Which AI-native services can take over healthcare prior authorization workflows from a human operations team?

If by “take over” you mean replace a human PA operations team for most of the workflow—intake → determine whether auth is required → gather clinical documentation → submit → chase payer → respond to requests → track status → escalate exceptions—the market has a handful of credible AI-native contenders.

The important distinction is which side of the transaction you sit on: payer/health plan vs. provider/health system vs. pharma hub.

The strongest candidates

1. Cohere Health — closest to a true payer-side replacement

Cohere is probably the most interesting company if you're talking about a health plan replacing an internal utilization-management/PA operation rather than simply buying a better PA submission tool.

Its platform is explicitly designed for touchless prior authorization, with AI applying health-plan policies and clinical criteria. Cohere currently claims 85% of authorizations can be approved in real time, with the remaining cases going to clinicians. It also offers a delegated model in which Cohere provides the operational/clinical staff rather than merely software.

That last point is important: software + delegated clinical operations is much closer to “replace my PA department” than most AI vendors.

2. Waystar — strongest provider-side incumbent

Waystar is attacking PA from the provider/revenue-cycle side. Its Authorization Manager uses automation to determine authorization requirements, submit requests and attach clinical documentation; industry reporting describes it as processing tens of millions of authorization transactions annually.

This is less “AI replaces nurses making clinical determinations” and more:

AI replaces the staff whose job is to figure out whether auth is needed, assemble the packet, submit it, monitor it and work the queue. That's potentially a huge portion of a provider organization's PA headcount.

3. Rhyme — infrastructure for making PA nearly touchless

Rhyme is particularly interesting because it sits at the EHR ↔ payer connectivity layer. Its partnership with Cohere was explicitly aimed at automating the end-to-end PA process: Rhyme handles EMR integration while Cohere handles the clinical authorization intelligence.

So I wouldn't necessarily view Rhyme as a standalone “AI ops replacement.” I'd view it as an important piece of an agentic PA stack.

4. Infinitus — potentially enormous for phone-based PA

Infinitus is interesting for a different reason: voice agents.

A surprising amount of PA work still consists of humans calling payers, navigating IVRs, checking status, obtaining information and documenting conversations. An AI voice agent can perform those interactions continuously and feed the results back into the PA workflow.

For organizations whose PA team spends a lot of its time on the phone rather than making clinical decisions, this can have unusually high labor-replacement potential.

5. Anterior — clinical-review automation

Anterior is more focused on the clinical reasoning/document-review portion of authorization. This is the piece that traditional RPA can't easily automate: interpreting records, extracting relevant evidence and determining whether the documentation supports medical necessity.

It's therefore more complementary to the administrative automation vendors than a replacement for the entire workflow.

6. Elsai — the “AI agent” approach

Elsai is one of the newer companies explicitly positioning itself around agentic prior authorization, rather than simply digitizing forms.

That distinction matters. The emerging architecture is essentially:

AI agent receives request → reads chart → determines requirements → retrieves evidence → fills payer workflow → submits → monitors → responds to requests → escalates only exceptions.

A 2026 industry comparison identifies Elsai alongside Cohere, Waystar, Rhyme, Navina and Infinitus as notable PA-AI players.

---

The really interesting distinction

I would divide the market into three categories:

A. “Replace my PA operations team”

Look hardest at:

• Cohere — payer-side
• Waystar — provider-side
• Infinitus — phone/call-center-heavy workflows
• Rhyme — EHR/payer workflow automation
• Elsai — emerging agentic approach

B. “Make my existing PA nurses 3–10× more productive”

Look at:

• Anterior
• Cohere Review Assist
• MCG
• Navina/Abridge-type clinical intelligence
• AI document/evidence extraction layers

Cohere itself describes Review Assist as surfacing the relevant clinical evidence for nurses inside their existing UM environment rather than removing the human reviewer altogether.

C. “Replace a pharma patient-access/hub operation”

This is a somewhat different market. Infinitus and Neon Health are particularly relevant; industry analysis currently places them in the direct hub-automation category, while Cohere is principally a payer-side platform rather than a pharma-hub replacement.

