Amazon Health Connect Sends Agents to the Back Office

If the competition for the back office was already hot, it’s a certified wildfire after last week’s debut of Amazon Health Connect. 

Amazon is pitching Amazon Connect Health as a purpose-built agentic AI solution for the administrative work that gets in the way of care. That’s definitely not fun to read for all the companies that had the same tagline on their booth at ViVE.

It comes with five core capabilities straight out of the box: 

  • Patient verification
  • Appointment scheduling 
  • Pre-visit summaries
  • Ambient documentation
  • Medical coding 

What’s the primary use case? AWS Director of Healthcare AI Naji Shafi says it’s the entire patient journey.

  • When a patient calls to book an appointment, Amazon Connect Health answers immediately, confirms their identity, checks their coverage, and lines up the visit while they’re still on the line.
  • Before the visit, it reviews their complete medical history across care settings, then surfaces previsit insights like active conditions or trends relevant to closing care gaps.
  • During the visit, it drafts clinical notes for provider review in real-time, with every detail linked back to the moment in the conversation where it was discussed.
  • After the visit, it generates patient-friendly summaries and the medical codes needed for billing, allowing the visit to be payor-ready and submitted within minutes.

But wait, there’s more. Amazon Connect Health integrates natively with Epic, and connects to 100+ EHRs and 35+ HIEs through data integration partners like Redox.

  • It’s also built entirely on AWS HealthLake, the cloud giant’s FHIR data repository that’s now getting new agentic capabilities to help convert records into standard formats.

Early users love it. Amazon One Medical was the perfect sandbox for polishing Amazon Connect Health in clinical settings before opening it to outside partners. It shows in the results.

  • UC San Diego Health is saving a minute per call, diverting 630 hours a week from patient verification to direct support, and slashed call abandonment by 30%.
  • Netsmart’s EHR supports more than 1,300 community provider orgs, and it saw ambient documentation adoption skyrocket 275% – and better staff retention as a result.

The Takeaway

There were already tons of agentic AI solutions competing to automate healthcare’s administrative waste, and now there’s one that’s bankrolled by the biggest bookstore in human history. It’s a crowded space, but $1 trillion per year is also enough bloat to go around.

Anterior Closes $40M to Take AI to the Largest Plans in the Country

The AI race between payors and providers is healthcare’s Kentucky Derby, and Anterior just closed $40M to help turn the dark horses into the frontrunners.

Anterior uses AI to ease the back-office burden on health plans. It started with a laser focus on prior authorizations, translating huge amounts of unstructured data into the information that’s actually needed to make quicker decisions.

  • When Anterior helps payors deploy AI in their clinical and operational workflows, it doesn’t just dump a bunch of models on them and disappear into the sunset.
  • It embeds its own clinicians and engineers alongside the platform to support its partners, optimize accuracy, and drive a measurable impact.

Trust is a differentiator. Payors are a cautious crowd, and they aren’t exactly known for trusting new friends with their critical workflows. 

  • Anterior’s clinicians are its secret sauce. They make up about 40% of the company, and many of them have even started contributing directly to the platform’s code base.
  • This hands-on support why partners build trust, and that hard-earned resource is what allowed Anterior to take the same tech underpinning its prior auth tools and expand it to other workflows.

New partners lead to new proof points. New proof points lead to new use cases. 

  • Anterior’s early successes – from both its people and technology – have allowed it to quickly land and expand into areas like payment integrity and risk adjustment. 
  • Since closing its $20M Series A in June 2024, Anterior has deployed its AI across major payors like Geisinger Health Plan, and worked alongside enterprise technology partners like HealthEdge to build out key strategic integrations.
  • The platform now supports orgs representing over 50M covered lives, and the fresh funds will help it use those case studies to pry open the door to the biggest national plans in the business.  

The Takeaway

Anterior’s earliest partners had to gamble on an unproven platform without any real-world evidence to back it up. Now, the proof is in the success stories, and Anterior just landed another $40M to go after the largest and most risk-averse payors in the country.

