For risk adjustment and HCC coders
HCC coding software that checks the support before the auditor does
Describe the note. Claire maps the diagnoses to ICD-10-CM codes and HCCs, walks through the MEAT evidence for each, and points out the ones that would not hold up.
Claire IT AI is HCC coding software for coders and audit leads who have to show that every HCC is supported in the note: the HCC for each code comes straight from the CMS mapping tables, never from the model’s memory, and the reasoning about support is laid out for you to check.
- No PHI needed
- CMS-HCC V24 and V28
- No card to start
Built by credentialed coders: meet the team
HCCs from the CMS tables
CMS-HCC V22, V24 and V28, ESRD, RxHCC and HHS-HCC, payment years 2019 to 2027.
The support, laid out
Claire walks through Monitor, Evaluate, Assess and Treat for each diagnosis you describe.
The right payment year
The model year follows the date of service, so a 2026 visit maps with the 2027 payment year.
Try a case · first answer free
Check a note for MEAT, free
Pick a sample or describe a note without identifiers. Claire lists the codes, maps each to its HCC from the CMS tables and explains what in the note supports it.
- The same answer pipeline as the full app: retrieval, the answer, then a check of the codes against the official code set.
- No sign-up for the first answer. Your question isn’t saved to an account or our database.
- No PHI: leave out names, dates of birth and record numbers.
Pick a sample case
Describe the scenario only. No names, dates of birth or record numbers.
0/1500Claire answers in the panel at the bottom right of the page.
What gets in the way
The questions behind every risk adjustment review
In risk adjustment the code is only half the work. The other half is proving the condition was managed at the visit.
“Is this HCC actually supported?”
Claire walks through which MEAT element the note meets (Monitor, Evaluate, Assess or Address, Treat) and points to the documentation it relies on, so you can confirm it in the record.
“Which codes would fail a RADV review?”
Ask Claire to review the conditions in a note and it points out the ones with no supporting evidence, such as a diagnosis that appears only on a problem list or medication list.
“Which model and year does this map to?”
The HCC is read from the CMS mapping tables for the model and payment year that match the date of service, and shown in a table built from those tables, so a category is never guessed.
Proof of the path
Type 2 diabetes with CKD stage 3a: codes, HCCs and MEAT
A prospective or retrospective review works the same way: the note, the codes, the support.
The case
The visit note documents type 2 diabetes with diabetic chronic kidney disease, stage 3a. The provider reviews the A1c and eGFR and continues metformin.
ICD-10-CM Official Guidelines, I.A.15 “With”
The classification presumes a causal relationship between two conditions linked by the term “with” in the Alphabetic Index or Tabular List, unless the documentation clearly states the conditions are unrelated.
- 1
Code the linked conditions
The classification links diabetes and CKD through “with”, so the combination code applies, plus a code for the CKD stage as the Tabular instructs.
E11.22N18.31 - 2
Map each code to its HCC
For a 2026 date of service, from the CMS-HCC V28 tables for payment year 2027: E11.22 to HCC 37 (Diabetes with Chronic Complications) and N18.31 to HCC 329 (Chronic Kidney Disease, Moderate, Stage 3 except 3B).
- 3
Confirm MEAT for each
Monitored and evaluated through the A1c and eGFR review; treated by continuing metformin. Long-term use of oral hypoglycemics is coded too.
Z79.84 - Result: E11.22 (HCC 37) and N18.31 (HCC 329), both supported, with Z79.84.
How it works
Three steps from note to defensible HCCs
- Step 1
Describe the case
Type or paste the clinical scenario, without identifiers.
For you: “Type 2 DM with CKD 3a; metformin continued; A1c and eGFR reviewed.”
- Step 2
Claire shows codes and the path
Codes, sequencing and the guidelines behind them, with the codes checked against the official code set.
For you: E11.22 and N18.31, HCC 37 and 329, with the MEAT evidence for each.
- Step 3
You decide and copy
Check the reasoning, make the call and copy the codes into your workflow.
For you: You confirm the support in the record and document your decision.
Compared
How Claire compares with the ways HCC review is done today
Claire sits between a paper checklist and an enterprise platform: fast to start, with the reasoning shown.
| Feature | Claire IT AI | Manual MEAT checklists | Enterprise risk adjustment platforms | General AI chatbot |
|---|---|---|---|---|
| Shows the reasoning for each condition | ||||
| HCC from the CMS mapping tables | ||||
| Works without EHR integration | ||||
| Set up in minutes | ||||
| Free plan to start |
A general comparison of tool types, not of any one product; features vary by vendor. Claire does not calculate RAF scores.
Trust
Built to support your judgment, not replace it
Claire IT AI shows its reasoning so you can check it, and it never needs patient identifiers to help.
No PHI, by design
Describe the clinical scenario without names, dates of birth or record numbers. Our terms ask that no PHI is ever entered.
