AI Prior Authorization Platform Comparison: Six Vendors Reshaping PA Automation

Head-to-head analysis of six AI-native prior authorization platforms — Infinitus Systems, Tandem AI, Cohere Health, Coral AI, Neon Health, and SuperDial — spanning voice AI, document AI, and agentic workflow architectures, with implications for CMS-0057-F compliance and pharma manufacturer hub operations.

Thesis

The strategic thesis is that CMS-0057-F raises the floor for electronic PA but does not eliminate the administrative work around specialty drug access. Payer-side AI such as Cohere benefits directly from API and decisioning modernization. Voice and document vendors remain relevant where commercial, Part D, fax, portal, and exception workflows persist. Hub-embedded vendors such as Neon matter when the manufacturer needs AI to connect BV, PA, copay, and pharmacy routing rather than optimize a single PA step.

Executive Summary

Prior authorization remains one of the largest administrative cost centers in U.S. healthcare. A Health Affairs analysis estimated that drug utilization management costs payers, manufacturers, physicians, and patients approximately $93.3 billion annually. The AI vendor question is narrower than that system-wide burden: which workflow layer is being automated, who controls the data, and whether the automation remains useful as CMS-0057-F pushes payer-side electronic PA APIs into production.

Within the emerging AI PA vendor landscape, six companies represent the most strategically distinctive plays, each betting on a different architectural thesis:

Vendor	Architecture	Primary Customer	Total Funding	Valuation
Infinitus Systems	Voice AI (payer calls)	Pharma manufacturers, specialty pharmacies	$102.9M	~$600M
Tandem AI	Full medication access (PA + pharmacy + affordability)	Healthcare providers	$137M	$1B
Cohere Health	Payer-side AI auto-determination	Health plans / payers	$200M+	Private
Coral AI	Document AI (intake / PA prep)	Providers (DME, infusion, SP)	$2M (seed)	Not disclosed
Neon Health	End-to-end hub automation	Pharma manufacturers, hub partners	$6M (seed)	Not disclosed
SuperDial	Voice AI (provider-side)	RCM companies, DSOs, MSOs	$20M+	~$20M post-money

Three structural observations organize the rest of this analysis. First, these vendors are not direct competitors across the board — they address different segments of the PA workflow (submission vs. verification vs. decisioning) and sell to different buyers (provider vs. payer vs. manufacturer). Second, capital raised has decoupled from operating maturity: Tandem AI’s $1B valuation on $137M raised and zero disclosed revenue stands in sharp contrast to Infinitus’s more conservative $600M valuation on ~$100M raised despite 44% Fortune 50 penetration. Third, the CMS-0057-F mandate is the biggest single demand catalyst, but its impact is asymmetric — it favors payer-side platforms (Cohere) and FHIR-ready submission tools while potentially disintermediating phone-based intermediaries over a 3-5 year horizon.

Two specialist entrants sit just outside the core six but matter for buyer pathing. 100ms brings voice-first intake, benefits, and authorization workflow automation with a specialty-access orientation, while RISA Labs is building an oncology-specific operating system spanning benefits verification, prior auth, denial management, and status tracking. They are not yet broad market leaders, but they are precisely the kind of focused entrants a buyer can miss if the analysis layer only surfaces the largest funded names.

Architectural Segmentation

The PA workflow breaks into five discrete sub-workflows. Each vendor maps to one or more:

1. Benefit verification (BV): Confirm deductible, copay, coverage details before PA submission
2. PA submission: Assemble clinical documentation, complete payer-specific forms, transmit to payer
3. PA decisioning: Clinical review and auto-determination by the health plan
4. PA status follow-up: Track pending PAs through approval/denial; handle peer-to-peer scheduling
5. Pharmacy / fulfillment routing: Match approved PA to optimal pharmacy for fill

