweep on aged AR.
Human gateAR specialist verifies coverage and rebuilds the claim on the new payer. Found coverage is never auto-billed without confirmation.
Full model card
M-05 / SURVEILLANCE
Claim Status Surveillance
live · 276/277 + IVR
Continuous surveillance across the claim lifecycle. 276/277 polling on payer cadence, IVR/portal scraping fallback, payer-behavior baselining, aging-bucket triggers. Routes only human-needed cases to specialists.
Human gateAR specialist works every claim that breaches its expected adjudication window. AI never auto-writes off, auto-appeals, or auto-resubmits.
Full model card
M-06 / WORKFLOW
AR Workflow Engine
live · $-weighted
Optimizer that prioritizes the queue by recovery probability × dollar value. Routes by skill, specialty, payer, complexity. Logs every action with timestamp and actor for a compliance-grade audit trail per claim.
Human gateSenior AR lead reviews escalations, sampled QA, and any account flagged compliance-sensitive. Engine recommends, lead approves writes.
Full model card
M-07 / YIELD
Denial Root-Cause Analytics
live · CARC/RARC + owner
Maps every denial past CARC/RARC code into its upstream owner: registration, coding, eligibility, auth, or payer policy. Detects repeat-denial patterns across provider × payer × code. Fix at source, not at appeal.
Human gateSenior partner walks the trends with your CFO every month. Provider retraining and payer escalations are human-led, AI-evidenced.
Full model card
M-08 / CLOSE
Cash Reconciliation AI
live · daily 3-way
Daily three-way reconciliation across bank deposits, ERA/EFT, and posted transactions. Auto-posts 835s with adjustments + contractual write-offs. Variances flagged inside 24 hours, not at month-end.
Human gateAR specialist resolves every flagged variance and any orphan deposit. Refunds + credit balances are compliance-tracked, never auto-issued.
Full model card
How we sit in the market
The four kinds of vendor on your shortlist. And which one we are.
Every AI-RCM vendor falls into one of four archetypes, and each archetype has a structural gap. We’re calling the gaps out by category, not by name. Diligence the rest yourself.
Archetype
What they pitch
Where it breaks
ASP-RCM
Pure-AI software vendors
SaaS · subscription · “autonomous”
“Replace your billing team with AI. Pay per claim. Watch the dashboard.”The deck is a model card and a logo wall.
No accountable senior partner. When the model is wrong on 1% of charts at hospital scale, that’s an audit risk no dashboard can absorb. You own the cleanup.
We sell outcomes. Not software. A senior partner is named on your account and their bonus is tied to your collections, not your seat count.
Clearinghouse + AI bolt-on
legacy pipes · features added on top
“We already move your claims. Now there’s an AI tab.”AI is a feature line on the renewal.
AI is bolted onto a switching pipe, not built into a service. The model touches the claim once. Nobody on their side gets paid to fix the upstream cause.
AI is built into the service. Not bolted onto pipes. Every layer feeds the next, and a senior partner closes the loop on root cause monthly.
Single-point AI tools
one capability · stitched into your stack
“Best-in-class AI coder.” “Best-in-class AI denial predictor.”Pick three. Integrate. Maintain.
You become the system integrator. Three vendors, three contracts, three audit trails, three different definitions of “done.” Nobody owns the cycle.
One Suite, one accountable senior partner. One audit trail per claim from eligibility to cash posted. The integration is our problem, not yours.
Offshore RCM, low-tech
labor arbitrage · scripts & macros
“We do it cheaper at scale.”The pitch is FTE rate cards.
Volume problem solved with more people. Quality problem unsolved. Eight hold-line FTEs is not a strategy; it’s a margin until the payer changes a rule.
We run AI on volume and a credentialed human gate on judgment. The cost-to-collect line bends because the work-mix bends, not because we hired more.
Trust & governance
The part of the AI conversation most vendors skip.
Independent audit cadence, role-defined human gates, model-retraining policy, and what happens when the AI is wrong. This is the section that closes with a security team.
Independent audit
Quarterly external CPC review
An external CPC firm pulls a meaningful sample of charts the AI coded and reviews them blind. Findings publish to client leadership and feed back into model retraining. No marking your own homework.
cadence: quarterlyHuman gate, by role
Defined credential per layer
Coding gate: CPC. Eligibility/discovery gate: Patient Access lead. AR escalations: Senior AR lead. Compliance-sensitive accounts: Senior partner + Compliance Officer. Every gate is named and credentialed.
CPC · CRCR · CCS-P · COCRetraining cadence
Models stay current with payers
Denial prediction retrains weekly on your 835 remits. Eligibility heuristics drift-checked monthly. Coding models updated against quarterly CPC audit findings. Reason-code taxonomy refreshed on payer policy bulletins.
weekly · monthly · quarterlyCompliance & data
HIPAA, SOC 2 Type 2, ISO 27001, HITECH
U.S. data residency for production claim data. PHI access logged at the user/action level. Independent SOC 2 Type 2 audit annually. ISO 27001 certified ISMS. Model training never on identifiable PHI.
U.S. residency · audited annuallyWhen the AI is wrong
Named escalation path
Every account has a written SLA on response time and a named senior partner accountable for outcome. AI errors caught at the gate go back into retraining; AI errors that reached a payer get worked, root-caused, and reported in the QBR.
SLA in 90 days · named partnerWhat we don’t do
The autonomy lines we won’t cross
No AI auto-write-offs. No AI auto-appeals. No AI auto-refunds. No coding submitted without a CPC gate on the charts that matter. No model output presented to a payer without a human-attributable action log behind it.
human-in-the-loop, by policyHow a chart actually moves through the Suite, with the gate visible.
Coding example · M-02STEP 01
Chart enters queue
EHR pushes documentation. Specialty + provider + payer attached as features.
actor: system
STEP 02
AI assigns codes
Full coding (CPT, ICD-10, HCPCS, modifiers, sequencing) with confidence per code.
actor: M-02
STEP 03 · GATE
CPC review on complex
Below confidence threshold or specialty-flagged → routes to credentialed coder. Human approves, edits, or rejects.
actor: CPC
STEP 04
QA sample
Random + risk-based sampling on auto-coded charts. Senior coder scores, errors flow back into retraining.
actor: Senior coder
STEP 05
Quarterly external audit
Independent CPC firm pulls a sample blind. Findings published to client leadership.
actor: External CPC
Want to see the Suite on your data?
30-minute walk: a senior partner runs your last 90 days of denials and remits through the Suite and shows you which models would have caught what. No demo theater, no NDA needed.
Book a 30-min audit
Pick a slot. No SDR triage.
Talk to a senior ASP-RCM partner about applying AI inside your RCM. 30 minutes. Bring a candidate use case. We will walk through pilot scope, exit criteria, and the math behind production deployment.
FOR CIO, CFO, RCM DIRECTOR, IT LEADERSHIP
30 minutes with a senior partner.
Free. Same calendar a senior partner and directors live in. The next available slots are below.
Or call 469-393-0083 or email [email protected].