Companies: Assort Health · Nextech
Assort Health and Nextech are codeveloping an integrated AI agent offering for ophthalmology practices, with a shared API roadmap and deeper access to the workflows inside Nextech's specialty platform. The important signal is not simply that another EHR has an AI partner. It is that patient-access AI is becoming part of the specialty operating system itself.
At DuPage Ophthalmology, a Nextech practice, Assort says its AI handles roughly 500 patient calls each week. The practice reports more than 60% of call time saved and expanded the system into after-hours coverage.
A patient calling an ophthalmology practice rarely knows the exact appointment they need.
The correct answer may depend on subspecialty, prior testing, injection cadence, visit history, insurance, referral status, location, urgency, and clinical escalation rules.
That makes ophthalmology a useful stress test.
An agent that can answer the phone but books the wrong retina visit simply moves work downstream.
An agent that understands the specialty well enough to complete the transaction removes work.
The market is moving from “can AI hold a healthcare conversation?” to “can AI complete the correct specialty workflow inside the system of record?” That is a much higher bar — and a much more valuable product.
Assort already had a proprietary Nextech integration.
The new partnership turns that integration into a shared product roadmap for Nextech customers.
That changes the distribution model.
Instead of Assort selling around the EHR, it can increasingly sell through the EHR ecosystem.
For Nextech, the partnership adds an AI execution layer without requiring the company to build every agentic workflow internally.
For the practice, the agent can use patient, appointment, provider, location, and scheduling data already inside the operating system.
The AI gets more useful because it gets closer to the workflow.
The announcement is framed around patient experience, but the economics sit directly in front-end RCM.
If an inbound call is not answered, there may be no appointment.
If a referral is not worked, there may be no visit.
If a patient is routed to the wrong provider or appointment type, staff has to fix the error later.
If a balance needs resolution before scheduling and the workflow does not catch it, the financial issue simply moves downstream.
Assort now supports scheduling, intake, referrals, after-hours coverage, medication requests, payments, and task management.
Those are not peripheral features.
They increasingly define how demand becomes a billable encounter.
One line in the announcement is especially important.
Assort argues that the next generation of healthcare AI should be measured by requests resolved, not conversations handled.
That is the right standard.
Healthcare automation has spent years celebrating call containment and deflection.
But a contained call that leaves a task for staff tomorrow is not true automation.
The better question is:
Did the patient get to the right next step without creating new work?
Assort says customers across its platform have seen a 5% lift in appointment volume, a 115% increase in labor capacity, up to $4.3 million in revenue captured per 100 providers, 97% resolution rates, and 79% referral-to-appointment conversion.
Those are company-reported results and should be evaluated practice by practice.
But they show why patient access is increasingly being bought as a financial product, not just a call-center product.
The buyer is not only asking whether the phones are answered.
The buyer is asking whether more demand converts into care with less labor.
Scheduling is the obvious wedge.
Referrals are potentially more valuable.
A referral can fail at multiple points:
Assort's new Referrals agent is designed to work the referral through scheduling rather than stopping after document intake.
That is important because referral leakage is revenue leakage.
An AI system that closes the loop can affect both access and downstream collections before the claim exists.
There is a broader market implication.
General-purpose voice models are improving quickly.
That means the defensible value may migrate away from the underlying model and into:
Ophthalmology is a good example because workflow nuance is operationally meaningful.
The winning product does not just understand English.
It understands how a retina practice actually runs.
Assort and Nextech are moving toward a model where AI is no longer an overlay on the specialty practice.
It becomes part of the operating system.
Nextech owns the core specialty workflow.
Assort owns the agent layer.
The integration connects patient intent directly to the system where the transaction gets completed.
That is where front-end RCM is heading: fewer standalone bots, more agents embedded into the platforms that already control scheduling, referrals, intake, payments, and patient flow.
The strategic question for specialty EHR vendors is no longer whether they add AI.
It is which agent layer becomes native to their installed base.
Daily analysis on patient access, AI, payer behavior, deals, and the infrastructure reshaping revenue cycle.
Source: Assort Health / Nextech via PR Newswire · RevCycleAI analysis · October 11, 2026