Subscribe | Advertise | PRO Member Hub

TOGETHER WITH ZINGAGE

Perspectives from the people building the future of health AI…

We sat down with Victor Hunt, Co-Founder and CEO of Zingage, to talk about what it takes to actually run home care operations at scale and why most existing software falls short.

Zingage is building an AI-powered “operator” for home care agencies, designed to handle the day-to-day logistics of scheduling, staffing, compliance, and communication. Instead of layering automation on top of existing systems, the company is focused on orchestrating the full workflow end-to-end across patients, caregivers, and payers.

Victor shared how Zingage approaches automation vs control, what has changed in AI to make this possible now, and why solving home care requires rethinking how day-to-day operations are actually run.

Let’s start from the top. What’s the founding story behind Zingage?
We didn’t find home care. Home care found us.

I grew up inside my grandmother’s home care agency in New York, so I saw how these businesses actually run. Later, my mom had a traumatic brain injury and we needed caregivers at home, which gave me the patient-side view as well.

My co-founder Daniel had a similar connection. His grandfather had Alzheimer’s, and his family struggled to find consistent care during COVID. Even when they did, call-outs meant the problem was never fully solved.

When we started looking at what production-grade AI could actually do, home care stood out immediately. The EMRs had improved slightly, but the real work of running the business was still manual, reactive, and dependent on people holding everything together. This wasn’t a problem for another SaaS tool. It needed to be rebuilt from first principles.

What did growing up around home care teach you that outsiders often miss?
Home care is not a staffing marketplace. Matching caregivers to patients is maybe 10% of the work. The other 90% is everything in between.

A caregiver calls out. A patient’s schedule changes. A license expires. A note doesn’t get signed. Each of these becomes a real-time problem that has to be resolved quickly or the agency loses revenue and the patient loses care.

Agency operators are very good at their jobs. They are not behind on AI because they are lazy or technophobic. They’re behind because the tools don’t reflect how the work actually happens. EMRs were built for documentation. What agencies needed was something to run the operation.

Daniel spent three months answering 5am call-outs before writing any code. That’s where we learned the EMR is often out of date, most operational knowledge lives in people’s heads, and admins spend hours each day on the phone acting as the integration layer between systems and people.

What is fundamentally broken about home care today, and why is this solvable now?
Home care has the demand, reimbursement, and workforce. What it doesn’t have is an orchestration layer.

Medicaid, Medicare, the VA, and private insurers authorize hundreds of billions of dollars of care each year for people to receive care at home, but a huge amount of that care never gets delivered. The average agency services only about 70 hours for every 100 hours of authorized care. That gap is operational loss.

Things break across intake, scheduling, staffing, and compliance because those are not separate workflows. They are the same workflow showing up at different points. If a call is not answered, a shift is not filled, a clock-in is not compliant, or a note is not written, the agency loses revenue and the patient may lose care.

Three things changed recently. First, the models got good enough to hold context, reason across systems, and take action. Second, Electronic Visit Verification (EVV) mandates created structured compliance data that did not exist five years ago. Third, buyers are ready. A year ago, agencies asked whether AI works. Now they ask how fast they can deploy.

Give us the quick overview. What is Zingage and what does Operator do?
Zingage is the AI operator for home care. We give agencies a full back office delivered as AI agents trained on how that agency actually runs.

Operator is the name of the full platform. It has three pillars: Care Delivery, Revenue & Growth, and Workforce Development. Care Delivery is live everywhere today, and Revenue and Workforce are live in some agencies.

The way I describe it to agencies is simple: we are not replacing your EMR, we are giving your back office leverage. EMRs are systems of record and document what happened. Zingage is the system of action and makes things happen.

Today, we manage more than 10,000 patient visits per day, handle more than 11,000 inbound and outbound conversations per day across phone, SMS, and EMR messaging, and resolve nearly all workflows autonomously. The customers that turn us on for 24/7 coverage are growing significantly faster than the national average, and 100% of them are growing.

Walk us through a typical workflow. What actually happens when Zingage handles a call-out?
The cleanest example is a 5am caregiver call-out for a 6am shift. That is probably our single most frequent workflow.

