A case study in AI-assisted patient billing — with methodology, source data, and intellectually honest caveats disclosed. Every number is observed, not modeled, unless explicitly flagged.
I'm the operator of the imaging group where this platform was built and measured. Same person, two roles: I run Crown Valley Imaging (15 years as CEO), and I built this platform through Veredge. That's not a bug in this case study — it's the whole point.
The platform exists because I had the problem first. I got tired of watching real money age out of my own AR while waiting for vendors to solve it, and I built the fix for my own business. Then I ran it on my own ledger for 123 days. Every number in this report is from that ledger.
You could reasonably ask: “is this really an independent case study?” Honestly — no, and I'm not going to pretend otherwise. What it is: a real operator, a real production business, real patient balances, a real measurement window, and a reproducible pipeline anyone can audit. The data passes whatever test you want to run it against. What it isn't: a third-party testimonial. Those come as we license the platform to other practices.
Veredge exists becauseof what this platform did for Crown Valley Imaging. I'm extending it to other practices because the method works — and because the only thing more dishonest than a puff-piece case study is pretending I discovered this solution anywhere other than in my own business.
The platform replaced a legacy patient-billing platform inside a diagnostic imaging group on December 15, 2025. In the 123 days of observed operation immediately following launch, the platform delivered measurable improvements across every financial and operational KPI we track — without increasing labor cost.
| Metric | Legacy platform (equal 123-day window) | The platform (post-launch) | Change |
|---|---|---|---|
| Total patient-responsibility dollars collected | $233,471 | $535,623 | +129.4% |
| Collections per paying account | $133.87 | $182.62 | +36.4% |
| Daily collections run-rate | $1,898 / day | $4,355 / day | +129.4% |
| Paying-account count | 1,744 | 2,933 | +68.2% |
| 60-day responsibility collection rate | 51.0% | 66.4% | +15.3 pts |
| 90-day responsibility collection rate | 53.1% | 75.8% | +22.6 pts |
| Invoice-to-payment conversion rate | 49.3% | 73.4% | +24.1 pts |
| Invoices fully paid off by end-of-window | 47.8% | 72.5% | +24.7 pts |
Incremental gross collections vs. legacy equal-window baseline.
Retained vs. industry-standard 35% collection-agency fee on the same base.
The platform operating cost: $7,426 over 123 days (comms + human follow-up, excl. platform licensing).
The $535,623 total is channel-inclusive — drawn from the platform's system-of-record ledger (online-initiated, cash-at-front-desk, POS-card, mailed-check, and payment-plan draws). Legacy baseline is constructed identically from that platform's own disbursement ledger. Under a stricter digital-rail-only attribution (only payments the platform processed on its own rail), the platform collected $275,155 — still +17.9% vs. legacy baseline. Full channel composition and three alternative attribution views are in the caveats section.
U.S. patient financial responsibility has grown dramatically over the last decade as high-deductible health plans have become standard — the average single deductible reached $1,790 in 2024, up from $584 in 2006.1 Patients now shoulder a much larger share of every medical bill.
At the same time, 41% of U.S. adults currently hold some form of medical debt, and an additional 16% have paid medical debt off in the last five years — meaning most adults (57%) have experienced owing money due to medical or dental bills.2
Industry research shows provider collection rates for small-dollar patient liabilities from insured patients run at 50–70%, falling to ~10% for self-pay — yet McKinsey's consumer research finds >90% of patients are willing and able to pay balances under $500 when given convenient payment mechanisms and structured options.3The gap isn't willingness — it's billing workflow.
The legacy operational model in most medical practices combines paper statements mailed 30–60 days after service, voice-call follow-up at 15–25 minutes per account, eventual handoff to a collection agency at 25–40% contingency,4 and A/R aging that routinely exceeds 60 days on patient balances.5
The platform closes the loop between invoice creation, multi-channel communication, AI-assisted triage of patient responses, and operational execution — without adding headcount. The rest of this case study quantifies that difference.
This is the strongest apples-to-apples comparison in the report and the best signal of platform effectiveness — linkage-independent, seasonality-matched, driven entirely by in-window behavior.
| Metric | Legacy (123-day equal window) | The platform (123-day post-launch) | Change |
|---|---|---|---|
| Total collected | $233,471 | $535,623 | +129.4% |
| Paying accounts | 1,744 | 2,933 | +68.2% |
| Collections per paying account | $133.87 | $182.62 | +36.4% |
| Transactions per paying account | 1.492 | 1.495 | +0.1% (tie) |
| Median payment size | $50.00 | $63.89 | +27.8% |
| 90th-percentile payment size | $230.58 | $330.19 | +43.2% |
Transactions per paying account are statistically tied (1.492 vs 1.495). Patients aren't being nagged more — they're paying more per contact. That's the signature of better timing, channel fit, and messaging clarity, not more outreach.
Higher median ($50 → $63.89) means the middle of the distribution pays more. Higher P90 ($230.58 → $330.19, +43%) means larger balances convert at a higher cadence. The platform isn't winning only on small balances.