Vendor	BV	Submission	Decisioning	Follow-Up	Pharmacy
Infinitus Systems	Yes	Partial	No	Yes (voice)	No
Tandem AI	Yes	Yes	No	Yes	Yes
Cohere Health	No	Indirect	Yes	Yes	No
Coral AI	Partial	Yes (prep)	No	No	No
Neon Health	Yes	Yes	No	Yes	Partial
SuperDial	Yes	Partial	No	Yes (voice)	No

Voice AI (Infinitus, SuperDial, Neon) bets that much of PA’s latency comes from payer phone calls — IVR navigation, hold time, representative conversation — and that conversational AI can reduce manual call-center work. Infinitus and Neon target the pharma hub / specialty pharmacy workflow; SuperDial targets provider-side RCM.

Document AI (Coral) bets that PA’s latency comes from document assembly and payer-specific form completion — that faxes, referral packets, and insurance card images are the choke point. Coral automates the document-processing layer and uses portal automation (not voice) to submit.

Agentic decision AI (Cohere, Tandem) bets on higher-order automation. Cohere sits on the payer side and auto-approves PAs against clinical criteria before they reach a human reviewer. Tandem sits on the provider side with a broader “medication access” scope including pharmacy routing and affordability — its product is free for providers and patients, monetized through downstream fees.

Specialist Entrants Worth Watching

100ms sits between the voice-AI and intake-orchestration camps. It is not as proven at payer-call scale as Infinitus Systems, but it is more naturally positioned for buyers who want one automation layer across intake, benefits, prior auth triage, scheduling, and specialty-access coordination rather than a single post-submission call bot.

RISA Labs is the oncology specialist in the field. The product posture is closer to a cancer-service-line operating system than a generic PA tool: it spans benefits, authorizations, denials, and status management across cancer centers, infusion networks, and specialty pharmacies. For oncology-heavy manufacturers or providers, it belongs on the shortlist even if it is not yet a general-market platform.

Competitive Deep Dive

Infinitus Systems — The Voice AI Category Leader

Founded in 2019 by Ankit Jain (ex-Google AI) and Shyam Rajagopalan (ex-Snap), Infinitus pioneered the voice-AI-for-payer-calls category. As of April 2025 it had crossed 6 million total calls and 100 million conversation minutes automated. Eight of the top 10 pharma companies are customers. 44% of Fortune 50 healthcare companies use the platform. Its flagship capability is fully autonomous outbound voice AI that navigates IVRs, waits on hold, conducts live conversations, and returns structured data via API.

Differentiators:

• Proprietary knowledge graph: payer-call history continuously trains IVR navigation, payer procedures, and real-time coverage rules
• Discrete action space architecture: Hallucination-prevention layer sitting on top of LLMs to restrict agent behavior to approved flows — a durable compliance moat
• Scale: 100M+ automated conversation minutes publicly reported
• Enterprise integrations: Salesforce Life Sciences Cloud, Epic via SMART on FHIR, Google Vertex AI, IBM Consulting partnership

Weaknesses: Voice-only architecture leaves document AI and payer-side decisioning to others. Phone-channel dependency could be disintermediated by CMS-0057-F FHIR APIs over 3-5 years (Infinitus has started positioning itself as complementary to FHIR for non-compliant payers).

Pricing (directional): Enterprise per-call / per-automation pricing; not publicly disclosed.

Tandem AI — The Billion-Dollar Wildcard

Founded in 2023 by Sahir Jaggi (ex-Oscar Health product leader), Tandem emerged from stealth with a distinctive positioning: free for providers and patients, monetized through the value chain rather than direct fees. In January 2026 it raised $100M at a $1B valuation led by Accel, with Thrive Capital and General Catalyst participating — the highest valuation of any AI PA startup.

