Healthcare value chain
Follow the money through 17 handoffs. Then watch AI reroute it.
Healthcare is a $4.5 trillion industry that moves every dollar through 17 distinct activities, from the moment a patient registers to the final collection call. Each activity has its own margin structure, its own workforce, and its own exposure to AI. This is not a survey of use cases. It is a map of where the profits sit and where they are heading.
The value does not disappear from healthcare. It migrates. The map below shows exactly where.
Reading the pool
The visualization below maps healthcare's 17 activities by two dimensions. Bar width shows each activity's share of total revenue. Care delivery takes 35 cents of every dollar, while payment posting takes one cent. Bar height shows operating margin. Claims adjudication runs at 18%, while hospital care delivery runs at 6%.
The product of width and height is the profit pool: the actual dollars captured. Care delivery generates the most total profit ($94.5B) despite thin margins because it handles the most revenue. Claims adjudication generates $121.5B in payer profit despite handling only 15% of revenue because its margins are three times higher.
Most healthcare AI coverage asks "which activity gets automated first." The profit pool asks the better question: where does the displaced margin go?
The healthcare profit pool
Revenue share and margin concentration across 17 healthcare activities. Bar width = revenue share. Bar height = operating margin. Color = player concentration.
Margin concentration across the chain
Three activities capture disproportionate margin. Claims adjudication runs at 18% because payers keep the gap between premiums collected and claims paid. Regulated at 80-85% medical loss ratio, the remaining 15-20% is operating margin. Denial management runs at 20% because overturning a denial recovers $2K-50K per claim against $10-15 in labor cost. Charge capture runs at 15% because coding accuracy is the single largest determinant of whether care gets reimbursed correctly.
All three are pattern-matching activities. Read document, match rules, produce output. The same structure that makes them high-margin makes them high-exposure to AI.
The highest-margin activities in healthcare are the most exposed to AI displacement. They are high-margin precisely because they require specialized knowledge that is expensive to hire. AI now replicates that knowledge at marginal cost.
Margin concentration map
Each activity's margin plotted against revenue share. High-margin activities on the right capture disproportionate profit relative to their revenue.
Four ways AI hits the value chain
Displaced: the activity's core function is pattern matching: read input, apply rules, produce output. AI automates 70-90% of volume. Benefits verification, prior auth, coding, claims submission, adjudication, denial management, payment posting, and collections all fit this pattern. The margin compresses because the labor cost that justified it evaporates.
Accelerated: AI makes the human faster and better, but the human remains essential. Clinical documentation, care coordination, scheduling, patient billing, quality reporting, and clinical decision support. Physicians still diagnose; AI captures the note. Coordinators still manage transitions; AI predicts which patients will bounce back.
Augmented: care delivery itself. The physician-patient encounter remains human. AI assists at the edges: diagnostic suggestions, drug interaction alerts, risk scores. Margin impact is indirect. Better documentation downstream means better coding means higher reimbursement.
Compressed: patient engagement and acquisition. AI reduces the cost to acquire and retain patients, but the savings get competed away. Every health system gets the same AI marketing tools, so the advantage is temporary. Margins compress toward a new, lower equilibrium.
Eight of the 17 activities face displacement. Five get accelerated. One gets augmented. One gets compressed. The net is not fewer jobs. Different jobs at different margins. at different margins.
Healthcare profit pool data
Revenue share, margin, AI impact, and key players for each of 17 value chain activities.
| Activity | Revenue share | Margin | Profit ($B) | AI impact | Players |
|---|---|---|---|---|---|
| Patient access | 3.0% | 2.0% | $2.7 | low | Health systems (60%)Vendor platforms (25%)Staffing firms (15%) |
| Benefits verification | 2.0% | 3.0% | $2.7 | high | RCM vendors (45%)Payer platforms (35%)Health systems (20%) |
| Scheduling | 2.0% | 2.0% | $1.8 | medium | Health systems (50%)Scheduling platforms (30%)Call centers (20%) |
| Prior authorization | 2.0% | 1.0% | $0.9 | high | Health systems (40%)UM vendors (35%)Payers (25%) |
| Clinical documentation | 5.0% | 3.0% | $6.8 | high | Ambient AI (Abridge, Nuance) (35%)Health systems (40%)Scribe services (25%) |
| Clinical decision support | 2.0% | 4.0% | $3.6 | medium | EHR vendors (Epic, Oracle) (45%)CDS platforms (30%)Health systems (25%) |
| Care delivery | 35.0% | 6.0% | $94.5 | medium | Health systems (50%)Physician groups (30%)Telehealth platforms (20%) |
| Charge capture & coding | 3.0% | 15.0% | $20.3 | high | RCM vendors (40%)Coding services (35%)Health systems (25%) |
| Claims submission | 4.0% | 12.0% | $21.6 | high | Clearinghouses (40%)RCM vendors (35%)Health systems (25%) |
| Claims adjudication | 15.0% | 18.0% | $121.5 | high | Commercial payers (55%)Government programs (30%)TPA vendors (15%) |
| Denial management | 3.0% | 20.0% | $27 | high | RCM vendors (45%)Health systems (35%)Specialty appeal firms (20%) |
| Payment posting | 1.0% | 8.0% | $3.6 | high | RCM vendors (50%)Health systems (30%)Clearinghouses (20%) |
| Patient billing | 4.0% | 5.0% | $9 | medium | Health systems (40%)Patient payment platforms (35%)Collection agencies (25%) |
| Collections | 3.0% | 10.0% | $13.5 | high | Collection agencies (40%)RCM vendors (35%)Health systems (25%) |
| Quality & compliance | 3.0% | 3.0% | $4.1 | medium | Health systems (45%)Quality analytics vendors (30%)Consulting firms (25%) |
| Care coordination | 6.0% | 4.0% | $10.8 | medium | Health systems (45%)Care management platforms (30%)Post-acute providers (25%) |
| Patient engagement | 7.0% | 14.0% | $44.1 | medium | Health systems (35%)Digital health platforms (30%)Marketing agencies (20%)CRM vendors (15%) |
The 24-month forecast
Watch the margins move in real time
Claims adjudication margins fall from 18% to 10%. Clinical documentation generates 3-8% more revenue through coding accuracy. The AI shift timeline maps the sequence: which activities to rebuild first, what tools to use, and where the margin compounds.