---

If I were evaluating this as a buyer

I'd actually score vendors against six autonomous steps, rather than asking “does it have AI?”:

1. Auth determination — Does it know whether PA is required?
2. Evidence acquisition — Can it autonomously find the required clinical documentation?
3. Submission — Can it submit electronically and handle payer portals/fax/phone?
4. Payer interaction — Can it answer requests for additional information without a human?
5. Status management — Can it continuously chase and reconcile cases?
6. Clinical exceptions — Does it route only genuinely ambiguous cases to a nurse/clinician?

The vendor that can do all six is capable of replacing an operations team. A vendor that only automates #3 is basically a better PA clearinghouse.

My short list for an actual 2026 enterprise evaluation would be Cohere, Waystar, Rhyme, Infinitus, and Elsai, with the shortlist changing substantially depending on whether you're a payer, provider, or pharma company. The market is also increasingly converging on AI agents + payer connectivity + clinical intelligence, rather than one monolithic PA application.

If you tell me whether you're looking at this from a payer, health-system, specialty-pharmacy, or pharma-hub perspective, I can map the vendors much more precisely—including which ones can actually eliminate FTEs vs. merely augment them, approximate pricing/business models, and who has real production deployments.

This is less “AI replaces nurses making clinical determinations” and more:

AI replaces the staff whose job is to figure out whether auth is needed, assemble the packet, submit it, monitor it and work the queue. That's potentially a huge portion of a provider organization's PA headcount.

3. Rhyme — infrastructure for making PA nearly touchless

Rhyme is particularly interesting because it sits at the EHR ↔ payer connectivity layer. Its partnership with Cohere was explicitly aimed at automating the end-to-end PA process: Rhyme handles EMR integration while Cohere handles the clinical authorization intelligence.

So I wouldn't necessarily view Rhyme as a standalone “AI ops replacement.” I'd view it as an important piece of an agentic PA stack.

4. Infinitus — potentially enormous for phone-based PA

Infinitus is interesting for a different reason: voice agents.

A surprising amount of PA work still consists of humans calling payers, navigating IVRs, checking status, obtaining information and documenting conversations. An AI voice agent can perform those interactions continuously and feed the results back into the PA workflow.

For organizations whose PA team spends a lot of its time on the phone rather than making clinical decisions, this can have unusually high labor-replacement potential.

5. Anterior — clinical-review automation

Anterior is more focused on the clinical reasoning/document-review portion of authorization. This is the piece that traditional RPA can't easily automate: interpreting records, extracting relevant evidence and determining whether the documentation supports medical necessity.

It's therefore more complementary to the administrative automation vendors than a replacement for the entire workflow.

6. Elsai — the “AI agent” approach

Elsai is one of the newer companies explicitly positioning itself around agentic prior authorization, rather than simply digitizing forms.

That distinction matters. The emerging architecture is essentially:

AI agent receives request → reads chart → determines requirements → retrieves evidence → fills payer workflow → submits → monitors → responds to requests → escalates only exceptions.

A 2026 industry comparison identifies Elsai alongside Cohere, Waystar, Rhyme, Navina and Infinitus as notable PA-AI players.

---

The really interesting distinction

I would divide the market into three categories:

A. “Replace my PA operations team”

Look hardest at:

The healthcare landscape has seen an emergence of AI-native and AI-powered platforms designed to reduce or completely take over the heavy lifting of prior authorization (PA) workflows from human operations teams. Rather than just acting as basic optical character recognition (OCR) or manual form-fillers, these modern services leverage Large Language Models (LLMs), intelligent document processing, and FHIR-based APIs to interpret complex clinical criteria, match them against payer medical policies, pre-fill documentation, and automate submissions.

Major AI-Driven Services & Platforms

• Cohere Health : Specializes in AI-powered clinical intelligence and collaboration for both health plans and providers. It automates prior authorization and utilization management end-to-end by digitizing clinical criteria matching and reducing friction between the request and approval.
• Availity AuthAI : Integrates clinical intelligence grounded in codified medical policies with nationwide multi-payer connectivity. It analyzes patient charts to automatically structure data and recommend or process routine prior authorizations in near real time.
• Thoughtful.ai : Provides AI-driven "digital workers" and intelligent automation agents specifically tailored for healthcare revenue cycle management. This includes automating complex, repetitive back-office workflows like prior authorizations, eligibility checks, and claims status.
• Innovaccer Galaxy UM : An AI-native utilization management platform launched to help health plans and large provider ecosystems completely automate prior authorization processing and drastically cut administrative overhead.
• Waystar : Incorporates advanced AI and machine learning into its revenue cycle suite to proactively identify PA requirements, predict and prevent denials, and streamline the submission lifecycle.