Differentiators:

• Full medication access scope: Beyond PA, covers pharmacy routing and affordability navigation — a broader workflow than any other vendor in this set
• Zero-cost adoption model: Eliminates the primary procurement friction for providers
• AI-native architecture: Built from scratch without legacy fax/phone workflows
• Capital: $137M raised in three years vs. Infinitus’s $102.9M in six years — most capital of any AI PA startup

Weaknesses:

• Revenue model opacity: The free-for-provider approach raises monetization questions (pharmacy network fees? manufacturer partnerships? payer analytics?). No disclosed revenue.
• Early stage: Founded 2023, ~40-50 employees — product maturity and specialty drug PA depth unproven at enterprise level
• Crowded market: Competing against well-funded rivals (Infinitus, CoverMyMeds with McKesson backing)
• Valuation risk: $1B on $137M raised implies high expectations with limited revenue evidence — the closest analog (GoodRx monetizing via pharmacy fees) took 10+ years to build that network

Strategic significance: If Tandem scales, its broad scope and free-for-provider model could become the category defining platform for primary care PA — but specialty drug PA (where pharma manufacturer value is concentrated) remains an open question.

Cohere Health — The Payer-Side Agentic Incumbent

Founded in 2019 by Siva Namasivayam, Cohere is the only vendor in this set that sits firmly on the payer side of the PA transaction. It processes more than 12 million PA requests annually for 600,000+ providers through its health plan customers and reports up to 85% real-time auto-approval rates. Series C of $90M in May 2025 (led by Temasek) brought total funding above $200M.

Differentiators:

• Clinically trained AI model library: Auto-determination across multiple clinical domains, with medical director oversight for edge cases
• Payer-side positioning: Directly addresses the health plan’s need to modernize UM amid CMS-0057-F regulatory pressure
• Platform expansion: September 2025 ZignaAI acquisition extends Cohere’s AI from pre-service (PA) into post-service (payment integrity) — a pre-service-to-post-service clinical AI pipeline that legacy UM vendors lack
• Provider-experience positioning: a deliberate differentiator vs. the adversarial reputation of legacy UM outsourcers

Weaknesses:

• Scale gap vs. eviCore: eviCore covers 100M+ lives; Cohere’s footprint, while growing rapidly, is significantly smaller
• Payer-only model limits TAM: No direct access to pharma manufacturers or provider revenue pools
• Execution risk on ZignaAI: Payment integrity has a different buyer (CFO/ops) and workflow — integration is non-trivial

Strategic significance: CMS-0057-F mandates electronic PA for Medicare Advantage and Medicaid plans by January 2027. Cohere is the best-positioned AI-native payer-side vendor to capture health-plan share from legacy UM incumbents — and its API-plus-AI stack (vs. API-only middleware) is the most defensible value proposition for payer buyers.

Coral AI — The Document-AI Dark Horse

Founded in 2024 by Ajay Shrihari and embedded with clinic staff for months before writing code, Coral AI is a document intelligence platform — not voice AI. It uses computer vision, OCR, NLP, and LLMs to read inbound referral packets and faxes, extract structured data, and prepare PA requests, all as an overlay on providers’ existing systems. Seed round of $2M in early 2025 led by Lightspeed India Partners.

Differentiators:

• Within-existing-systems architecture: Overlay approach means no rip-and-replace — fast time-to-value
• Document-heavy workflow focus: Automates the PA prep that happens before payer phone calls — complementary to Infinitus/SuperDial
• Rapid revenue growth: 8x revenue growth in 7 months; 500K+ patient workflows per month by late 2025
• Strategic advisor: Ashish Singh (Senior Partner, Bain & Company; founder of Bain’s healthcare practice) joined October 2025

Weaknesses:

• Severely undercapitalized: $2M seed vs. Tennr’s $162M for the same document-AI-for-healthcare-intake problem — Coral either needs a Series A quickly or risks losing the category
• No disclosed compliance credentials: SOC 2 / HIPAA certifications not public, potential enterprise procurement barrier
• Commoditization risk: Document processing and portal automation are increasingly commoditized capabilities — EHR vendors, clearinghouses, and better-funded startups can build similar solutions
• Small customer base: DASCO Home Medical Equipment is the primary named customer; no pharma manufacturer or health plan logo yet

Acquisition path: Coral is the most likely acqui-hire target in this set — a larger RCM company, EHR platform, or Infinitus/SuperDial could acquire its team and document AI stack to round out a voice-plus-document full-stack PA solution.