You make the call
Claire explains and suggests. The final code is always the coder’s decision.
Codes checked
Claire checks the codes in its answers against the official code set, and corrects or flags a code that does not exist or does not fit.
Current code sets
Updated to the FY2027 ICD-10-CM and ICD-10-PCS code sets and the October 2026 HCPCS quarter.
The question we hear most
“Can an AI judge clinical support?”
It should not, and Claire does not. It lays out the evidence it found and the rule it applied; the coder checks the record, makes the call and documents it. The HCC itself is never the model’s opinion: it is read from the CMS mapping tables.
Plans for risk adjustment coders
Individual coders can start on the Free plan or Pro. Audit leads and plans with a team of coders: book a demo and we will walk through a rollout.
What's included:
Basic code lookups
Limited chat queries
Query Generator Tool (Beta)
File uploads
Claire with CPT Access
Priority Support
Code Libraries
ICD-10-CM
HCPCS
ICD-10-PCS
CPT
What's included:
Pro code lookups
Limited chat queries
Query Generator Tool (Beta)
File uploads
Claire with CPT Access
Priority support
Code Libraries
ICD-10-CM
HCPCS
ICD-10-PCS
CPT
What's included:
Paste clinical notes for coding
File uploads (text, PDFs)
Search and save favorite codes
Priority support
Claire with CPT Access
Code Libraries
ICD-10-CM
HCPCS
ICD-10-PCS
CPT
What's included:
Everything in Premium
Unlimited tokens
Early access to new features
Priority support
Claire with CPT Access
Code Libraries
ICD-10-CM
HCPCS
ICD-10-PCS
CPT
FAQ
Questions coders ask
Short, straight answers. Can't find yours? Ask Claire a case above.
What is MEAT in HCC coding?
Monitor, Evaluate, Assess or Address, Treat: the evidence auditors look for that a condition was managed at the visit. A condition with none of them documented is hard to defend.
AI HCC gaps in clinical documentationWhich HCC model does Claire use?
It maps from the CMS tables for CMS-HCC V22, V24 and V28, ESRD V21 and V24, RxHCC V05 and V08, and HHS-HCC, for payment years 2019 to 2027, choosing the payment year from the date of service.
Can Claire calculate a RAF score?
No. Claire maps codes to HCCs and explains the hierarchy, but it does not hold the coefficients or calculate a risk score.
HCC coding and accurate RAF scoresDoes it support prospective and retrospective reviews?
Yes. Describe the note in either case: before the visit to see which conditions need addressing, or after it to check which codes the documentation supports.
Is a medication list enough to support an HCC?
Usually not on its own. A medication implies a condition but does not show it was monitored, evaluated, assessed or treated at the visit.
Can it replace our coding audits?
No. It speeds up review and lays out the reasoning, while auditors keep the final decision.
AI-powered medical coding auditsHow do we roll it out to a team?
Each coder has their own account today. Book a demo and we will walk through a rollout for your team.
Claire for coding teamsOfficially Licensed & Partnered With Industry Leaders
Claire is officially licensed and partnered with leading medical authorities, ensuring trusted standards in coding and CDI.
Supporting compliant, audit-ready healthcare workflows.
Keep reading
From the Claire blog
- AISep 24, 2026
AI HCC Gaps in Clinical Documentation: 2026 Guide
How NLP and large language models surface conditions the record supports but the claim never carried: FHIR ingestion, extraction split from HCC mapping, V28 trumping logic, and why MEAT validation and the compliant query stay with the coder.
Read article
Medical CodingNov 2, 2026Most commonly missed HCC codes: a CDI checklist
Discover the most commonly missed HCC codes, why they slip, the MEAT and ICD-10 specificity to capture them, and RADV audit risk.
Read article
Medical CodingJul 29, 2026Claire AI for HCC Coding and Risk Adjustment: Capturing Accurate RAF Scores
Claire AI helps risk adjustment teams capture accurate RAF scores—identifying all qualifying HCC conditions, validating documentation specificity for RADV audits, and generating compliant queries while avoiding overcoding.
Read article
Medical CodingJun 9, 2026AI Medical Coding Audits: 100% Review vs Random Sampling
AI-powered coding audits review 100% of encounters in real-time, catch 85-95% of errors before submission, and reduce error rates 40-60% — moving audit teams from random sampling to targeted investigation.
Read article- ComplianceApr 28, 2026
Medical Coding Compliance in the AI Era: Navigating Regulations and Risk
Medical coding compliance in the AI era requires maintaining human accountability while leveraging artificial intelligence for accuracy improvements.
Read article
Know which HCCs will hold up before anyone asks.
Check a note free, or book a demo to plan a rollout for your review team.
The AI Medical Coding Assistant,
Built for Real-World Clinical Workflows
© 2026 CLAIRE IT AI. All rights reserved.