How These Services Operate Under the Hood

Modern AI-native systems handle tasks that traditionally required an army of human intake coordinators or clinical nurses:

Note on Human Oversight: While these platforms drastically reduce the need for manual operations teams, fully autonomous "auto-approval" is usually restricted to routine or highly standardized cases. Complex or borderline clinical denials still often route to a streamlined human-in-the-loop queue where AI acts as a clinical review assistant.

Would you like to explore solutions tailored specifically for providers/health systems vs. payers/insurance plans , or focus on a particular EHR integration (like Epic or Cerner)?

Yes. A growing group of AI-native companies is specifically trying to replace much of the work currently performed by prior authorization specialists, rather than simply providing workflow software or electronic prior authorization (ePA) routing.

The market naturally breaks into two categories:

Provider-side AI companies

These are the closest to replacing a human operations team.

Payer-side AI platforms

These focus on utilization management rather than provider operations.

• Cohere Health
• Anterior
• Case Health AI
• Genzeon Prior Authorization Agents

Typical capabilities include:

• Intake from fax, API, portal, and X12
• Clinical summarization
• Guideline matching
• Medical necessity review
• Provider outreach
• Routing to physician reviewers only when necessary
• Full audit trails for regulatory compliance

What "taking over" actually means

The newest AI-native systems generally automate much more than traditional RPA. A mature workflow can include:

1. Reading the order from the EHR.
2. Determining whether prior authorization is required.
3. Collecting clinical documentation.
4. Completing payer-specific forms.
5. Logging into payer portals.
6. Uploading documentation.
7. Calling payers when required.
8. Monitoring authorization status over several days.
9. Responding to requests for additional information.
10. Updating the EHR.
11. Escalating only unusual or ambiguous cases to humans.

This differs from older ePA vendors, which primarily route transactions and still rely on staff to manage exceptions.

Which vendors appear most AI-native?

If your definition is "could replace an operations team member rather than assist one," the strongest examples today are:

• Honey Health
• Mandolin
• Ascertain
• GenHealth
• Beija
• Silna Health

These position themselves as autonomous or agentic systems that execute work across existing EHRs and payer portals, rather than requiring users to operate another workflow application.

Enterprise readiness considerations

When evaluating these vendors, organizations typically look beyond automation claims and assess:

• Breadth of payer coverage (commercial, Medicare Advantage, Medicaid)
• Support for medical, pharmacy, imaging, and specialty-drug authorizations
• EHR compatibility (e.g., Epic, Oracle Health/Cerner, athenahealth)
• Exception-handling rates (how often humans still intervene)
• Auditability and explainability
• HIPAA and SOC 2 compliance
• Ability to operate through portals, APIs, EDI, fax, and phone—not just one submission channel

In practice, the leaders differ by setting: Mandolin and SamaCare are especially strong for specialty drugs and infusion workflows, while platforms such as Honey Health, Ascertain, and GenHealth target broader provider-side prior authorization across multiple service types.

Several AI-native and AI-powered platforms have emerged to shoulder or fully automate prior authorization (PA) workflows, reducing the need for heavy manual operations teams . These services address different sides of the ecosystem—some target provider-side submission and chart extraction, while others handle payer-side clinical reviews and decision support.

Prominent AI-driven platforms taking over these workflows include:

• 1. Myndshft — A prominent automated platform built on machine learning and generative AI that handles both medical and pharmacy benefit prior authorizations. It connects directly to EHRs or pharmacy systems to automatically check whether an authorization is required, extract required data, and submit the paperwork dynamically through the optimal electronic channels.
• 1. Cohere Health — Widely recognized as an AI-native clinical intelligence platform, Cohere heavily focuses on streamlining and automating complex medical prior authorizations, primarily operating at the intersection of health plans and risk-bearing provider groups. It embeds evidence-based clinical criteria directly into the digital workflow, reducing peer-to-peer administrative friction and automating routine approvals.
• 1. Availity AuthAI — Availity leverages an analytical AI approach grounded in codified medical policy. Using FHIR APIs (such as CRD, DTR, and PAS), AuthAI automatically queries whether an authorization is needed, pulls the exact clinical documentation required from the provider's EHR, and processes transparent, auditable case recommendations for routine approvals in near real time.
• 1. Waystar — Known for embedding robust AI automation directly into revenue cycle management (RCM) infrastructure, Waystar’s prior authorization solutions automate status tracking, minimize documentation errors, and streamline submission workflows to mitigate delays and downstream denials.