Paying accounts grew +68% even against a peak-season legacy window (Aug–Dec 2025, including Q4 deductible-reset surge). The platform reached more unique patients in the identical 123-day period.
Per HFMA MAP Keys, best-in-class patient collections-per-account benchmarks for imaging groups of comparable size cluster in the $110–$150 range.5 The platform's observed $182.62 per paying account places the practice above the top of that published benchmark range.
This view shows platform performance over time within the launch window. It's the clearest indication that gains are real and compounding — not a one-off driven by launch novelty.
| Metric | First 30 days (Dec 15 – Jan 13) | Last 30 days (Mar 18 – Apr 16) | Change |
|---|---|---|---|
| Daily collections | $2,371 / day | $5,958 / day | +151.3% |
| Paying accounts | 493 | 986 | +100.0% |
| Median payment size | $50.00 | $72.28 | +44.6% |
The last-30-day run-rate of $5,958/day is 3.1× the legacy equal-window baseline of $1,898/day.
If sustained, the annualized collections run-rate is approximately $2.17M — vs. ~$693K/year on the legacy equal-window daily rate. That's a projected ~$1.48M annual uplift on the same patient panel.
Caveat: the last-30-day window includes Q2 2026 deductible-reset activity and the first full cycle of AI send-time optimization after enablement. Both are legitimate, repeatable drivers — but a full year of data is the right basis for a binding financial plan.
This is the direct head-to-head on invoice-level performance — matched windows, identical KPI definitions on both sides.
| Collection horizon | Legacy platform | The platform | Delta |
|---|---|---|---|
| Within 60 days | 51.0% | 66.4% | +15.3 pts |
| Within 90 days | 53.1% | 75.8% | +22.6 pts |
The platform collects 75.8% of patient responsibility by 90 days vs. 53.1% on the legacy platform. That's a 22.6-point absolute advantage, or 42.7% relative. By 60 days, the platform is already +15.3 points ahead.
HFMA MAP Keys reports patient responsibility collection rates of 50–60% at 90 days as the industry average for non-integrated billing workflows.5 The platform's observed 75.8% at 90 days is approximately 15 percentage points above the reported industry median.
For this comparison we hold gross collections constant at the platform's observed $535,623 and compare operating cost and net-after-cost across four alternative operating models.
| Operating model | Basis | Gross collected | Operating cost | Net after cost | Cost as % of collections |
|---|---|---|---|---|---|
| The platform (observed) | Observed | $535,623 | $7,426 | $528,198 | 1.39% |
| Manual phone + mail, 2 FTE (modeled) | Modeled | $535,623 | $43,269 | $492,354 | 8.08% |
| Manual phone + mail, 3 FTE (modeled) | Modeled | $535,623 | $59,433 | $476,190 | 11.10% |
| Collection agency handoff (modeled) | Modeled | $535,623 | $187,468 (35% contingency) | $348,155 | 35.00% |
The platform's operating cost is $7,426 over 123 days — 1.4% of collections.
A modeled 2-FTE manual operation achieving the same gross costs $43,269 — 5.8× the platform. The platform's net exceeds the 2-FTE net by $35,844 even giving the manual team perfect execution.
Against collection-agency handoff at 35% contingency (industry midpoint4), the platform preserves $187,468 of gross collections over 123 days.
Operating cost excludes the platform platform licensing — add subscription for fully-loaded cost-to-collect.
Implied annual uplift vs. legacy run-rate (~$693K/year extrapolation):
| Scenario | Comms volume | Human time | Net the platform cost | Savings vs manual | Revenue retained vs 35% agency |
|---|---|---|---|---|---|
| Lean (2 SMS + 1 email / account, 3 min each) | 4,318 SMS + 2,159 email | 108 hrs | −$604 (net positive) | $26,649 | $187,468 |
| Base case (4 SMS + 2 email, 6 min each) | 8,636 SMS + 4,318 email | 216 hrs | $2,129 | $23,916 | $187,468 |
| High-touch (8 SMS + 3 email, 12 min each) | 17,272 SMS + 6,477 email | 432 hrs | $7,595 | $18,450 | $187,468 |
Even in the highest-touch scenario, total 123-day operating cost stays under $7,600 while preserving $187K that a typical agency referral would have skimmed at 35%. The lean scenario is actually revenue-positive once patient-paid processing-fee revenue is netted against cost.
The platform isn't “statement software plus automation.” It's an AI-assisted patient-billing operating system with twelve live, observable capabilities, each measurable and toggle-controlled. This is the mechanism by which the financial results above were produced.
Generates prioritized, approval-gated settlement campaigns. AI scoring with operational guardrails (cooldown, template validation, opt-out respect). Every candidate volume, balance, and skip reason is instrumented.
Analyst-driven hybrid mode. AI recommendations combined with deep filters (days past due, balance, DOS range, payment plans, communication attempts). Supports AB cohorts.
Multi-option settlement notices with configurable discounts, terms, and deadlines. Email, mail PDF, or both. Patient response portal captures selections and timing.
Classifies inbound SMS into structured intents — payment confirmation, link request, confusion, payment-plan request, dispute, stop, wrong number. Converts unstructured messages into operational signals.