PHTI Breaks Down Barriers to Clinical AI

PHTI’s new Clinical AI report delivered exactly what we’ve come to expect from their research: top tier industry analysis through the lens of actual stakeholders.

They assembled the A Team for this one. The report was built from an in-person workshop that PHTI convened with senior industry leaders – from health systems and health plans to tech firms and federal agencies – to explore what’s needed to safely scale clinical AI.

  • The workshop underscored the policy, reimbursement, and evidence gaps holding back adoption, with several key themes emerging from the discussion around their example use cases (hypertension management and mental health chatbots).

Theme 1: Evidence standards should compare AI to current standards of care and scale with risk.

  • That means comparing AI to the care that patients actually receive today rather than idealized care, then having different standards that align with the clinical risk of using the tool.
  • Highlight: Evidence should assess whether the full workflow (including multiple models, devices, and human oversight) improves outcomes, not merely model performance.

Theme 2: Performance benchmarks should be based on clinical outcomes, and safety standards should adapt as the evidence grows.

  • Ambiguity around what constitutes “good” performance is a persistent barrier. Metrics need to be anchored to specific clinical outcomes instead of vague process measures.
  • Highlight: Across both use cases, participants emphasized the need not only to set benchmarks but to set minimum safety floors, which could adjust dynamically over time on the basis of observed outcomes, changing patient risk profiles, & emerging evidence.

Theme 3: New technologies may be initially tested in lower-risk populations, but should scale quickly to high-risk populations to maximize impact.

  • Low-risk patients are tempting on-ramps, but AI’s greatest benefits come from reaching the high-need patients, and reaching them carries higher evidence expectations and more clinical risk.
  • Highlight: For mental health, engagement and retention are huge barriers to treatment. Participants cautioned that overly restrictive AI deployments risk limiting access and instead emphasized the need for appropriate care routing following LLM engagement.

The Takeaway

Even the most effective clinical AI tools still have plenty of questions to address before adoption can scale, and PHTI just crowdsourced some promising answers straight from the boots-on-the-ground in the healthcare trenches.

ViVE 2026 Recap and Major Announcements

Not even our second generational snowstorm of the year could stop ViVE Los Angeles from bringing the heat.

Agentic AI was the theme of the show, but it’s clear that category lines are only going to keep getting blurrier as agents break down the barriers to entry.

It felt like every startup had just finished supercharging their engine with the latest frontier models, and even though most healthcare orgs are open to adopting faster solutions to their problems, the toughest competition in the exhibit hall might have been buyer bandwidth.

As always, ViVE kept the spotlight on the innovation, so we’ll go ahead and follow their lead with our recap of the biggest announcements from the show.