Are you looking to implement these tools on the provider side (to streamline clinic submissions and reduce rejections) or on the payer side (to automate utilization management and clinical reviews)? If you share your specific workflow bottlenecks, I can help narrow down the ideal platform fit.

Yes. If by “take over” you mean replace most of a human PA operations team rather than merely give staff a better work queue, the market is increasingly credible—but the vendors attack different parts of the workflow.

The strongest provider-side candidates

The important distinction is that Waystar + Infinitus is closer to an “AI PA operations team” than any single product. Waystar handles the electronic/transactional layer, while Infinitus handles the cases where the payer still requires a phone conversation. Waystar says its Authorization Manager covers 30+ service lines and 70M annual authorization transactions, with exception-based worklists and automated status retrieval. Waystar Waystar Infinitus specifically automates payer calls and PA follow-up using voice agents.

Where Cohere is different

coherehealth.com is particularly interesting if you're asking about replacing utilization-management operations, rather than merely replacing the provider's authorization coordinators.

Its platform performs clinical-information extraction and medical-necessity decisioning, and Cohere currently advertises 85% of prior authorizations approved in real time. It explicitly says its AI handles administrative work while humans retain control of critical clinical decisions.

So I'd characterize the landscape this way:

• Provider PA department: Waystar / Rhyme / Infinx
• Payer UM department: Cohere
• Human payer phone-call work: Infinitus
• Medication/eRx PA: Surescripts
• True end-to-end autonomous PA: still emerging; usually requires combining several of the above

What I would evaluate if your goal is headcount elimination

The key question isn't “Does it have AI?” It's:

What percentage of PA cases can go from order → auth determination → documentation assembly → submission → payer interaction → status → approval without a human touching it? I'd score vendors on six automation stages:

1. Auth-required determination
2. Clinical/document extraction
3. Payer-specific form completion
4. Electronic submission
5. Payer follow-up / phone calls
6. Exception handling and escalation

Most vendors are strong on 1–4. #5 is where Infinitus is unusually differentiated, while #6 is still where human operations teams tend to survive.

Waystar is explicitly moving toward agentic AI that executes revenue-cycle tasks with minimal intervention, including PA support.

Bottom line: If you're building a shortlist for a provider organization trying to replace a 10–100 person PA operations team, I'd start with Waystar, Infinitus, Rhyme, and Infinx. If you're a health plan, I'd put Cohere at the top of the list instead. The most interesting architecture may be Waystar (electronic rails) + Infinitus (voice/exception rails) + an agentic clinical/document layer, rather than betting on one monolithic vendor.

If you tell me provider vs. payer, specialty, approximate PA volume, and whether you want to eliminate FTEs or just reduce workload, I can narrow this to the 3–5 vendors most likely to actually replace the operation.

Several AI-native and agentic platforms are designed to take over repetitive prior authorization (PA) workflows from human operations teams . Rather than acting merely as electronic submission forms or basic fax routers, these services use large language models (LLMs) and intelligent agents to extract unstructured clinical evidence from EHRs, match it against complex payer policies, submit via APIs or browser/voice automation, and track statuses end-to-end.

The leading AI-native and deeply automated platforms addressing this space include:

• Honey Health operates specialty drug and surgical prior authorizations as an autonomous AI "staff member" that handles the complete request lifecycle across medical and pharmacy benefits rather than just routing a form.
• Humata Health (acquired by R1 RCM) focuses on touchless prior authorization for procedures and specialty drugs, using AI policy matching and automated status tracking to drive down manual intervention.
• Cohere Health provides clinically-trained AI for utilization management and surgical/musculoskeletal authorizations, combing through unstructured clinical records to satisfy payer policies and achieve high rates of real-time approvals.
• Flexbone AI deploys multi-channel AI agents that pull orders from EHRs, navigate disparate payer web portals via browser agents, call payers via voice agents for edge cases or peer-to-peers, and write determinations back to the chart.
• Prosper AI utilizes healthcare-tuned conversational voice agents to automate the phone-based and administrative legwork of calling payers, verifying benefits, and checking PA statuses.
• Availity AuthAI applies analytical AI grounded in codified medical policy to recommend approvals in near real time, emphasizing transparent, auditable logic designed for major payer-provider networks.