First, our AI receptionist picks up on the first ring, recognizes the caregiver, confirms the call-out, asks diagnostic questions, and logs everything to the context graph. Then the agent pulls the patient’s care plan, authorization hours, acuity, and preferences from the EMR. If it is a must-staff patient, like a VA patient with dementia, the system flags it as escalated immediately.

From there, the agent runs parallel outbound calls to eligible caregivers ranked by license, proximity, prior patient match, preferences, and overtime status. A human admin might call three caregivers one by one. We can call 20 in parallel.

Once a replacement confirms, the agent updates the schedule, notifies the family, messages the care circle, and creates the EVV-compliant audit trail. At the end of the shift, it verifies clock-in and clock-out against the care plan, flags exceptions, writes the compliant note, and routes anything ambiguous to a human for QA.

Every one of those steps used to be an admin on the phone. Now the system handles the workflow in minutes.

Where does AI show up in the product, and what does the system look like under the hood?
AI shows up across the product in three main ways.

First is voice agents. Our AI receptionist handles inbound and outbound calls across caregivers, patients, families, referral partners, and payers, with context carried across conversations.

Second is decision agents. These manage scheduling, coverage, escalation routing, prior auth navigation, and credentialing. For a single call-out, the system can make dozens of decisions across the EMR, payer systems, and the agency’s internal context.

Third is documentation agents. These handle EVV enforcement, clock-in and clock-out validation, note generation, and compliance audit trails.

The key difference is that we are not automating individual tasks. We are orchestrating the full workflow end to end, which allows the system to actually run the operation rather than just assist it.

Under the hood, the system is built around three layers. The context graph is a real-time model of caregivers, patients, preferences, and relationships. On top of that is the agent layer, which executes workflows, and the integration layer, which connects into EMRs and external systems so actions can be taken directly.

We use frontier models from OpenAI and Anthropic as the reasoning backbone, fine-tune where needed, and build the orchestration layer, context graph, and voice stack in-house.

How do you decide what is automated versus what stays human-in-the-loop?
Today, over 90% of workflows resolve autonomously. For those that escalate to a human, either a customer-side manager or our QA layer, it’s based on provider preference and confidence thresholds.

Automation covers scheduling, call-outs, intake, compliance, and routine communication. Humans stay involved for clinical judgment, sensitive conversations, and edge cases.

There are three rules: everything must be auditable, customers control thresholds, and automation is earned over time by proving it on lower-risk workflows first. Most agencies start with workflows they’re comfortable handing off then expand as they build confidence in our agents’ ability to reason and operate.

Agencies still want control over relationships. Patients, caregivers, referral partners, and the voice of the agency are theirs. Most customers even rebrand the agents under their own coordinators’ names. What they do not want is to keep managing back-office execution manually.

How does Zingage integrate with existing systems, and what does implementation look like?
We layer on top of existing systems. The EMR remains the system of record, and Zingage becomes the system of action.

We have partnerships with the major home care EMRs, including WellSky, HHAeXchange, AlayaCare, Axxess, and SwyftOps. The goal is not to rebuild those systems. The goal is to run the business on top of them.

For smaller businesses, we can go live very quickly, sometimes in less than 24 hours, by starting with best practices and industry patterns. Over time, it takes about 6 weeks for the context graph to fully acclimate to that agency’s patient and caregiver base and become autonomous.

For larger or more sophisticated businesses, implementation can take longer because we go deep into workflows and unpack business logic that lives in systems of record, paper files, and people’s heads. Sometimes the biggest blocker is not technical. It is internal alignment. Different leaders may have different views of how the business should run. We help reconcile that by adapting the AI to their needs rather than forcing a top-down workflow.

Who is your core customer, and what are the main use cases today?
We serve both SMB agencies, usually from early-stage up to $25M in revenue with one to three locations, and enterprise agencies, typically $50M to $1B, often multi-state and Medicaid-heavy.

The most common use cases today are scheduling and call-out coverage, compliance documentation, patient and family communication, credentialing, intake and referral conversion, caregiver engagement, and HR triage.

The anchor use case is care coordination. Once we’re running scheduling and call-outs, customers see value in week one, which earns us the right to expand.

Customers don’t buy AI. They buy outcomes. They don’t care that your system can call 100 caregivers if the shift doesn’t get filled. They care about billable hours, coverage, growth, and whether their best people are back in the field instead of on the phone.