  • athenahealth launched new agentic patient communication tools across its provider network, giving patients around-the-clock access to a virtual assistant at their doctor’s office that can help with tasks like answering basic questions or scheduling appointments.
  • Artisight integrated Epic MyChart Bedside TV with its smart hospital platform, transforming in-room televisions into an Epic-aware hub for inpatient care and patient engagement.
  • b.well Connected Health kept its hot streak going with bailey, a white-label health AI assistant. Orgs can embed bailey directly into their own apps to deliver modern patient experiences – finding care, managing medications, scheduling appointments, navigating benefits – without doing the heavy lifting of building the AI from scratch.
  • Canvas Medical showcased its new Canvas Plugin Assistant that effectively eliminates any coding skill requirements for customizing the Canvas EMR platform or building agents on top of it.
  • Care Continuity debuted CarePath IQ to give health systems better visibility into their patients’ follow-up care plans, navigation pathways, and provider handoffs – inside and outside the network.
  • CLEAR rolled out its CLEAR1 platform at Mount Sinai Health System to give patients and employees a single, secure identity across the entire ecosystem.
  • DiMe announced that CMS made the DiMe Seal a required on-ramp for digital health applications in the upcoming Medicare App Library, meaning they’ll have to complete a defined benchmark across evidence, security, and usability if they want to get in front of 86M beneficiaries.
  • Dock Health deployed its productivity platform at Mayo Clinic to streamline referral workflows and optimize operations across its cardiovascular, econsult, and specialty contract programs.
  • Fabric took the lid off Evo, a nationwide virtual care benefit that consolidates high-demand services into a unified experience: Urgent Care, Talk Therapy, Mental Health Med Mgmt, and Weight Loss.
  • Heidi was everywhere at this one. They scored a hat trick with the launch of a fully integrated Heidi Evidence tool that brings clinical evidence to workflows without any advertising baggage, the acquisition of UK-based clinical AI pioneer AutoMedica, and the debut of Heidi Comms to give care teams an AI partner for coordinating patient communications.
  • Hyro equipped its AI agents with clinically validated content and decision logic courtesy of WebMD Ignite, moving conversational AI flows beyond simple Q&As with guided actions like specialty routing and appointment scheduling.
  • Innovaccer joined forces with Allina Health Minneapolis Heart Institute to expand access to guideline-directed heart failure management from Story Health, its recently acquired virtual specialty care and patient monitoring platform that not-so-coincidentally has one of the best HF management programs in town.
  • Kontakt.io added to its flurry of new solutions with Patient Flow Agent, an orchestration agent that puts real-time operations signals from RTLS and the EHR in context so frontline caregivers can make the best decision for the patient and hospital. The RTLS component stands to unlock some big improvements for length of stay and delays, AKA revenue and experience.
  • Luma Health shared updates on its Operational AI platform that executes complex healthcare workflows from start to finish rather than optimizing isolated tasks. Over 50 health systems used Luma-powered AI workflows to save two million staff hours in 2025.
  • NewDays laid out its unique cognitive treatment platform that blends human and AI elements to help patients delay symptoms and preserve independence. The approach combines clinical assessments and psychotherapy with an AI companion named Sunny for exercises and support between visits. Their CIO Daniel is also awesome.
  • RevSpring debuted its first dedicated MCP server to give developers a grounding layer that connects AI models with data like provider quality, real-time availability, plan networks, and cost transparency – without the usual agent lag. The launch arrives as the ink is still drying on RevSpring’s acquisition of Trust Commerce.
  • TigerConnect took the lid off its new AI-powered Operator Console to replace legacy operator favorites (spreadsheets and Post-Its) with a cloud-native smart switchboard. Operator Console centralizes calls, code activations, and facility alarms in a unified interface, along with intelligent call routing and AI-recommended next best steps.
  • Wheel expanded its Horizon virtual care platform with a Clinical Action Layer that ingests patient and partner data (AI, wearables, labs, records), generates clinician-ready summaries, and intelligently routes patients into orchestrated workflows. It also debuted its new WheelX exchange that connects enterprises with the AI experiences built on Horizon.
  • Withings Health Solutions was showcasing its BPM Pro 2 connected blood pressure monitor and Body Pro smart scale, the dynamic device duo that MedStar Health is bringing to its Signature concierge medicine service to make the patient and provider experiences feel like consumer experiences. Their VP of VBC Patrick Sheehan also happens to be a fantastic interview.
  • Wolters Kluwer opened up its expert-curated medication data to AI developers with its Medi-Span Expert AI MCP that lets them easily spin up their own agentic AI workflows. Medi-Span’s safety guardrails have made it the go-to medication support for pretty much everybody that values accuracy over off-the-shelf convenience.

Many thanks to all of our awesome readers who caught us up on the latest and greatest at the show, and we’re looking forward to running it back at HIMSS in a couple weeks. Smash that reply button and let’s set something up!

LLMs Still Struggle With Medical Misinformation 

The Lancet Digital Health just published one of the largest-ever stress tests on medical misinformation in LLMs, and it looks like most models still struggle to separate fact from fiction.

