The origin of XiFin is a complaint disguised as a company. In the 1990s, Lâle White was running finance inside the laboratory business and confronting a peculiar mess: enormous numbers of relatively small claims, each exposed to a thicket of payor rules, coding requirements and exceptions. The available commercial software did not fit. Her teams built their own systems, three generations of them, but an internal IT department always had louder priorities. In 1997, White left to build the product she could not buy.
That detail explains more about XiFin than its current AI vocabulary. The company did not begin by asking where software might be applied. It began with an expensive operational irritation that horizontal software treated as an edge case. White co-founded XiFin with product chief Mike Coats and technology chief Jeffrey Yates. The first commercial platform arrived in 2001, sold as an internet-delivered service before “cloud” became mandatory copy on every software homepage.
Today, XiFin sits between a patient encounter and the money owed for it. Its systems validate insurance, submit claims, interpret payor responses, route denials, assemble appeals, estimate patient responsibility, manage documents and show finance teams what is clogging the pipe. It also sells laboratory information software, patient and provider portals, pharmacy clinical tools, and outsourced billing. If healthcare delivery is the performance, XiFin works the windowless room where the box office reconciles every ticket.
The first thing that failed was the obvious thing
XiFin’s first failed system was not XiFin. It was the succession of in-house tools White’s laboratory teams had been forced to maintain. They knew the workflow but could not give it sustained engineering attention. That is a familiar enterprise-software opening: the spreadsheet or internal app is good enough until transaction volume, regulation and exceptions turn maintenance into a second business.
The company’s initial wedge was revenue-cycle management for labs, where a single missing field can postpone payment across thousands of claims. The economic buyer is usually a CFO, revenue-cycle leader, lab executive or operations chief. Users range from billing staff and coders to patient-service teams and data analysts. The customer is not buying “payments” in the cheerful checkout-page sense. It is buying fewer touches, faster reimbursement, better visibility and a lower cost to collect.
That focus produced a defensible kind of expertise. A laboratory claim is not the same as a hospital stay, a radiology read or a vaccination billed under the medical benefit. But they share a family resemblance: specialty rules, high volume, reimbursement pressure and a chain of systems that rarely agree on the first try. XiFin learned the claim and then followed the pattern.
Two acquisitions widened the same bottleneck
In July 2021, XiFin bought Computerized Management Services, a radiology billing specialist. Five months later it acquired OmniSYS, whose technology served more than 30,000 pharmacies, including four of the five largest national chains at the time. On paper, imaging centers and pharmacy counters look like separate businesses. From XiFin’s desk, both throw off complex medical claims, documentation requirements and patient communication chores.
The acquisitions changed the company’s addressable market and its product surface. Radiology brought billing, coding, credentialing and consulting. OmniSYS brought medical-benefit claim processing, clinical documentation and patient messaging inside pharmacy workflows. XiFin now says it serves 11 of the largest public laboratories, four of the five top integrated delivery networks and more than 30,000 pharmacies. Its named customers and partners include Sonic Healthcare divisions, Caris Life Sciences, Northwell Health Laboratories and Ambry Genetics.
XiFin’s business model is a mixture of recurring software and people-heavy services. Customers can license the platform, outsource portions of billing, or combine technology with managed RCM, coding and consulting. There is no public price card. This is negotiated enterprise software; cost changes with claim volume, interfaces, specialty, implementation and service scope. That opacity makes comparison harder, but publishing a fake per-seat number would be worse.
The AI pitch works because the boring layer came first
XiFin renamed its flagship RPM product Empower RCM in 2025, after reporting 19 percent revenue growth for 2024. That September it raised an undisclosed growth-capital round led by Goldman Sachs External Investing Group alongside majority investor Avista Healthcare Partners and other institutions. In 2026 came Empower AI, an ecosystem meant to coordinate predictive models, generative tools and task-specific agents across the billing lifecycle.
The appeals agent is the cleanest example. It can ingest a case, normalize plan and encounter data, inspect live payor portals, identify coverage policy, draft a cited appeal letter, assemble supporting records and flag cases that need judgment. XiFin says it can reduce manual appeals work by as much as 90 percent. The company also applies AI to insurance mapping, patient estimates, payor-response interpretation, document extraction and exception prioritization.
This is where XiFin differs from a generic AI billing entrant. Its raw material is not simply a language model. It is the data model, payor connectivity, configured edits, adjudication history and specialty workflow built over decades. That does not make the system infallible. It makes the company’s claim more testable: can the intelligence move an exception to the right next action, inside a secure production workflow, without inventing facts or dropping the audit trail?
A customer changed its mind when manual work stopped scaling
One XiFin case study offers a useful miniature. A molecular diagnostics company acquired a complex test. For the first year, it left billing with the vendor that already handled it. Then the cracks appeared: 75 percent of test volume required prior authorization, the vendor relied heavily on manual processes, documentation retrieval was clumsy and appeals were frequent. The buyer’s problem was not that outsourcing had failed in principle. The workflow had outgrown the vendor’s automation and integration.
The company moved the billing in-house in phases on Empower RCM. Prior authorization could fire from the test order; documents moved from the laboratory system without manual chasing. The customer reported cutting billing costs by 50 percent. What can another operator copy? Stage the migration, automate at the earliest reliable event, connect source documentation, and measure cost, denials and reimbursement instead of celebrating task counts.
When the playbook breaks
- A small practice with simple claims may not need an enterprise specialty platform.
- Bad order data and brittle interfaces merely send mistakes through the system faster.
- Agentic workflows need permission controls, audit logs and a human path for judgment-heavy appeals.
- A team unwilling to redesign ownership and exception queues will keep the old bottleneck under new software.
- Organizations wanting transparent, self-serve pricing will not find it here.
The bill is also a patient experience
Revenue-cycle software is usually described in finance language, but the patient sees its mistakes. An inaccurate estimate can turn a necessary test into a surprise bill; a missing authorization can postpone the answer; a confusing statement can send someone to a call center that has no useful context. XiFin’s portals separate the needs of patients, ordering physicians and provider organizations, an architectural choice that recognizes they are not interchangeable users.
Ambry Genetics gives the strategy a human shape. After evaluating multiple RCM vendors, the molecular diagnostics company selected XiFin and prioritized out-of-pocket estimates and the patient portal. The two companies developed near-real-time text and email notices when an estimate was ready. A patient could review the amount, approve the test and prepay through the portal. For Ambry, the same workflow promised lower service cost and fewer write-offs while making a genetic test less financially mysterious. It is a useful reminder that a clean claim and a clear patient message are two views of the same transaction.
A specialist in a market full of giants
XiFin competes with broad RCM platforms and service operators such as Waystar, R1 RCM, Optum and Change Healthcare, Oracle Health, Experian Health, FinThrive, athenahealth and Infinx. Several can offer more hospital-scale breadth, clearinghouse reach or outsourced labor. XiFin’s argument is narrower: diagnostics, ancillary medicine and pharmacy contain enough peculiar economics to reward a purpose-built system.
That niche is getting broader. White has said XiFin wants to transfer its model into more healthcare sectors, and recent capital is intended for product development and growth. The risk is familiar: expanding the platform until specialty depth becomes generic breadth. Its two 2021 acquisitions worked because they mapped onto the same bottleneck. Future moves will need the same discipline.
XiFin’s most useful lesson is not to sprinkle AI on accounts receivable. It is to earn the right to automate. The company spent years converting payor behavior and operator knowledge into rules, workflows and data. The models arrived after the plumbing. In healthcare, where a confident mistake can delay care or cash, that order of operations may be the whole story.