Neon Health — The AI-Native Hub Automation Layer

Founded in 2023 by Stedman Blake Hood (ex-YC, multiple exits) and Harry Bleyan (MIT, ex-Tesla, ex-Google), Neon Health (legal entity: Belay, Inc.) targets the specialty pharmaceutical hub-services market. Rather than competing with hubs, Neon sells an AI automation layer that sits on top of existing hub infrastructure. Seed round of $6M in September 2025 led by NFX, with DigiTx Partners (Astellas/MPM Capital JV) participating.

Differentiators:

• Full end-to-end coverage: BV, PA, copay enrollment, provider referrals, adverse event intake, adherence support — broadest automation scope of any vendor in this set
• Hub-embedded GTM: Neon’s CareMetx relationship (marquee case study: 300% ROI, 70% cost reduction, 2x throughput) demonstrates land-and-expand potential selling into hubs rather than competing for hub contracts
• Published performance: 1.92M+ human hours saved, 100B+ tokens processed, up to 99% of calls zero-human-intervention
• Strategic investor alignment: DigiTx’s Astellas backing provides pharma-strategic distribution

Weaknesses:

• Severe customer concentration: CareMetx is the primary proof point; loss of that relationship would be existential
• Capital gap: $6M vs. Infinitus’s $102.9M — limited runway for enterprise sales cycles
• Competitive replication risk: EVERSANA launched an AI Accelerator in 2025; CareMetx is building internal AI alongside Neon; Infinitus has a multi-year scale advantage
• Limited pharma brand count: Six pharma brand clients as of September 2025

Strategic significance: Neon represents the “AI-native layer on legacy hubs” thesis — if it works, it validates the idea that AI doesn’t need to displace hubs, just automate 80% of their labor layer. The next 12-18 months determine whether Neon is a meaningful standalone business or an acquisition target for a larger hub (EVERSANA, ConnectiveRx, or CareMetx deepening the existing relationship into ownership).

SuperDial — The Provider-Side RCM Voice AI

Founded in 2021 as Haven Health/SuperBill by Sam Schwager (Stanford CS/AI, ex-McKinsey) and Harrison Caruthers (Stanford Symbolic Systems, ex-Amazon), SuperDial relaunched late 2023. Targets RCM companies, dental service organizations, and large provider groups — not pharma manufacturers. Series A of $15M in June 2025 led by SignalFire (first investment from their new $1B applied AI fund). January 2025 MajorBoost acquisition deepened IVR navigation. March 2026 Omega Healthcare partnership for RCM outsourcer distribution.

Differentiators:

• Customizable scripting: Fully configurable call flows — broader than more rigidly templated competitors
• Human-in-the-loop fallback: Human call center steps in for unresolved calls, generating labeled training data
• Bidirectional EHR integration: Epic, Cerner, eClinicalWorks, athenahealth, AdvancedMD via HL7/FHIR APIs
• RCM channel strategy: Simpler procurement, faster time-to-value vs. pharma hub sales cycle

Weaknesses:

• Capital gap vs. Infinitus: ~$20M total (including $3M debt) vs. Infinitus’s $102.9M
• Scale gap: Millions of calls (trending 10M/yr by end 2025) vs. Infinitus’s 100M+ minutes
• Data advantage fragmentation: Customer-specific workflows mean SuperDial’s training data is less concentrated than Infinitus’s standardized therapy-specific flows
• FHIR disintermediation risk: Same structural threat as Infinitus, but with less capital to pivot

Strategic significance: The Omega Healthcare partnership is the most interesting 2026 development — B2B2B distribution through a Best in KLAS RCM outsourcer. If it works, SuperDial could capture meaningful RCM market share without the high CAC of direct provider sales.

