← All Industry solutions
AICost Industry Solutions

Hospitals and health systems

3 AI automations this industry runs today, each with the process, the AICost engine chain that prices and governs it, and the evidence behind it.

Cost engine Governance engine Each step priced and controlled. Click any engine to open its calculator or playground.
Ambient clinical documentationPatient message (In Basket) triage and draft repliesRevenue cycle: coding, denials and appeal letters

Ambient clinical documentation

Owner: CMIO / CNIO office, physician and nursing informatics, clinical operations; used by physicians, APPs and increasingly nurses

See a chain run live Build this chain for our stack
Scale: Ambient documentation is the largest healthcare AI spend category, about $600M in 2025 (Menlo Ventures); Abridge expects more than 50M medical conversations in 2025; Northwell is scaling to its 20,000+ physicians

How the work runs

  1. Get patient consent and start the ambient session from the EHR mobile app
  2. Record and transcribe the encounter
  3. Generate a structured note (HPI, exam, assessment and plan) plus orders and codes
  4. Clinician reviews, edits and signs in the EHR
  5. Downstream: patient instructions, coding suggestions, quality capture

AICost engine chain

6 cost engines 4 governance engines

Each step priced and controlled. Click any engine to open its calculator or playground.

Where cost leaks

  • Per-clinician seat licenses bought enterprise-wide while active use is low (TPMG: 3,442 of about 10,000 offered physicians adopted)
  • Long encounters and re-generation of notes by specialty template increase token and audio minutes
  • Paying separately for overlapping ambient tools (EHR-native and third party)

Where risk sits

  • Recording without proper consent (class actions against Sharp HealthCare and other systems)
  • Hallucinated exam findings or meds signed into the legal record
  • Upcoding risk from AI-suggested codes

Rules that apply

  • HIPAA (BAA, minimum necessary)
  • state two-party recording consent laws (e.g., California CIPA)
  • California AB 3030 for AI-generated patient communications
  • ONC HTI-1 decision-support transparency where applicable

Runs on

Epic (Art / Haiku, Epic native ambient)AbridgeMicrosoft Dragon Copilot (DAX)Ambience HealthcareOracle Health Clinical AI Agent

Evidence

abridge.com ↗ama-assn.org ↗fisherphillips.com ↗

Patient message (In Basket) triage and draft replies

Owner: Ambulatory operations, nursing triage pools, physicians; patient access / digital front door team

See a chain run live Build this chain for our stack
Scale: Providence: about 150,000 monthly active users on its navigation platform, 30% of administrative MyChart messages deflected (Becker's)

How the work runs

  1. Patient sends a portal message (MyChart) or chats with a digital assistant
  2. AI self-service deflects administrative requests (scheduling, refills, navigation)
  3. Classify and route the remaining messages to the right pool (clinical, billing, refill)
  4. Draft a reply using chart context
  5. Clinician edits and sends, with an AI-disclosure footer where required

AICost engine chain

3 cost engines 4 governance engines

Each step priced and controlled. Click any engine to open its calculator or playground.

Where cost leaks

  • Drafting replies for every message, including ones the clinician deletes
  • Pulling the full chart into context for simple administrative messages
  • Volume spikes (portal messages keep rising) on per-message pricing

Where risk sits

  • Unsafe clinical advice in a draft that is sent without a real edit
  • Missing AB 3030 disclosure
  • Urgent symptoms misrouted or deflected by a chatbot

Rules that apply

  • HIPAA
  • California AB 3030 (disclaimer plus a way to reach a human unless a licensed provider reviewed the message)
  • state AI disclosure laws (e.g., Utah)

Runs on

Epic In Basket Art (AI draft replies)MyChartProvidence Grace / ProvARIAMicrosoft Azure OpenAI

Evidence

beckershospitalreview.com ↗emarketer.com ↗telehealth.org ↗

Revenue cycle: coding, denials and appeal letters

Owner: Revenue Cycle (HIM / coding, patient financial services, denials management), CFO

See a chain run live Build this chain for our stack
Scale: Coding and billing automation was about $450M of 2025 healthcare AI spend (Menlo Ventures)

How the work runs

  1. Autonomous or assisted coding from the documentation
  2. Pre-bill claim scrubbing and denial-risk prediction
  3. Work denials: classify the reason and pull supporting chart evidence
  4. Draft the appeal letter with clinical and payer-policy citations
  5. Submit, track and feed results back into prevention rules

AICost engine chain

5 cost engines 3 governance engines

Each step priced and controlled. Click any engine to open its calculator or playground.