Note on Industry Operations: While these tools significantly deflect human workloads (often achieving 80% to 90%+ touchless or real-time processing for routine cases), complete end-to-end replacement of human teams remains nuanced. Regulatory requirements—such as state-level restrictions on pure-AI medical necessity denials and complex peer-to-peer appeals—still mean that edge cases and subjective denials are typically routed back to human clinicians or operations specialists.

If you'd like, let me know:

• Are you looking for a solution from the provider side (clinics/health systems) or payer side (insurers)?
• What is your primary specialty or service line (e.g., surgical procedures, oncology/infusion, general ambulatory)?
• Which EHR system (e.g., Epic, athenahealth, eClinicalWorks) do you need to integrate with?

I can help narrow down which platform fits your specific operational volume and tech stack.

Several AI-native platforms and intelligent automation services can handle or significantly reduce human operations workloads in healthcare prior authorization (PA). These tools use optical character recognition (OCR), large language models (LLMs), and FHIR/API integrations to extract clinical facts, match them to specific payer rules, pre-populate forms, and track requests.

While fully autonomous, "straight-through" processing is increasingly common for routine policy-aligned requests, most services maintain a human-in-the-loop governance model for complex clinical edge cases.

Leading AI-Driven Prior Authorization Services

• Cohere Health : An established AI-native platform focused on clinical intelligence and automated approvals, using LLMs to scan unstructured clinical attachments against medical policies, enabling high rates of real-time determinations.
• Prosper AI : Specializes in high-friction operational workflows, deploying healthcare-tuned voice and data agents that handle phone trees, interact with payer IVRs, initiate requests, and track statuses across both medical and pharmacy benefits.
• Availity AuthAI : Availity’s native AI engine connects providers across a massive network of health plans to parse clinical documentation and deliver plan-specific authorization recommendations rapidly.
• Humata Health : Focuses on eliminating administrative friction by embedding intelligent document processing and automated packet assembly directly within existing electronic health record (EHR) workflows.
• Myndshft : Provides a unified real-time automation platform spanning both medical and pharmacy benefit prior authorizations, reducing manual data entry and status polling.
• Optum Digital Auth Complete : A comprehensive enterprise option that connects to hundreds of payer systems to surface medical policies, bundle clinical criteria, and track the full authorization lifecycle.

If you'd like to narrow down these options, please share:

• Are you looking for a solution for a provider health system or a payer organization?
• Which EHR or billing software (such as Epic, athenahealth, or Cerner) do your operations teams use?

If the goal is to replace a human prior authorization operations team, the relevant category is not just “AI documentation tools.” You need AI agents that can execute the PA lifecycle: identify whether authorization is required, gather clinical evidence, complete payer forms, submit through portals/APIs/fax, monitor status, respond to requests, escalate exceptions, and document outcomes.

The market is splitting into a few architectures:

Provider-side AI agents (closest to replacing PA coordinators)

Honey Health

Best fit: ambulatory groups, MSOs, specialty practices that want an “AI employee” handling PA work.

Capabilities:

• Detect PA requirements from EHR workflows
• Extract clinical information
• Submit requests to payer portals
• Track pending authorizations
• Follow up and update charts

It is positioned as an end-to-end provider-side PA agent rather than a simple rules engine.

Cohere Health

Best fit: payers and utilization management organizations.

Capabilities:

• AI-supported clinical review
• Policy matching
• Automated determinations for eligible cases
• Human review routing for complex cases

Cohere focuses more on the payer decision side than replacing provider PA coordinators.

Waystar

Best fit: health systems wanting PA automation inside broader revenue-cycle operations.

Capabilities:

• Authorization workflows
• RCM automation
• Integration with billing and administrative workflows

(Compared with newer agentic vendors, Waystar is more of an established RCM platform expanding automation.)

---

Specialty medication / infusion PA automation

These companies focus on the hardest PA categories: biologics, oncology drugs, infusion therapies, and specialty pharmacy.

Infinitus Systems

Best fit: payer calls and access workflows.