CMS-0057-F Readiness

The CMS-0057-F Interoperability and Prior Authorization Final Rule requires Medicare Advantage, Medicaid, and CHIP plans to implement FHIR-based PA APIs by January 2027, with 7-day standard and 72-hour urgent response timelines. The rule excludes pharmacy benefit PA and commercial plans but creates a massive forced-adoption wave for payer-side platforms.

Vendor	CMS-0057-F Fit	Rationale
Cohere Health	Dominant	Tier 1 FHIR API + AI auto-determination on top of APIs is the most defensible payer-side stack
Infinitus Systems	Complementary	Voice AI fills the non-FHIR commercial/ERISA/Part D gap; likely complements FHIR rather than replacing it outright
SuperDial	Complementary	Same non-FHIR coverage thesis as Infinitus, provider-side
Tandem AI	Indirect	Provider-side platform benefits if FHIR reduces PA friction; no direct FHIR product disclosed
Neon Health	Indirect	AI-native orchestration can consume FHIR or voice — architecture-agnostic
Coral AI	Indirect	Document AI works for non-FHIR payers and the legacy fax/portal layer; FHIR doesn’t eliminate document prep

The key CMS-0057-F insight: FHIR APIs do not eliminate the PA decision — they just move the data. Vendors with only API plumbing (Availity, Edifecs) leave payers with the hardest problem (actually making the clinical determination). Cohere’s API-plus-AI stack is the most defensible answer. Voice AI vendors (Infinitus, SuperDial) remain relevant for the ~160M commercially-insured and ~55M Part D lives outside the mandate.

Market Positioning Summary

Dimension	Infinitus	Tandem	Cohere	Coral	Neon	SuperDial
Customer side	Pharma / SP	Provider	Payer	Provider	Pharma / Hub	Provider / RCM
Architecture	Voice AI	Agentic multi-modal	Decision AI	Document AI	End-to-end AI	Voice AI
Primary differentiator	Scale + knowledge graph	Free-for-provider + full scope	Payer-side AI auto-approval	Within-existing-systems overlay	Hub-embedded land-and-expand	RCM channel + HITL
Total funding	$102.9M	$137M	$200M+	$2M	$6M	$20M+
Valuation	~$600M	$1B	Private	N/A	N/A	~$20M
Revenue maturity	6M+ calls, 44% F50	Not disclosed	12M+ PA / year	Early revenue traction	Early customer proof points	Early revenue traction
Enterprise readiness	High	Emerging	High	Low	Emerging	Moderate
CMS-0057 fit	Complementary	Indirect	Dominant	Indirect	Indirect	Complementary

Buyer Decision Matrix

The shortlisting logic depends on which side of the PA transaction the buyer controls. A manufacturer, hub vendor, provider group, specialty pharmacy, and payer are not buying the same product even when every vendor uses the phrase “AI prior authorization.”

Buyer	Primary problem	Best-fit architecture
Pharma manufacturer	Reduce hub labor, time-to-therapy, PA abandonment, and payer-call friction across specialty brands	Voice AI plus hub workflow automation; Infinitus and Neon are the most directly relevant in the current source set
Hub / patient services vendor	Lower FTE burden without losing program-specific SOP control	Embedded automation layer that can sit on top of Salesforce / hub CRM and preserve manufacturer reporting
Provider / RCM group	Intake, benefits, documentation, portal submission, status follow-up, and denial reduction	Document AI, voice AI, and RCM-distribution models such as Coral and SuperDial
Specialty pharmacy	Referral intake, payer verification, PA status, and patient routing at scale	Voice / document AI plus pharmacy-management-system integrations
Payer / health plan	CMS-0057 compliance, auto-determination, UM modernization, and provider satisfaction	Payer-side decisioning and API workflow such as Cohere

Two diligence questions should be mandatory. First, does the vendor automate a workflow that still exists after FHIR APIs mature, or is it dependent on today’s phone / fax / portal friction? Second, who owns the resulting status and decision data: the manufacturer, provider, payer, hub, or platform vendor? The first question determines durability; the second determines whether the automation improves pharma commercial intelligence or merely reduces vendor labor cost.

