Vendor reference · 2026
Every serious risk adjustment software vendor, on one page.
Ten platforms that suspect, capture, code, or audit HCCs for Medicare Advantage plans, IPAs, MSOs, and ACOs. Each is compared on capture model, AI approach, RADV defensibility, submission-lifecycle visibility, and whatever independent validation exists. Every claim carries a date and a link. There is no composite score here; recommendations are organized around the job you are hiring the software to do. The publisher is disclosed below.
Picks by job
Start from the job, not the brand.
KLAS handed its 2026 risk adjustment awards to three different vendors for three different jobs. Below are the strongest fits we can defend for each job as of August 2026.
-
Unified, real-time risk operations
Prospective and retrospective capture reconciled on one member record shared with quality, pharmacy, and care management, V28 trumping applied live, and each diagnosis tracked from 277CA acknowledgement through the MOR. Tradeoffs: founded in 2025, no KLAS rating, and a smaller install base than anything else on this page.
Disclosure: Pelica Health publishes this directory. Weigh this pick accordingly; the competing picks below are stated just as plainly.
-
Coding & retrieval throughput
Two 2026 Best in KLAS wins, Risk Adjustment (Coding, Retrieval and Compliance Solutions) at 90.5 and Outsourced Coding at 94.9, plus a self-reported 4.5 million charts coded a year at 98% accuracy. Shortlist it first when the bottleneck is records retrieved and charts coded before a deadline.
-
Scaled retrospective review & RADV response
Built for high-volume chart review, with an all-in-one RADV Audit product shipped October 2025 and managed CMS and RADV-IVA submissions. Since EVE Hybrid AI launched in January 2026 it competes on prospective suspecting too.
-
Point-of-care & in-home capture
2026 Best in KLAS in Risk Adjustment: Point of Care and In-Home Health Assessments at 91.2. Its PayerOne Risk product sits exactly where V28 pushed the capture decision: the encounter itself.
-
Enterprise payer data platform
Risk adjustment as one workflow on a unified payer data model, with 2026 Best in KLAS in Data Analytics Platform for Payers at 90.5 and a shift toward outcome-based, per-task pricing announced April 2026. The right shape for an enterprise plan standardizing its data layer first.
Side by side
The full comparison.
Positioning is summarized from each vendor's public materials and dated announcements. What any given customer actually receives depends on contract and scope, so verify current details with the vendor before you buy.
| Vendor | Capture model | AI approach | RADV & audit | Submission lifecycle | 2026 KLAS | Best fit |
|---|---|---|---|---|---|---|
| Apixio (Datavant) | Retrospective, prospective, and concurrent | Machine-learning code prediction over the chart, reviewed by certified coders | AI-powered HCC auditing and compliance tools | Coding-cycle focused | Within Datavant's awards | MA and ACA plans that need coding completeness at volume |
| Arcadia | Retrospective-led analytics | Data lakehouse with risk analytics | Analytics view, not a coding engine | Analytics view | Not awarded in category | Health systems and ACOs building a data foundation |
| Cozeva (Vatica) | Prospective, point-of-care and in-home | PayerOne Risk, built on native-FHIR certified measure engines | Point-of-care documentation support | Not its primary focus | 91.2 POC & in-home | Plans and groups prioritizing capture at the encounter |
| Datavant | Retrospective coding and retrieval at volume; also sold as an outsourced service | Clinical Insights Platform; NLP predictions reviewed by certified coders | Compliance solutions; audit-grade coding review | Coding and retrieval focused | 90.5 coding/retrieval · 94.9 outsourced | Plans that hand off retrieval and coding volume |
| ForeSee Medical | Prospective, point-of-care | NLP disease-detection engine inside the EHR | InstaVu links each suspect to its source page | Not its primary focus | Not awarded in category | Practices that want suspects raised inside their EHR |
| Innovaccer | Both, through the Galaxy payer platform with Gravity agents on top | Unified data model with an AI agent layer | Handled inside broader risk and quality workflows | Tracked inside the platform | 90.5 payer analytics | Enterprise plans consolidating onto a single data layer |
| Navina | Prospective, point-of-care | EHR-native condition detection presented as a clinician copilot | Suspects link back to chart evidence | Not its primary focus | Won Clinician Digital Workflow in 2025, not 2026 | Physician groups whose capture happens in the exam room |
| Pelica | Both, in real time, against a single member record | Risk Adjustment Copilot plus an action layer that executes follow-through | Per-HCC chain-of-custody; trumping computed live | Per diagnosis, 277CA through MOR in a single view | Not yet rated (founded 2025) | Risk-bearing IPAs, ACOs, and plans consolidating point tools |
| RAAPID | Retrospective and concurrent | Neuro-symbolic AI with knowledge-graph clinical NLP | MEAT-based audit trail; dedicated RADV solution | Coding and audit focused | Not awarded in category | Teams prioritizing defensible, traceable coding at scale |
| Reveleer | Retrospective abstraction, plus prospective suspecting added January 2026 | EVE Hybrid AI: deterministic rules written by clinicians layered over generative extraction | All-in-one RADV Audit SaaS; CMS and RADV-IVA submissions managed for the plan | Submission and audit focused | Not awarded in category | Plans with high chart-review and audit-response volume |
The directory
All ten vendors, alphabetically.
