Health plans turn to new data platform as errors and compliance risks pile up

Provider directories may be one of the least visible pieces of health plan infrastructure, but they shape some of the most consequential moments in care: who patients can see, how quickly they get appointments, and whether the bill that arrives later matches what their plan promised. 

Yet provider directory accuracy remains at 45–55 percent for most payers. For years, the system has limped along on outdated addresses, closed practices, and clinicians listed as accepting insurance they dropped months ago. Insurers have tried periodic clean-ups, providers have been pressed to update their details across a dizzying number of portals, and none of it has produced durable accuracy.

Now, one startup says it has a fix. Certify, a provider data intelligence company, has launched Provider Hub, a platform built to give health plans a single, continuously updated source of truth about physicians and facilities. It’s one of the many startups that aims to clean up healthcare’s data problems and better connect patients to the care they need. 

The cracks in healthcare’s data infrastructure

A Centers for Medicare & Medicaid Services (CMS) audit of Medicare Advantage plans found that 52 percent of provider locations contained at least one inaccuracy. Errors ranged from providers not being at listed locations (33 percent) to wrong phone numbers to false claims about accepting new patients. For patients, those inaccuracies can mean hours wasted, surprise bills, or delays in care.

“Provider data is the invisible backbone of healthcare, yet it remains broken,” said Anshul Rathi, founder and CEO of Certify, in an exclusive interview with DHI. “Every stakeholder operates from a different playbook. Health plans maintain their own systems, providers are asked to enter the same information across multiple portals, and outdated technology cannot keep pace.”

Those weaknesses have become harder to ignore under a wave of new regulations. The Transparency in Coverage Rule requires plans to publish negotiated rates and cost-sharing information in machine-readable formats. The No Surprises Act, in effect since 2022, requires health plans to maintain accurate provider directories and establish verification processes, with plans facing financial penalties for publishing incorrect network information.

Meanwhile, the National Committee for Quality Assurance (NCQA) announced shortened credentialing verification timeframes in its 2025 updates, released in August 2024, to enable organizations to work with more current data and accelerate network enrollment.

Rathi argues that these rules expose a deeper structural flaw in healthcare’s data ecosystem: the mixing of disconnected sources that undermines the accuracy of records. “Think about it as a dirty pool of water,” he said. “Once all these disconnected inputs mix together, it is hard to know what is right and what is wrong.”

Cleaning up provider data with AI tools and human expertise

Certify aims to clean up that pool by synthesizing data from credentialing systems, directories, claims, rosters, and other sources, then applying validation rules built from more than 1,600 primary data feeds. AI and human reviewers work together to spot duplicates, resolve conflicts, and maintain what the company calls a continuously updated “source of truth.” Health plans can set custom rules to prioritize which data takes precedence when discrepancies appear.

The company, which entered the market in 2021 as a credentialing platform, has already signed multiple customers seeking to reduce duplication and compliance risk. It reports metrics such as an 83 percent reduction in credentialing turnaround time and an 85 percent drop in quality-checking labor, though it has not disclosed which organizations achieved those results or under what conditions.

The Provider Hub platform integrates with credentialing, contracting, directories, and claims systems while also continuously monitoring for license expirations, sanctions, and outdated records. This ongoing surveillance is designed to meet rules requiring monthly credential checks.

“At CertifyOS, we use AI as a powerful tool, but always pair it with human expertise and a continuously updated source of truth so health plans can trust the data behind every decision,” Rathi said. Artificial intelligence, he added, can detect errors faster than humans, but “human oversight remains essential for reliability.”

Technology is only part of the solution

Directory errors can erode member trust, inflate administrative costs, and trigger penalties. Studies have tied those inaccuracies to surprise medical bills, particularly in behavioral health; a 2020 Health Affairs study found that 53 percent of patients who used mental health provider directories encountered inaccuracies, making them more likely to receive surprise out-of-network bills.

Rathi said Provider Hub differs from legacy systems that merely digitize old workflows, but rather replaces them. “Because we are a data company at heart, we enable organizations to go from storing data to unlocking its enterprise-wide potential.”

Whether the approach can finally solve health care’s provider-data problem is less clear. Years of escalating federal and state mandates have yet to yield reliable directories. Policy experts have suggested that without stronger penalties or standardized data formats, no single technology will fix the systemic flaws.

Health plans are now operating under new Medicaid and CHIP provider-directory requirements that took effect July 1, 2025, under the Consolidated Appropriations Act of 2023, which mandate quarterly updates and standards-based APIs. Those rules arrived the same day as NCQA credentialing changes, a convergence that has pushed more payers toward automation.

Still, technology may not be enough to resolve a problem rooted in misaligned incentives and fragmented accountability. This new generation of platforms may deliver lasting accuracy across a fractured system, but the industry’s progress will likely hinge less on what software can automate and more on whether health plans and providers can finally align around a shared definition of trust in their data.

As Rathi put it, the challenge also represents an opportunity, “When provider data is continuously accurate, every workflow downstream improves.”