Capabilities:

• Voice AI calling insurers
• Benefits verification
• Authorization status checks
• Access workflow automation

Neon Health

Best fit: specialty-drug hub workflows.

Capabilities:

• Benefits investigation
• Prior authorization workflows
• Patient access coordination

---

Health-plan utilization management automation

For a payer trying to shrink a nurse reviewer or UM operations team:

Cohere Health

• AI clinical intelligence
• Policy-based decision support
• Automated approvals for appropriate cases
• Human escalation for exceptions

UiPath

Best fit: large payers that want orchestration across legacy systems.

Capabilities:

• Intake automation
• Document extraction
• Clinical summarization
• Workflow orchestration
• Human-in-the-loop review

UiPath’s approach is usually to automate the workflow around clinicians rather than eliminate clinical judgment.

---

If you want the closest thing to “replace the PA team”

A realistic target operating model in 2026 is:

AI handles

• 80–95% of routine submissions
• payer portal navigation
• missing-information checks
• status calls/messages
• documentation assembly
• routine approvals

Humans handle

• ambiguous medical necessity cases
• peer-to-peer reviews
• appeals strategy
• payer escalations
• unusual clinical scenarios

Vendor shortlist by buyer

Large health system / MSO

1. Honey Health
2. Waystar
3. UiPath
4. Cohere (if payer relationships matter)

Specialty clinic (oncology, rheumatology, dermatology, infusion)

1. Honey Health
2. Neon Health
3. Infinitus
4. Specialty-focused access vendors

Health plan

1. Cohere Health
2. UiPath
3. Other utilization-management AI platforms

The key diligence question is: does the vendor merely draft PA paperwork, or does it actually operate the workflow (submit, monitor, chase, and close the loop)? The latter category is what can materially reduce a human operations team.

Yes. The market has moved beyond “AI that helps a coordinator fill out a PA form” toward agentic systems that can own substantial portions of the workflow. But the vendors differ dramatically in how much of the human ops team they can actually replace.

The strongest candidates

1. Silna is probably the closest to what you're describing

Silna explicitly positions its product as managing the entire prior-auth process: monitoring expirations, identifying payer-specific documentation requirements, generating/submitting forms, following up, communicating approvals, and surfacing escalations.

That makes it interesting if the objective is:

“Give us the PA queue and let the software run it, rather than giving our coordinators another tool.”

Its remaining human work appears to be primarily exceptions—additional-information requests, unusual cases, resubmissions, and denials.

2. Infinitus attacks the ugly part that software often misses

Infinitus is particularly interesting because it doesn't stop at electronic PA.

Its agents actually call payers, navigate the conversation, ask for requirements/status, and return the information to the provider's system. Infinitus says its PA agents cover specialty drugs, procedures and diagnostics and have expertise across 1,000+ therapies.

That's important because a PA “workflow” often looks like:

EHR → payer portal → fax → payer phone call → hold → transfer → missing document → another call → status check → resubmission

Infinitus can automate that telephone layer rather than forcing humans to remain the integration glue.

3. Waystar is the enterprise incumbent to take seriously

Waystar has a much broader installed-base strategy. Its Authorization Manager uses AI/automation to determine whether authorization is required, initiate authorizations, analyze EHR/HIS/PM orders, attach clinical documentation and automate certain submissions and notifications.

I'd categorize it as “AI-enabled PA infrastructure” rather than a pure AI-native workforce replacement.

That's potentially a feature rather than a bug for a health system: it is more likely to fit into existing RCM infrastructure.

4. Cohere is different—and extremely important

Cohere Health is much more oriented toward the payer/UM side.

Its AI analyzes structured and unstructured clinical records against payer policies and says its automation handles the majority of PA decisions, with 85% of authorizations approved in real time depending on specialty.

So if by “take over PA” you mean replace the payer's utilization-management operation, Cohere is arguably one of the strongest companies in the category.

If you mean replace the provider's 10-person PA department, it's not the same product.

5. SuperDial is a strong “AI employee” for calls

SuperDial offers agents for insurance verification, PA requirements/status, claims follow-up, credentialing and other payer/provider calls.

I'd view it as particularly compelling when your PA operation is already fairly good electronically but burns enormous amounts of labor on calling insurance companies.