Here’s the setup. Researchers probed 20 LLMs with over 3M prompts containing medical information from three different sources: social media posts, simulated clinical vignettes, or real hospital discharge notes with a single fabricated recommendation inserted.

  • Each prompt was presented in multiple versions, once with neutral wording to establish a baseline, then with a series of variations that were emotionally charged or leading.
  • Ten logical fallacies were also used to test how framing influences model behavior, such as appeals to authority (a physician said…) or popularity (everyone agrees that…).

LLMs love fake news. The susceptibility was shockingly high across all models, with the medical misinformation accepted in 32% of the neutral base prompts.

  • That jumped to 46% when the misinformation was embedded in formal discharge notes, but at least the models were more skeptical of the social media content (9%).

Other findings were more counter-intuitive. Eight of the 10 logical fallacies ended up reducing the misinformation acceptance rate rather than increasing it like the authors expected.

  • Only appeals to authority (+2.9 percentage points above the base prompts) and slippery slope prompts (+2.2pp) increased susceptibility, a relatively small impact considering appeals to popularity slashed it by nearly 20pp.
  • Larger models were generally safer, although the language and phrasing had a far greater influence than the parameter count alone. 
  • It was also surprising to see that the medical models performed worse than the general purpose models, with many having weaker lie detectors despite the specialization.

Improving LLM safety is about more than making bigger models. It’s about knowing how information gets presented by actual humans, and having guardrails in place that hold up even when that information is wrong.

The Takeaway

Benchmark performance isn’t real-world performance, and this study provides another reminder that a model’s ability to separate fact from fiction is often more important than its test scores.

ACCESS Might Be InACCESSible

The wait for CMS’ new ACCESS model payment rates is finally over, but the debate over whether or not they’re financially viable is just getting started.

Advancing Chronic Care with Effective, Scalable Solutions. ACCESS was designed to move more Medicare beneficiaries away from fee-for-service toward outcomes-driven models.

  • The program’s core mechanism for accomplishing that is Outcome-Aligned Payments (OAP), a per-beneficiary annual allowed amount to cover integrated care management for chronic conditions.
  • The end goal is to get more tech-forward companies to lean in on Medicare by rewarding them for using technology to improve patient outcomes.

That goal might be hard to reach. Here are the annual OAPs by clinical track and care period:

  • Early Cardio-Kidney-Metabolic (eCKM) – $360 initial, $180 follow-on
  • Cardio-Kidney-Metabolic (CKM) – $420 initial, $210 follow-on
  • Musculoskeletal (MSK) – $180 initial, N/A follow-on
  • Behavioral Health (BH) – $180 initial, $90 follow-on

Those numbers present some real challenges. They’re considerably lower than expected, and many of the companies that had already announced plans to participate are now being forced to reevaluate the decision.

  • For the sake of comparison, Medicare’s average annual Part B spending for a diabetic patient is around $700 under fee-for-service.
  • Asking providers to deliver comprehensive, tech-enabled care for half of that is a tall order, especially for services-heavy companies with humans in the loop.
  • Companies with an AI-first approach and an established patient pipeline might perform better, but even then the rates are so low that they’ll likely do little to motivate new entrants to Medicare given the infrastructure needed to comply with the program and achieve the desired outcomes.

The Takeaway

CMS has made it clear that it’s going to start taking bigger steps to control costs, but it also has to find rates that actually encourage companies to participate. Striking that balance is an unenviable task, but the initial consensus seems to be that ACCESS missed the mark.

The Patient You Lost Before They Ever Walked In

Thousands of patients are referred for procedures but vanish into the void because no one called them back within 48 hours.

By Shani Fargun, VP Healthcare at StackAI
Sponsored by StackAI

While the headlines at major cardiology conferences focus on AI that can read angiograms or predict arrhythmias, a quieter, unsexy revolution is happening in the back office, and it might be the key to actually using those advanced clinical tools.