Where cost leaks

  • Full-chart context for every claim line instead of only the relevant encounter
  • Agentic payer-portal bots that retry and loop
  • An AI-vs-AI 'battle of the bots' with payers that drives up volume on both sides

Where risk sits

  • Systematic upcoding or unsupported codes (FCA liability)
  • Fabricated clinical citations in appeals
  • PHI sent to non-BAA tools

Rules that apply

  • HIPAA
  • False Claims Act exposure for upcoding
  • CMS and OIG coding compliance
  • payer contract terms

Runs on

Epic ResoluteOracle Health RCMWaystar (AltitudeAI)R1 RCMOptum3M/Solventum 360 EncompassExperian Health

Evidence

hfma.org ↗waystar.com ↗healthcare.digital ↗

Reference chains are a starting point. Your volumes, models and vendors change the numbers; the engines recompute them from live prices.

Your process is not on the list, or runs differently? These chains are a starting point. We assemble the right engines for your workflow, your models and your vendors, and you run it over MCP or in the calculators.

Business solutions → Build my chain
📖 Data sources & methodology 172 text models · 9 embeddings · 44 vision · 62 audio · 8 vector DBs across 10 vendor pages · last verified 2026-10-15

Methodology

  • All prices are USD per 1 million tokens, current as of 2026-10-15.
  • Vendor-published values have no mark. Inferred/extrapolated values are marked with * and listed below.
  • Batch API discounts are 50% off standard rates across providers that offer Batch mode.
  • Prompt caching discounts vary by provider (typically 80-90% off cached input tokens).
  • Regional data-residency surcharges (Anthropic 1.1x, OpenAI 1.1x, Google regional tiers) are NOT included in base rates.
  • Long-context pricing tiers apply when input exceeds model threshold.
  • Embedding prices are input-only (no output tokens generated).

Primary sources

Last-verified date is the most recent successful daily snapshot (aicost_pricing_snapshots) or, when no snapshot exists yet, the latest successful crawler run (aicost_crawler_runs). 10 of 10 vendors are currently verified. Aggregator services (TokenCost, AI Pricing Guru, etc.) are not listed.

Anthropic
2026-10-14
https://www.anthropic.com/pricing
Daily snapshot since Sep 2023 · 700 days captured
Anthropic Docs
2026-10-14
https://platform.claude.com/docs/en/about-claude/pricing
Daily snapshot since Sep 2023 · 700 days captured
OpenAI
2026-10-14
https://openai.com/api/pricing/
Daily snapshot since Sep 2023 · 701 days captured
Google AI
2026-10-14
https://ai.google.dev/gemini-api/docs/pricing
Daily snapshot since Dec 2023 · 676 days captured
Google Vertex
2026-10-14
https://cloud.google.com/vertex-ai/generative-ai/pricing
Daily snapshot since Dec 2023 · 676 days captured
DeepSeek
2026-10-14
https://api-docs.deepseek.com/quick_start/pricing
Daily snapshot since May 2024 · 615 days captured
xAI
2026-10-14
https://x.ai/api
Daily snapshot since Nov 2024 · 533 days captured
Mistral
2026-10-14
https://mistral.ai/pricing
Daily snapshot since Dec 2023 · 674 days captured
Cohere
2026-10-14
https://cohere.com/pricing
Daily snapshot since Sep 2023 · 700 days captured
Voyage AI
2026-10-15
https://docs.voyageai.com/docs/pricing

Inferred values (marked with * in calculator tables)

Derived from industry conventions, not directly published by the vendor. Typical conventions: cached input = 10% of base (90% off), Batch API = 50% of base (50% off).