---

The important distinction: “AI PA software” vs. “AI PA workforce”

This is the distinction I'd use when evaluating the market.

Tier 1 — Copilot

“Here's the PA form and the clinical evidence you should attach.”

Doesn't replace many people.

Tier 2 — Workflow automation

“I'll determine whether PA is required, assemble the packet, submit it and track the status.”

Can substantially reduce headcount.

Tier 3 — Agentic PA operations

“Give me the order. I'll figure out the payer requirements, obtain missing information, assemble the clinical packet, submit it through whatever channel is necessary, call the payer when required, monitor the authorization, respond to routine requests, update the EHR and escalate only exceptions.”

This is the category that can actually replace an operations team.

Silna + Infinitus are particularly interesting when viewed through that lens: one is closer to owning the workflow, while the other is exceptionally strong at the payer-communication layer.

But I wouldn't assume “fully autonomous” yet

There is a meaningful gap between vendor demos and completely unattended healthcare administration. Two 2026 benchmarks are especially revealing: HealthAdminBench found the best computer-use agent achieved only 36.3% end-to-end task success across realistic healthcare-admin tasks, while CHI-Bench found the best configuration resolved only 28% of its long-horizon healthcare workflows.

That suggests the realistic 2026 architecture is:

AI handles 70–95% of routine cases → humans handle an exception queue, rather than “fire the entire PA department.”

And that's actually the business model I'd want to buy. The vendor should be willing to contract around touchless authorization rate, exception rate, turnaround time, and accuracy, not merely “AI-assisted productivity.”

My shortlist

If I were evaluating this for a provider organization today, I'd put the first calls into:

1. Silna — strongest match for end-to-end provider PA operations
2. Infinitus — strongest for payer phone interactions and follow-up
3. Waystar — strongest enterprise/RCM incumbent
4. SuperDial — strong AI workforce for payer calls
5. Cohere — strongest if you're actually on the payer/UM side

The key diligence question for each isn't “How much does your AI automate?” It's:

“If I give you 10,000 PA cases tomorrow, exactly which steps happen without a human touching them, which steps require a human, and can you show me the touchless rate by payer and specialty?”

That's the metric that separates a genuine AI-native replacement for a PA ops team from another RCM productivity tool.

AI-native platforms are shifting prior authorization (PA) from a manual, human-heavy operational hurdle to an automated, "straight-through processing" workflow . These services use large language models (LLMs), natural language processing (NLP), and specialized AI agents to extract clinical facts from electronic health records (EHRs), map them against complex payer rules, submit the paperwork, and track statuses.

The leading AI-native and AI-enhanced services capable of substantially taking over these workflows from human operations teams include:

• Cohere Health — Focuses heavily on clinical intelligence and utilization management, deploying advanced machine learning and clinically trained AI models that autonomously source structured and unstructured data to approve a massive percentage of medical prior authorizations in real time.
• Humata Health — (Working closely with Optum integration ecosystems) embeds real-time approvals and intelligent automation directly into clinical and EHR workflows to capture and clear authorization requirements without human intervention.
• Prosper AI — Utilizes specialized healthcare-tuned voice AI agents to handle cumbersome payer phone calls, navigate interactive voice response (IVR) systems, initiate PAs, and perform real-time status checks.
• EasyPA — Connects to major EHRs via FHIR APIs, auto-verifies eligibility across thousands of payers, and compiles AI-prepared PA packets to process authorizations in minutes rather than days.
• Infinitus AI — Deploys voice and data AI agents specifically tailored to take over tedious follow-up tasks, checking PA statuses, and updating case files across complex specialty medication and medical procedure categories.
• Myndshft (part of DrFirst) — Unifies medical and pharmacy benefit prior authorizations into a single automated platform, evaluating criteria and handling electronic submissions dynamically.
• Waystar — Integrates cloud-native "agentic intelligence" and automated workflow solutions (like Auth Accelerate) to manage the entire prior authorization lifecycle from initial detection to final clearance.
• Banjo Health — Built with a flexible "CARE" engine for both providers and payers/PBMs, using no-code dynamic decision trees and AI-assisted clinical reviews to drastically cut down manual touchpoints.

If you are evaluating these for a specific setting, would you like to share whether you are looking from the provider side (hospitals/clinics) or the payer side (insurance/PBM) , and what EHR system you need to integrate with?