Read the AI shift timelineBy activity
Explore each part of the value chain
Clinical documentation→
Accelerated: Ambient AI captures the encounter. Physicians reclaim 1-2 hours/day. The scribe becomes the auditor.
Claims processing→
Displaced: Submission, adjudication, and denial management. 5.5B claims/year, 10-15% denied. AI auto-adjudicates routine claims.
Revenue cycle management→
Displaced: The end-to-end cycle from registration to collection. 17 handoffs, each with its own AI exposure.
Benefits verification & prior auth→
Displaced: The $35B bottleneck. AI reads plan documents faster than humans. 80%+ of verifications become zero-touch.
Scheduling & patient access→
Accelerated: 15-25% no-show rates, 75-85% utilization. AI predicts, backfills, and self-serves.
Care delivery→
Augmented: The one activity AI assists but does not replace. Diagnostic support, staffing optimization, virtual care.
Care coordination→
Accelerated: Referrals, transitions, chronic care. AI identifies high-risk patients early. Panel sizes increase 2-3x.
Clinical decision support→
Accelerated: From alert fatigue (90% override rate) to contextual, patient-specific guidance.
Medical coding→
Displaced: AI reads documentation and suggests codes at 95%+ accuracy. Human coders shift to auditing.
Patient billing & collections→
Displaced: The last mile of revenue. AI prioritizes by propensity to collect. AR days compress 5-10 days.
Quality & compliance→
Accelerated: Quality scores swing 2-9% of Medicare revenue. AI turns retrospective reporting into real-time gap detection.
Patient engagement→
Compressed: AI reduces acquisition cost from $1,200 to $400. But every competitor gets the same tools.
The profit pool maps where the money sits. The healthcare AI thesis maps where it moves. Read them together.
Common questions
What is a healthcare profit pool?
A profit pool maps an industry's total profit by activity. Instead of looking at one company's margins, it shows where profit concentrates across the entire value chain, showing which activities capture the most margin relative to their revenue share. The concept comes from Bain & Company; we apply it specifically to healthcare with AI shift projections.
How does AI affect healthcare margins?
AI affects margins four ways: it displaces high-margin pattern-matching activities (claims, coding, verification), accelerates human-centered activities (documentation, coordination), augments clinical care delivery without replacing it, and compresses patient acquisition margins by giving every competitor the same tools.
Which healthcare activities are most exposed to AI displacement?
Eight activities face displacement: benefits verification, prior authorization, charge capture, claims submission, claims adjudication, denial management, payment posting, and collections. All share a common trait: their core function is reading structured input, applying rules, and producing structured output.
Will AI replace healthcare workers?
AI changes which jobs exist, not how many. Displaced activities lose volume-processing roles but gain oversight and exception-handling roles. A medical coder becomes a coding auditor. A claims examiner becomes a fraud analyst. The net headcount may shrink, but the remaining roles require more judgment, not less.
Where does the profit pool data come from?
Revenue and margin estimates are derived from three public sources: Bureau of Labor Statistics occupational employment data (what roles exist and what they cost), SEC 10-K filings from UnitedHealth, HCA, and CVS/Aetna (segment-level revenue and margin disclosures), and HFMA revenue cycle benchmarks.
How is this different from McKinsey or Bain healthcare reports?
Three differences: it is interactive (not a static PDF), it maps AI impact per activity with specific margin projections (not "AI will transform healthcare"), and it is free and ungated. McKinsey and Bain publish broad surveys; this maps 17 specific activities with revenue, margin, and displacement trajectory for each.