Nights and weekends are where value shows up fastest. Calls don’t stop at 5pm, but staff often does. Turning on Zingage 24/7 means the owner can finally stop being on-call.

How do you measure success and ROI?
Operationally, we track autonomous resolution rate, visits managed, conversations handled, and fill rate on called-out shifts. Today that’s over 90% autonomous resolution, 10,000+ visits managed per day, 11,000+ conversations handled daily, and ~80% fill rates on call-outs.

At the customer level, it comes down to two numbers: weekly billable hours and cost to deliver care. Billable hours up, opex down.

Senior Helpers is a clear example: 35% revenue growth and 12% opex reduction in 12 weeks. We also have agencies, like one in Florida, using Zingage data on staffing continuity, rehospitalizations, and fill rates to negotiate better reimbursement rates and drive more volume with payers.

At day 7, we’re typically handling core scheduling workflows. By day 45, we’re often live 24/7 and the scheduler team is redeployed. By day 60, the context graph is deep enough that we can help agencies do things they never had capacity for before like proactive patient and family engagement.

How does the system learn over time, and where do you see this going long term?
The system improves through three nested loops.

Within each customer, every call, schedule change, and EMR write updates the context graph. By month three, we know the agency better than a new hire.

Across customers, regulatory changes, payer updates, and EVV mandates are captured once and propagated across the network. No individual agency could build that alone. Across agents, every human correction feeds back into the system.

We’re not just producing raw scheduling data. EMRs already have that. We’re building every agency’s context graph: the relationships, preferences, and operational knowledge that define how an agency runs.

Long term, the cost to deliver home care dramatically drops leading to a lot more healthcare happening in the home. The back office should become close to fully automated, while relationships and clinical judgment stay human. Every hour we remove from scheduling is an hour a care coordinator can spend onboarding new patients and caregivers.

Five years from now, the goal is for Zingage to expand how much care happens in the home by empowering providers with an operating layer they can scale on. Every authorized hour of care, every caregiver schedule, every payer authorization, every EVV compliance check, flowing through one orchestration layer.

What stood out, what’s tricky, and why it matters…

Zingage is going after one of the least visible but most consequential problems in healthcare: the day-to-day logistics of actually delivering care in the home.

As more care shifts out of hospitals and into home settings, the operational burden does not go away. It gets harder. Agencies are coordinating caregivers, patients, payers, and compliance requirements in real time, often over phone calls and fragmented systems. The result is a system where a meaningful portion of authorized care never gets delivered, not because demand or supply is missing, but because coordination breaks down.

Zingage’s approach is to treat this as an orchestration problem. Instead of building point solutions for scheduling or documentation, they are building an AI operator that sits on top of existing systems and handles the full loop from intake to staffing to compliance.

The company is still early but moving quickly. Zingage raised a $12.5M seed round led by Bessemer Venture Partners, South Park Commons, the founders of Ramp and others. Victor is a repeat founder with firsthand experience in home care, and the team brings strong technical depth that shows up in how the product is designed and deployed.

What stands out is seizing the moment in this important market. This is not about making admins slightly more efficient. It is about reducing the need for manual coordination across large parts of the back office, while keeping humans involved where judgment and relationships matter.

What stood out

What’s tricky

Final thoughts

Zingage is tackling a layer of healthcare that is easy to overlook but critical to making the system function. As more care moves into the home, the ability to reliably coordinate that care becomes just as important as clinical quality.

What is interesting here is not just the use of AI, but where it is applied. Instead of focusing on documentation or decision support, Zingage is focused on execution. Making sure shifts are filled, visits happen, and care is actually delivered.

That is a harder problem, but also a more valuable one if it works.

They are approaching it with a clear point of view, starting with the highest-friction workflows and expanding from there. The combination of operator empathy, technical depth, and early traction puts them in a strong position to keep building.

It is also a large and timely market. Home care is already a $300B category, and the gap between authorized and delivered care remains significant. Closing that gap has real implications for access, cost, and outcomes across the system.

If they can continue to prove reliability at scale and deepen the system’s understanding of how agencies operate, Zingage has a credible path to becoming part of the underlying infrastructure for home-based care.