Vendor / Model Field Why it’s inferred
Anthropic — Claude Sonnet 4.6 cachedInput Derived at 10% of input rate — Anthropic publishes 90% cache-hit discount on this tier.
Anthropic — Claude Sonnet 4.5 cachedInput Derived at 10% of input rate; same 90% cache-hit convention as Sonnet 4.6.
Anthropic — Claude Sonnet 4.5 batchInput Derived at 50% of standard input — Anthropic documents uniform 50% Batch discount.
Anthropic — Claude Sonnet 4.5 batchOutput Derived at 50% of standard output — Anthropic documents uniform 50% Batch discount.
Anthropic — Claude Haiku 4.5 cachedInput Derived at 10% of input rate — Anthropic 90% cache-hit discount convention.
OpenAI — GPT-5.4 Mini cachedInput Derived at 10% of input — OpenAI documents automatic 90% discount on cache hits across GPT-5.x tier.
OpenAI — GPT-5.4 Nano cachedInput Derived at 10% of input — OpenAI 90% cache-hit convention.
OpenAI — GPT-5.4 Nano batchInput Derived at 50% of input — OpenAI Batch API uniform 50% discount.
OpenAI — GPT-5.4 Nano batchOutput Derived at 50% of output — OpenAI Batch API uniform 50% discount.
OpenAI — GPT-5.4 Pro cachedInput Derived at 10% of input — OpenAI 90% cache-hit convention.
OpenAI — GPT-5.4 Pro batchInput Derived at 50% of input — OpenAI Batch API uniform 50% discount.
OpenAI — GPT-5.4 Pro batchOutput Derived at 50% of output — OpenAI Batch API uniform 50% discount.
OpenAI — GPT-5.2 cachedInput Derived at 10% of input; no residency uplift.
OpenAI — GPT-5.2 batchInput Derived at 50% of input.
OpenAI — GPT-5.2 batchOutput Derived at 50% of output.
OpenAI — GPT-5 cachedInput Derived at 10% of input.
OpenAI — GPT-5 batchInput Derived at 50% of input.
OpenAI — GPT-5 batchOutput Derived at 50% of output.
OpenAI — GPT-5.5 Pro cachedInput Derived at 10% of input — OpenAI does not publish a cached rate for *-pro models; using the family convention.
OpenAI — GPT-5.5 Pro batchInput Derived at 50% of input.
OpenAI — GPT-5.5 Pro batchOutput Derived at 50% of output.
OpenAI — GPT-5.2 Pro cachedInput Derived at 10% of input — pro-tier convention.
OpenAI — GPT-5.2 Pro batchInput Derived at 50% of input.
OpenAI — GPT-5.2 Pro batchOutput Derived at 50% of output.
OpenAI — GPT-5.1 batchInput Derived at 50% of input.
OpenAI — GPT-5.1 batchOutput Derived at 50% of output.
OpenAI — GPT-5 Pro batchInput Derived at 50% of input.
OpenAI — GPT-5 Pro batchOutput Derived at 50% of output.
OpenAI — GPT-5 Nano cachedInput Derived at 10% of input.
OpenAI — GPT-5 Nano batchInput Derived at 50% of input.
OpenAI — GPT-5 Nano batchOutput Derived at 50% of output.
Google — Gemini 3 Flash cachedInput Derived at 10% of input — Google caching discount convention ~90%.
Google — Gemini 3.1 Flash-Lite cachedInput Derived at 10% of input — Google caching convention.
Google — Gemini 3.1 Flash-Lite batchInput Derived at 50% of input — Google Batch API uniform 50% discount.
Google — Gemini 3.1 Flash-Lite batchOutput Derived at 50% of output — Google Batch API uniform 50% discount.
Google — Gemini 2.5 Pro cachedInput Derived at 10% of input.
Google — Gemini 2.5 Flash cachedInput Derived at 10% of input.
Google — Gemini 2.5 Flash-Lite cachedInput Derived at 10% of input — Google caching convention.
Google — Gemini 2.5 Flash-Lite batchInput Derived at 50% of input — Google Batch API uniform 50% discount.
Google — Gemini 2.5 Flash-Lite batchOutput Derived at 50% of output — Google Batch API uniform 50% discount.
xAI — Grok 4 (legacy) cachedInput Extrapolated at 25% of base.

Pricing is cross-verified against the LiteLLM community registry when available. Daily snapshots are kept in aicost_pricing_snapshots; every change is logged to aicost_price_changelog with old & new values for full audit trail. Read the full methodology →