Alphabetical order, so nothing on this list reads as a ranking. Each entry links to a fuller profile with a fact sheet and dated sources.
-
Apixio
Coding engine · part of DatavantDatavant acquired Apixio's value-based care suite in September 2024. The practical effect for a buyer is that Apixio is now a component inside the Datavant platform, with contracts and roadmap running through Datavant. The engine itself still covers retrospective, prospective, and concurrent coding, with machine-learning predictions checked by certified coders and a separate AI-driven HCC auditing product.
-
Arcadia
Data platformArcadia sells a data foundation: a lakehouse, a longitudinal patient record across large populations, and analytics that include risk adjustment views. It makes no claim to be a coding engine and should not be scored as one. Nordic Capital took a majority stake on July 2, 2025. Organizations whose bottleneck is fragmented data tend to buy this layer first and add capture workflow afterward.
-
Cozeva
Point-of-care · merged with Vatica HealthCozeva merged with Vatica Health on October 31, 2025, and took the 2026 Best in KLAS award for Risk Adjustment: Point of Care and In-Home Health Assessments at 91.2. Its PayerOne Risk product works inside the visit and the in-home assessment, with native-FHIR certified measure engines underneath. The combined company pairs Cozeva's platform with Vatica's clinician-support model.
-
Datavant
Coding, retrieval & complianceTwo 2026 Best in KLAS wins, Risk Adjustment (Coding, Retrieval and Compliance Solutions) at 90.5 and Outsourced Coding at 94.9 against a category average of 88.4, give Datavant more independent validation than anything else listed here. Its own release cites 4.5 million charts coded a year at 98% accuracy. What it is good at is volume. Encounter workflow and per-diagnosis submission tracking are not part of the pitch, so most buyers run it next to something else rather than instead of something else.
-
ForeSee Medical
Point-of-care suspectingForeSee runs an NLP disease-detection engine inside the EHR and puts HCC suspects in front of the clinician during the visit. Its InstaVu feature links each suspect to the page of source documentation it came from, which saves a coder the hunt later. It is a focused tool rather than a platform: no chart-review volume, no audit response management.
-
Innovaccer
Enterprise platformInnovaccer treats risk adjustment as one workflow on its Galaxy payer platform, with Gravity AI agents on top and a 2026 Best in KLAS win in Data Analytics Platform for Payers at 90.5. In April 2026 it announced a move from software-access fees to outcome-based, per-task pricing and a $250M three-year commitment to the agent platform.
-
Navina
Clinician copilotNavina is the clinician-first option here. It condenses EHR, HIE, and claims data into something a physician can read before the visit, flags conditions worth confirming during it, and lets the clinician document without leaving the chart. Funding: a $55M Series C on March 25, 2025, led by Growth Equity at Goldman Sachs Alternatives. One detail that circulates wrong: its Best in KLAS award in Clinician Digital Workflow was the 2025 one. TransformativeMed took that category in 2026.
-
Pelica Risk Adjustment Copilot
Unified operations · the publisher's productPelica runs prospective and retrospective capture in real time against a single member record, and that record is the same one the quality, pharmacy, network, and care management teams work from. V28 constrained-group trumping recalculates as each condition lands, every HCC carries its evidence packet by default, and each diagnosis stays visible across the four CMS reports (277CA, MAO-004, MMR, MOR) in one place. Tradeoffs: the company was founded in 2025, holds no KLAS rating, and has a smaller install base than any other vendor here. Pelica Health publishes this directory.
-
RAAPID
Audit-transparent codingRAAPID is built around the audit trail. Neuro-symbolic AI and knowledge-graph clinical NLP produce MEAT-based support for every HCC, tracing each code back to the clinical language that justifies it, alongside autonomous retrospective review and a dedicated RADV product. Compliance officers are the buyer this design has in mind. It is a coding-and-audit tool, not an encounter workflow.
-
Reveleer
Retrieval, review & RADVReveleer built its business on high-volume retrospective chart review and has spent two years widening the footprint. Buying Curation Health in October 2024 brought point-of-care rules with Epic, Cerner, and Athenahealth integrations. An all-in-one RADV Audit SaaS shipped October 15, 2025. EVE Hybrid AI added prospective suspecting on January 20, 2026. An April 2026 Clinical Data Repository pulls member records together across risk, quality, RADV, and care management. It sells primarily to plans, so a delegated provider group typically meets Reveleer as something its payer runs rather than as a tool it owns.
How to evaluate
Five things that separate the 2026 buy from the 2024 buy.