Autonomous reply for low-risk, high-confidence intents (link-request, thank-you, wrong-number). Generates payment links on demand. Deflects work from the human queue.
Detects dispute intent, auto-creates dispute records, classifies type, and produces AI summaries. Dispute queues and reply workflows live inside the messaging UI.
Per-account risk score and recommended action. Staff worklist orders the queue in real time by risk and payment probability — predictive prioritization embedded in daily workflow.
Predicts payment probability per account. Aggregates expected collections 7 and 14 days out. Segments high-risk vs. likely-to-pay cohorts for cash-planning.
Patient-specific send-time recommendations using age band, ZIP-income profile, timezone, and communication history. Tracks strategy cohorts vs. control and measures conversion lift.
Recommends down-payment, installment count, monthly amount, and auto-draft suitability. Applies recommendation directly to create a plan — tied to execution, not just advisory.
Automatically creates AI-generated follow-ups for high- and medium-risk accounts. Load-balances assignments to available collections staff. Bridges prediction to task orchestration.
Downloadable CSVs — Invoice/AR, Payment Ledger, Communication Log, Invoice-to-Cash Timeline. Reproducible evidence base for every number in this case study.
This section exists because a case study with no caveats is not credible. Everything we know could be wrong with these numbers — disclosed explicitly.
The legacy platform stopped producing transactions on Dec 12, 2025 (cutover). We can compare the platform against the last 123 days of legacy operation but cannot test it on post-Dec-15 patient flow. If Aug–Dec 2025 was atypical (staffing shift, insurance-mix change, seasonal oddity), the baseline is off by that amount. Spot-checks of the full-period vs equal-window stats are directionally consistent ($147/account full-period vs $134/account equal-window), so we don't believe this is large — but it can't be ruled out.
The platform timestamps invoices at date-of-service; the legacy platform timestamps them at bill-send (typically ~30 days later, after claims processing). A +30-day legacy adjustment is applied where windows are compared (sensitivity at +15/+30/+45 in the appendix). We do NOT lead with speed-to-cash as a headline metric because of this — the hero claims here are per-account economics, 60/90-day collection rates, and cost-to-collect, all of which are maturity-adjusted or clock-independent.
We don't have a real parallel 2- or 3-FTE manual billing team to measure against. Manual numbers are built on HFMA labor benchmarks and BLS 2024 wage data. Directionally reliable (standard industry arithmetic) but should not be cited as observed field data.
Analysis treats all patient-responsibility dollars as fungible. High-deductible commercial vs. Medicare Part B vs. self-pay have very different collection dynamics. Practice mix is approximately stable across both windows (same clinic, same service lines), but a full multivariate controls model would be a reasonable follow-on.
The $535,623 total is system-of-record (channel-inclusive). Under digital-rail-only attribution (payments the platform processed on its own rail: $275,155), uplift is +17.9%. Under digital-rail + payment-plan ($304,767), uplift is +30.5%. Under all three framings, the platform exceeds the legacy baseline. The offline-recorded payments are a legitimate part of the platform-of-record total but are not claimed as AI-caused dollars.
123 days passes statistical significance on volume metrics and spans two deductible-reset cycles — but it's not full seasonality. Any pro-forma built on last-30-day run-rate × 365 should be treated as an upper-bound projection until at least six months of post-launch data exist.
Every benchmark range cited in this case study is drawn from a named, verifiable source below. Observed the platform numbers come from the platform's own export CSVs and are reproducible via the pipeline scripts in the report appendix.
Cited for: Average single-coverage deductible $1,790 in 2024; HDHP enrollment growth over the last decade.
Cited for: 41% of U.S. adults currently have some debt caused by medical or dental bills; 57% have had medical debt in the last five years.
Cited for: Provider collection rates of 50–70% for small-dollar insured balances; 90%+ of patients willing and able to pay balances under $500 with right payment mechanisms; inefficient billing and lack of financing options are primary drivers of uncollected revenue, not patient willingness.
Cited for: Collection agencies typically charge 25–40% contingency on medical debt; rates rise to 30–50% for debts aged 1–2 years, and 50%+ for older accounts.
Cited for: HFMA MAP Keys define 29 strategic KPIs across Patient Access, Pre-Billing, Claims, Account Resolution, and Financial Management — including cash collection as % of net patient service revenue and patient responsibility collection benchmarks.
Cited for: Median annual wage for medical records specialists: $50,250 (May 2024). Used as base-rate input for fully-loaded labor cost in the manual-comparison model.
Cited for: Best-practice cost-to-collect benchmark has moved from 3% to 2% of collections over the past 5–7 years, driven by digital-first billing platforms.
All numeric claims in this case study that are not directly observed from the platform's export data are sourced from the above. Observed the platform data is derived from four CSV exports (Invoice/AR Detail, Payment Transaction Ledger, Statement/Communication Log, Invoice-to-Cash Timeline) via the scripts/run_report.py and scripts/build_hero_report.py reproducibility pipeline.
External licensing opens Q2 2026. Twenty-minute call covers fit, integration with your PM system, and deployment timeline.