AI agents for workers' comp claims.
A licensed human at every decision.

ClaimLayer is a regulatory-aware execution layer that runs Claude-powered agents on top of your existing claims system — drafting the analysis, applying the statutory math, generating the notices, and tracking the deadlines, inside guardrails enforced in code.

6 specialized Claude agents
100% of model calls audit-logged
0 auto-deny pathways — by construction
1,242 automated tests · 83 suites
Product tour

See the whole claim lifecycle, end to end

Narrated walkthrough captured live from the running application — intake to settlement, with the AI decision audit trail in between.

Why ClaimLayer

Execution, not prediction

Most AI in claims scores risk. ClaimLayer carries out the regulatory workflow itself — and is built around the assumption that a wrong automated decision has legal consequences.

⚙️

Agents that do the work

Every incoming document ingested, summarized, and reviewed in the context of its claim; compensability analyses drafted; treatment requests evaluated against MTUS; Medicare interests screened; voice intake structured — before the adjuster opens the file.

🛡️

Guardrails in code, not prompts

The model physically cannot deny treatment, move a statutory deadline, write a reserve, or invent a statutory value. Hard limits live in the service layer, where no prompt injection can reach them.

🔌

A layer, not a replacement

Pluggable adapters sit on top of your retained system of record — ingesting claims, pushing diaries, documents, and notices back. No migration project required.

Before — react to paperwork

Medical reports pile into an inbox · work status reports wait to be read · legal documents get filed by hand · the adjuster works out what each one requires · notices and diaries typed one at a time.

After — decide, then it executes

Every document ingested, summarized, and reviewed in the context of its claim · required actions surfaced to the adjuster — including reserve adjustments · the adjuster works the queue · the AI handles everything after: decision documented, letters and notices sent, diaries completed, the next ones set, FROI/SROI state reports filed — all auditable.

The operating contract: as long as the adjuster works the generated action queue, statutory timelines are met, penalties are avoided, and the state reporting files itself. The adjuster's job becomes the two things that need a human: impactful decisions — with everything relevant presented, summarized, and organized — and taking care of people.
Platform

Everything a regulated claims operation needs

AI Agent

Compensability analysis

Score, priority, suggested reserves, red flags, and written rationale on every new claim — presented as a plain-language decision brief with the source documents and their AI summaries one click away. Accept or deny stays a licensed-human action.

AI Agent

MTUS treatment authorization

Every RFA reviewed against California treatment guidelines. The agent can only return auto_approve or physician_review — denial doesn't exist in its vocabulary.

AI Agent

Settlement pricing & MSA screening

C&R offers priced with written rationale, gated by a deterministic Medicare-interest screen that runs before the model ever sees the file — and a settlement package that refuses to generate until the MSA document is verified on the claim.

Engine

Statutory benefits engine

TD at two-thirds AWW with the 104-week cap, PD rating and advances, waiting periods, and calendar- vs business-day deadlines per code section — version-pinned to DWC sources.

Engine

Notices, diaries & automated FROI/SROI

Server-side PDF generation for statutory notices, penalty diaries that cannot be snoozed, and automated WCIS state reporting: as the claim moves — created, benefits start, rates change, benefits suspend or resume — the matching FROI/SROI EDI transactions enqueue themselves, deadline-tracked and acknowledgment-reconciled.

Trust

AI decision audit trail

Every model call logged with its input snapshot, raw output, tokens, latency, guardrails triggered, and human-override status — queryable in the in-app Agents console.

How it works

From injury to resolution

Report

The worker reports the injury by voice or guided form — bilingual, mobile-first, no login. The employer files the FROI from their own portal.

Prepare

Every incoming document — uploaded PDF, scanned fax, email attachment — is ingested, summarized, and reviewed in the context of the claim. The actions it requires, reserve adjustments included, surface on the adjuster's queue with rationale attached.

Decide

The adjuster works the queue: every consequential call, made with everything relevant presented, summarized, and organized. The AI recommends; the licensed human decides.

Execute & prove

The AI handles everything after the decision: documents it, sends the letters and notices, completes the diaries, sets the next ones, and files the FROI/SROI state reports — every step audit-logged.

Engineering decisions

Safe to point at a regulated workflow

01
No auto-deny pathway exists anywhere in the systemDenials are a licensed-human-only action by construction — the output schema doesn't include one.
02
Reserve changes require a licensed adjuster's approvalThe AI may suggest reserves; nothing is written to the financial system of record until an adjuster approves it.
03
Deterministic MSA screen gates every settlementMedicare-interest screening is never left to the model's discretion.
04
Statutory values are never model-generatedRating schedules, fee schedules, and caps come from version-controlled DWC publications.
05
Penalty diaries cannot be snoozedStatutory deadlines use the correct calendar/business-day basis per the governing code section.
06
Adversarially tested guardrailsThe test suite actively tries to push agents past their bounds — and asserts the guardrails hold.
Working with ClaimLayer

Three sides of the same claim

How each person actually uses the platform — and what changes against the claims software they're used to.