V28 finished phasing in for payment year 2026 and cut roughly a fifth of the ICD-10 codes that used to map to an HCC. Features that were nice to have while V24 stayed forgiving are the ones now worth paying for.
Both directions, one record
Prospective work gets a valid diagnosis documented inside the service year. Retrospective work recovers what was missed and supplies the audit answer. Buy them separately and you get two systems whose member views drift. In a demo, ask to see a single member with both the prospective suspects and the retrospective findings on the same screen.
Live V28 trumping logic
A V28 constrained group pays one coefficient no matter how many of its conditions you document, so extra codes add audit exposure and no RAF. The software should recalculate trumping as conditions land, which keeps a coder's queue limited to diagnoses that can still change payment.
Evidence attached at capture time
The source note with MEAT support, the labs or imaging, the attestation, and the coder review record all belong to the HCC from the moment it is captured. An evidence packet built by hand after the RADV notice arrives is a research project on a deadline, and a diagnosis that was genuinely supported can still fail because nobody can locate the support in time.
Lifecycle visibility, 277CA to MOR
CMS decides whether a diagnosis counted, and it says so across four reports. Ask for per-diagnosis tracking through the 277CA acknowledgement, the MAO-004 eligibility report, the MMR payment report, and the MOR model output. An MAO-004 rejection is the earliest warning a team gets that documentation will not hold up.
Specific, pre-claim prompts
A prompt that says "review chronic conditions" changes nothing. A prompt that says "CKD stage 3a documented in 2024, confirm staging today" changes the note. Sit through a live encounter during the demo and judge the prompts yourself rather than the screenshots.
FAQ
Common questions.
What does risk adjustment software actually do?
It helps Medicare Advantage plans, IPAs, ACOs, and medical groups document the chronic conditions their members actually have, so CMS payment reflects the real clinical risk of the population. In practice that means suspecting likely HCC diagnoses from claims and clinical data, surfacing them to clinicians or coders for confirmation, attaching audit evidence, and tracking each submitted diagnosis until CMS accepts it. Vendors specialize. One will be strongest inside the exam room, another at reviewing thousands of charts after the fact, another at analytics, and a handful attempt the whole cycle.
Do I need prospective or retrospective risk adjustment software?
Most risk-bearing organizations need both. Prospective tools work at or before the visit so a valid diagnosis gets documented inside the service year, which matters more under V28's tighter code-to-HCC mappings. Retrospective tools review charts after the fact to recover missed diagnoses and support audits. The question worth putting to a vendor is whether both run against one member record, because two separate tools produce two versions of the same patient and no way to settle which one is right.
How did CMS-HCC V28 change what to look for in a vendor?
V28 applies in full for payment year 2026. Its crosswalk recognizes about 7,770 ICD-10 codes against roughly 9,797 under V24, the model carries 115 condition categories, and CMS actuaries put the average MA risk-score effect at about a 3.12% reduction. What that changes for a buyer is timing. A tool that reports after claims settle is reporting a loss rather than a fixable gap, so the features that now separate vendors are V28-aware suspecting before the encounter, constrained-group trumping computed live, and per-diagnosis submission tracking.
Which risk adjustment vendors won 2026 Best in KLAS?
KLAS split the category across several awards in 2026, all announced February 4, 2026. Datavant won Risk Adjustment (Coding, Retrieval and Compliance Solutions) at 90.5 and Outsourced Coding at 94.9. Cozeva, now merged with Vatica Health, won Risk Adjustment: Point of Care and In-Home Health Assessments at 91.2. Innovaccer won Data Analytics Platform for Payers at 90.5. No single vendor swept the category, which supports choosing by the specific job you need done.
What makes an HCC capture RADV-defensible?
The auditor asks for evidence and the evidence is already sitting on the code. That means the source clinical note carrying MEAT support, whatever labs or imaging back it, the provider attestation where one exists, and the record of the coder who reviewed it, all bound to the HCC itself. Trumping has to be right too, so a constrained group sends only its highest-weighted condition to the claim. The way to test a vendor is to ask whether chain-of-custody is something the system maintains continuously or something a team assembles once a RADV notice arrives.
Who publishes this directory?
Riskadjustmentsoftware.com is published by Pelica Health, Inc., the company behind the Pelica Risk Adjustment Copilot, which appears in the directory alongside nine competitors. Vendor information is drawn from public sources that are dated and linked on every profile, competitors' strengths are stated plainly, and the ownership relationship is disclosed so readers can weigh the publisher's perspective.
Ownership disclosure
This directory is published by Pelica Health, Inc., which builds one of the ten products listed. That is a conflict of interest, and the way we manage it is procedural: vendor facts come from public sources that are dated and linked, competitors' awards and strengths are reported at full strength, and our own product's weaknesses are printed alongside everyone else's. Our evaluation criteria are documented on the methodology page. If you find an error, email directory@pelica.com and we will correct it.