The biggest bottleneck in modern cardiology is administrative friction. It’s the death by 1,000 faxes that occurs when a patient is referred for a TAVR, but the pre-op workup is trapped in a PDF from an external hospital. It’s the prior authorization that sits in a queue for weeks because a specific keyword was missing from the submission.

  • According to the AMA, 94% of physicians report that these administrative hurdles lead to delays in accessing necessary care.

Healthcare has a data problem. The industry runs on unstructured data. Referral letters, handwritten call notes, faxed labs, and denial letters make up the bulk of cardiac operations.

  • Nearly 80% of all healthcare data is unstructured and inaccessible to traditional automation. This forces highly trained clinical staff to spend hours acting as data entry clerks rather than treating patients.

Agentic AI is the solution. Agentic AI isn’t a chatbot or a diagnostic model, it’s a digital worker. 

  • Unlike traditional software that waits for a human to input data, Agentic AI can autonomously perform tasks across different systems.

How can agentic workflows change modern practices?

  • Patient Scheduling & Follow-Up – Agents autonomously handle the last mile of care coordination, reaching out to patients to schedule diagnostic testing, confirming procedure dates, and answering routine logistical questions without burdening clinical staff. This directly combats referral leakage, which costs health systems an estimated $971,000 per physician annually. 
  • Automated Prior Auth – Agents cross-reference patient charts against payer-specific guidelines to draft authorization requests that minimize technical denials. Download the free whitepaper of use cases for healthcare here.
  • Referral Velocity – Agents ingest incoming faxes and emails, extract clinical criteria, and draft the patient chart for review: reducing time-to-appointment from weeks to days.

The Takeaway

The future of healthcare starts with better flows. By automating the administrative burden, we allow interventionalists to focus on what they do best: treating patients.

Request a demo to see customized use cases for your organization here.

Garner Raises $118M and Becomes Care Navigation Unicorn

This year’s been good for the unicorns, and care navigation startup Garner just became the latest member of the herd after closing $118M of Series D funding at a $1.35B valuation.

Show us the incentives and we’ll show you the outcomes. Garner’s proprietary provider-ranking engine helps patients find the best doctors for their needs, then creates financial incentives to actually go and see them.

  • The engine runs on de-identified data from over 320M patients, which gives Garner the evidence it needs to identify doctors that produce the best outcomes and lowest costs.
  • Every metric answers an important question. Which doctors follow the latest research? Who avoids unnecessary procedures? Who gets patients healthy the fastest? 

Those top providers put up some impressive stat lines. 

  • 75% fewer complications
  • 60% fewer hospitalizations
  • 3x greater adherence to medical guidelines 

Garner takes it a step further. Instead of the usual services-heavy care navigation strategy, Garner’s engine has the incentives for real behavior change baked in. 

  • When employees choose a top provider, Garner picks up the tab – copays, deductibles, and even some procedures. 

The end result speaks for itself. Garner’s clients see a 12% average reduction in total healthcare spending, and their healthy employees are happy employees.

  • Garner already works with 700 organizations reaching 2.5M members, and revenue was up 130% last year as employers increasingly look beyond traditional approaches to combat rising costs. 

What’s next? The fresh funds were earmarked for bolstering the provider-ranking engine, scaling AI navigation capabilities, and growing its team to keep up with demand.

The Takeaway

If data on complication rates and hospital readmissions can help identify the best physicians, then it should also be able to reduce overall costs. Seems like a sound thesis, and Garner just scored another $118M to prove it.

Will Oracle Offload Cerner to Fund Datacenters?

The rumor mill was working overtime last week after a TD Cowen research note claimed that Oracle will have to offload Oracle Health – formerly known as Cerner – to fund its AI datacenter commitments.

It’s a tale of the times. A research note isn’t an official announcement, but the EHR market could be heading towards its biggest shakeup since Oracle first acquired Cerner in 2022.