The injured worker

Reports once, in their own words

  • Opens a secure single-use link on their phone — no account, no password, no app to install
  • Describes the injury by voice or typing, in English or Spanish; the agent extracts the facts — no forms
  • Picks a network doctor ranked by distance and tier, books the appointment, confirms it in the same session
  • The DWC-1 claim form is prepared from what they said and tracked to signature
vs. the usual claims systemWorkers are usually a data subject: a paper FROI, a phone tree, then silence. Here the worker drives their own intake in one sitting — and every step they complete is one the adjuster never has to chase.
The adjuster

Works the queue — the AI handles everything after

  • Incoming documents — uploaded PDFs, scanned faxes, email attachments — are ingested, summarized, and reviewed in the context of the claim; the actions they require, reserve adjustments included, surface on the queue
  • Each item opens as a plain-language brief: what to decide, why, what in the claim led here — everything relevant presented, summarized, and organized, with the originals one click away
  • After the decision, the AI does the follow-up: documents it, sends the letters and notices, completes the diary, sets the next ones, files the FROI/SROI state reports — all of it logged and auditable
  • What's left for the human is exactly what should be: the impactful decisions, and taking care of the people on the claim
vs. the usual claims systemLegacy systems are passive ledgers: the adjuster reads raw PDFs, recalls the code section, drafts each notice, calendars deadlines, and keys state reports by hand. Here the contract is simple — work the queue and the timelines are met, the penalties avoided, the reporting filed.
The employer

Files in minutes, sees everything

  • HR files the first report of injury in the portal; payroll wage data pulls automatically and the benefit rates compute themselves
  • The worker's intake link goes out immediately — the claim is moving the same day
  • Watches claim status, deadlines, and decisions in real time instead of calling for updates
  • Pulls loss runs and OSHA-ready reporting self-serve at renewal — with reserves honestly labeled as adjuster-approved or pending
vs. the usual claims systemThe usual experience is a fax, a week of silence, and a quarterly report. Here the employer sees the same live claim the adjuster does — scoped to their own data and nothing else.

Every behavior described above is implemented and demonstrated in the interactive demo, on synthetic data.

Screenshots

The product, as it actually runs

Captured live from the application on demo data. Click any screenshot to enlarge.

Adjuster console with ranked action queue
Adjuster consoleRanked action queue with AI priority, reserves, and deadlines
Claim file with AI compensability analysis
The decision loopApprove, edit, or decline — with the full aftermath shown before you commit
TD benefit periods timeline
Benefits trackingTD periods against the 104-week statutory cap
AI decision audit trail
Agents consoleEvery model call logged — overrides, guardrails, latency
RFA treatment authorization center
RFA centerMTUS evaluation — auto-approve or physician review only
Employee intake wizard
Employee portalBilingual voice-enabled injury intake
Legacy integrations console
IntegrationsAdapters over your retained system of record
In-app architecture playbook
Architecture viewThe system documents itself — agents, guardrails, lifecycle
Company

Built by claims people, with AI

ClaimLayer was designed, architected, and built by a California workers' compensation claims professional with a decade in the field — by directing Claude Code, Anthropic's AI coding agent. The hard part was never the keystrokes; it was knowing what to build, where the model must not be trusted, and how to make a regulated workflow safe.

The result is both an active project and a public worked example of a question we think matters: how do you get real leverage from AI agents in an environment where a wrong automated decision has legal consequences?

The answer ClaimLayer takes: agents execute, humans decide, code enforces the boundary, and everything is logged. The entire implementation — prompts, guardrails, statutory math, and 969 automated tests — is open for inspection on GitHub.

DomainCalifornia workers' compensation
AIAnthropic Claude (6 agents)
StackNode / Express · React · PostgreSQL
Tests1,138 across 75 suites
Audit trailEvery model call, queryable
StatusReference implementation · demo data
Contact

See it on your own claims workflow

Tell us a little about your operation and we'll set up a walkthrough — or open an issue on GitHub if you're here for the engineering.

Submitting opens a pre-filled email — nothing is stored on this site.

🐙  GitHubSource, issues, and discussions at aksiomatixx/ClaimLayer
🖱  Interactive demoLaunch the demo in your browser — the real application on synthetic claims, no install, no keys. Or clone the repo and run supabase start + npm run dev:demo for the full stack.

Click anywhere to close