Analyst Observations

1. Valuation divergence signals capital-market exuberance, not vendor maturity. Tandem AI at $1B with $137M raised and no disclosed revenue is the clearest example. Infinitus’s $600M on $102.9M raised with Fortune 50 customer base is the more grounded comparator. Two-to-three-year hindsight will reveal whether Tandem’s broader scope + free-for-provider model justifies the premium — or whether it was a top-of-cycle valuation moment.

2. The “voice AI vs. document AI vs. decision AI” trichotomy is real but converging. Infinitus (voice) is extending into patient-facing and provider-facing agents. Cohere (decision) is acquiring post-service AI (ZignaAI). Tandem (full-stack) is pursuing a horizontal “medication access” platform. The practical buyer implication is to diligence roadmap breadth and integration depth, not just today’s point-solution claim.

3. Pharma manufacturer relevance is highest for Infinitus, Neon, and Tandem. Infinitus has the deepest specialty drug PA workflow scars (hub integrations, specialty pharmacy relationships, 44% F50 penetration). Neon’s hub-embedded model is narrower but directly automates the labor-intensive hub workflow that pharma pays for. Tandem’s relevance to pharma depends on whether it develops specialty drug PA capabilities — currently positioned toward primary care / PCP workflow (GLP-1, chronic disease).

4. CMS-0057-F is the structural demand catalyst but not the only one. Commercial PA reform, state-level PA legislation (270+ utilization management bills in 2025 state legislatures alone), and manufacturer cost pressure on hubs all drive demand. Payer-side vendors benefit most from 2027 mandate timing; provider-side vendors benefit most from near-term cost pressure.

5. Consolidation risk belongs in the RFP. With substantial capital deployed across these vendors, plus legacy incumbents (CoverMyMeds, Surescripts, eviCore/Evernorth) mobilizing AI responses, buyers should contract for data portability, change-of-control notice, and continuity if the vendor is acquired or pivots.

Implications

Manufacturers should buy AI PA capability by workflow gap, not by vendor funding rank. A GLP-1 or autoimmune brand with PCP prescribing may need CoverMyMeds/Surescripts rails plus a document or voice layer for exceptions. A rare disease or oncology brand may need hub-embedded automation that escalates cleanly to clinical case managers. A payer-facing product strategy may require monitoring Cohere-like health-plan adoption because that determines what documentation and API paths the hub must support.

Contracting should force measurable integration. Require vendors to disclose which PA stages are automated, which payers and plan types are covered, what happens when CMS-0057 APIs do not apply, how human escalation is priced, and whether outputs can be audited by the hub or manufacturer. The buyer risk is not just failed automation; it is creating another workflow silo between ePA, payer calls, document intake, hub case management, and pharmacy routing.

Related Analyses

• AI Disruption in Pharma Services — Broader AI healthcare automation context
• CMS-0057 Vendor Readiness Matrix — Regulatory deep dive for all PA vendors
• AI Prior Authorization Vendors Comparison — Four-vendor head-to-head (Infinitus, SuperDial, Neon, Coral)
• 100ms — Voice-first intake, benefits, and PA-adjacent workflow automation
• RISA Labs — Oncology-specific AI access operating system
• AI Hub Operations ROI — Quantitative ROI framework for AI automation
• Hub Services Market Analysis — Hub market sizing and vendor landscape
• Payer Landscape Overview — PBM and health plan market structure

Related Wiki Vendors (Auto)

Auto-generated cross-references closing audit-surfaced link gaps. Vendors named in the prose above without inline links are listed here so the wiki graph is queryable.

• Availity
• ConnectiveRx
• Edifecs
• GoodRx
• Omega Healthcare
• Salesforce Life Sciences Cloud
• Tennr