  • The speculation revolves around Oracle’s massive $300B datacenter contract with OpenAI, which will apparently take $156B of capital expenditures to fulfill.
  • Add in contracts with Meta and Nvidia, and Oracle’s commitments swell to over $500B.

That’s a ton of CapEx. TD Cowen says Oracle will have to make some deep cuts to round up enough funds. That includes:

  • Selling Cerner to the highest bidder
  • Axing up to 30k jobs, about 15% of the current workforce
  • Exploring “bring your own chip” arrangements to lighten the load on Oracle’s books

Oracle’s back is against the wall. It’s already raised $58B in the last two months, and U.S. banks have started pulling back their lending.

  • Foreign banks are still supporting Oracle’s datacenter projects, but they’ve also raised their premiums to levels typically reserved for non-investment grade companies.
  • On top of that, Oracle is going to have a hard time recouping the $28B it just paid for Cerner. Since the acquisition, Cerner’s had a brutal VA implementation, a tough rollout with the DoD, and Epic’s been eating its lunch.

Who has deep enough pockets to acquire Cerner? It’s a short list.

  • Microsoft is a prime suspect. It’s already heavily invested in healthcare through Nuance and Azure, so an EHR could potentially create a compelling end-to-end cloud lineup for its existing customers.
  • Google and Amazon also probably wouldn’t mind having Cerner’s customer base as an anchor for their cloud ambitions. They both have full war chests and established healthcare ventures like Verily and One Medical, but they also share a track-record of expensive lessons in the industry. 

The Takeaway

Recent struggles aside, Cerner is one of healthcare’s true industry titans. It shaped decades of innovation and thousands of careers. Now it might end up as a line item to fund GPU clusters.

Epic Shakes Up Scribe Market With AI Charting

The wait is over. Epic’s scribe has arrived, and it’s packing a lot more than ambient notes.

“AI Charting” goes beyond transcriptions. The fully built-in feature not only listens during patient visits and drafts notes, it also queues up orders based on the conversation.

  • The initial release allows clinicians to personalize the note structure using voice commands (Ex. asking to format the history of present illness as a bulleted list).
  • Epic is positioning AI Charting as the killer app for its Art clinical copilot, which also has a pre-visit Insights tool that’s apparently already being used 16M times per month.

Distribution is king. Over 40% of U.S. hospitals are on Epic, and an AJMC study from just last week showed that two-thirds of those hospitals have already adopted ambient AI.

  • AI Charting is breaking onto the scene through one of healthcare’s biggest distribution channels, and Epic has a ton of levers it can pull with pricing and bundling to start stealing share (DAX Copilot, Abridge, and ThinkAndor accounted for ~80% of Epic hospitals in the recent study).
  • Rather than charging a per-user-per-month fee like most ambient AI platforms, STAT reports that Epic plans to have a separate license for AI Charting, with the price varying by org size and utilization to get the tool in as many hands as possible.

It’s time to differentiate. The race is on for established players to prove they can deliver value that Epic’s integrated approach can’t match.

  • That means tackling problems that are too messy for Epic to touch (Abridge bringing real-time prior auths to the point of conversation), or too specialized for it to get right with so many other plates spinning (Nabla raising the bar for AI safety with world models).
  • Epic is working closely with Microsoft to get new features online quickly, but nailing multiple specialties in countless languages could still prove to be a job that’s better suited for a company with a dedicated focus.
  • Epic might own the “operating system” almost as much as Microsoft owns Windows, but just because MS Paint exists doesn’t mean the world doesn’t need Adobe Photoshop.

The Takeaway

Ambient scribes proved how fast health systems would layer on their own AI if Epic couldn’t keep up, and we’ll now have to wait and see if the cost and experience of Epic’s scribe is enough to compete with the flock of ambient AI innovators dedicated to this problem.

Get the top digital health stories right in your inbox