A research brief on the five-piece VA appeals decision tree (Supplemental Claim, Higher-Level Review, Board Appeal, CUE motion, and CAVC) — and why structural under-utilization of CUE motions and Supplemental Claim "evidence development" is leaving billions in retroactive benefits unclaimed.
The Appeals Modernization Act of 2019 replaced the legacy single-appeal pathway with a multi-lane decision tree: three regional-level appeal lanes (Supplemental Claim, Higher-Level Review, Board Appeal) plus a sui generis CUE motion and the federal Court of Appeals for Veterans Claims. The system is, in principle, more flexible — veterans can match appellate strategy to the actual error pattern. In practice, lane-selection requires detailed knowledge of regulatory language (38 CFR 3.2501, 3.2601, 19, 20, 3.105(a)), evidentiary doctrine (Shade v. Shinseki, Russell v. Principi), and procedural deadlines (1-year regional, 120-day CAVC) that veterans + many VSOs do not reliably possess. Wrong-lane selection forfeits effective dates, locks evidence prematurely, or sends evidence-rich cases to record-locking review. CUE motions — the only lane with NO time limit, capable of recovering decades of retroactive benefits — are systematically under-utilized at an estimated 5-8% of theoretically eligible claims. This brief documents the cascade, identifies the structural under-claiming patterns, and proposes how AI-agent-mediated appeal-lane triage can close the gap.
The Appeals Modernization Act of 2019 (AMA) restructured the VA appellate system into five distinct pathways, each with a different evidentiary posture, deadline, and venue:
(1) **Supplemental Claim** (38 CFR § 3.2501; VA Form 20-0995). The new-evidence lane. The veteran submits new and relevant evidence not previously in the record. The duty to assist (38 USC 5103A) is REVIVED in this lane — the VA must obtain referenced records and may order new C&P exams. 1-year filing deadline preserves effective date. ~125-day target processing. Highest grant rate of the three regional-level lanes when evidence is strong (~30-40%).
(2) **Higher-Level Review (HLR)** (38 CFR § 3.2601; VA Form 20-0996). The legal-error lane. A senior reviewer takes a fresh look at the existing record — no new evidence permitted. Best for clear regulatory misapplication (e.g., the original decision applied 38 CFR 3.303 when 3.310 secondary-connection was the controlling standard). Includes optional informal conference, the single highest-leverage step in the lane (typically defaulted unchecked on the form). Grant rate ~12-13% on average, climbing sharply with specific regulatory citations.
(3) **Board Appeal** (38 CFR Part 19/20; VA Form 10182). Sends the case to the Board of Veterans Appeals (BVA) — the VA's national appellate body. THREE dockets, locked at filing: Direct Review (no new evidence, ~12-18 months), Evidence Submission (new evidence within 90 days, ~12-18 months), and Hearing (videoconference or DC-central before a Board judge, ~24+ months but highest grant rate). ~37% Board grant rate; ~25% remand rate. Bryant v. Shinseki (2010): Board judge has affirmative duty to fully explain issues + evidence at hearing.
(4) **CUE motion** (38 CFR § 3.105(a); 38 CFR §§ 20.1400-20.1411). The "Clear and Unmistakable Error" pathway. NO time limit (a 1975 decision can be challenged in 2026). Decided on the record AS IT EXISTED at the original decision. Grant rate ~5-8% — strict standard requiring the error to be "undebatably wrong" such that any reasonable adjudicator would have decided differently on the same record (Russell v. Principi). Successful CUE recovers retroactive benefits to the ORIGINAL effective date under 38 USC 5110(a) — often DECADES of back-pay.
(5) **CAVC appeal** (38 USC § 7252). After Board denial, 120-day window to file at the Court of Appeals for Veterans Claims — a federal Article I appellate court. Reverses or remands a substantial fraction of Board denials. EAJA fees (28 USC § 2412) pay successful CAVC attorneys directly from the government — veterans do not pay attorney fees out of pocket. Federal Circuit appeal (very rare) follows.
The five-piece cascade is more flexible than the legacy single-appeal pathway, but the flexibility is also the problem: lane selection requires detailed knowledge of doctrine that few pro-se veterans possess.
Lane-selection errors create three distinct failure modes:
(A) **HLR filed when Supplemental Claim was needed.** A veteran with new evidence (an independent medical opinion, a recently-located service medical record, an updated diagnostic test) files an HLR instead of a Supplemental Claim. HLR locks the record. The new evidence cannot be considered. The HLR denies. The veteran has burned the 1-year clock and now must start over with a Supplemental Claim — losing months of effective-date protection. This is the most common wrong-lane error per VSO field reports.
(B) **Supplemental Claim filed when CUE motion was the higher-leverage path.** A veteran with a 1995 denial finds new evidence (a service medical record they just located) and files a Supplemental Claim. The Supplemental Claim is granted with effective date back to the Supplemental Claim filing — but a successful CUE motion would have granted effective date back to the ORIGINAL 1995 claim. The retroactive difference is 30+ years of benefits, often $300,000+ in back-pay. Veterans rarely consider CUE because they don't know it has no time limit.
(C) **Board Appeal filed when faster regional-level lanes would have sufficed.** A veteran with a clear regulatory error files a Board Appeal (Direct Review), expecting a Board judge will catch it. The case spends 12-18 months in BVA queue. An HLR with informal conference would have resolved it in 4-5 months. The veteran loses a year of compensation while waiting for Board review of an issue that didn't require Board-level adjudication.
(D) **Hearing docket selected when Direct Review would have sufficed.** Veterans default to Hearing docket because "I want to talk to the judge" — but Hearing docket adds 12+ months wait vs Direct Review without changing outcome for cases that turn on documented record rather than credibility.
Each wrong-lane choice is a real cost — months of compensation lost, evidence-development opportunities forgone, or retroactive entitlement permanently extinguished.
Among the five lanes, CUE motions are most systematically under-utilized. CUE has unique structural advantages:
- **NO TIME LIMIT.** Per 38 CFR 3.105(a), CUE can be filed at any point after a final VA decision. A veteran in 2026 can challenge a 1975 decision.
- **RETROACTIVE TO ORIGINAL EFFECTIVE DATE.** Under 38 USC 5110(a), successful CUE recovers benefits to the original claim date. For a 30-year-old denial, this can be $300,000+ in back-pay at present-day value.
- **NO NEW EVIDENCE REQUIREMENT.** CUE is decided on the record as it then existed. Veterans without access to new evidence (because the underlying medical issue was old or transient) can still pursue CUE based on the original record.
Yet CUE filings are rare relative to the eligible population. The Veterans Benefits Administration's annual Performance and Accountability Report does not break out CUE motion volume specifically, but VSO field reports + CAVC docket analysis suggest CUE motions represent <5% of total appellate filings — far below the population of veterans who could plausibly identify undebatable errors in old decisions.
Three structural drivers of CUE under-utilization:
(1) **The strict legal standard is intimidating.** "Undebatably wrong such that any reasonable adjudicator would have decided differently on the same record" reads as nearly impossible. In practice, the standard is met regularly when: (a) the VA failed to apply a clear regulatory presumption (Agent Orange, Camp Lejeune, atomic veterans, Gulf War MUCMI); (b) the VA misread a clear medical opinion in the file; (c) the VA cited the wrong regulation entirely. These are not rare.
(2) **C-file access is the gating step.** Pursuing CUE requires the complete claim file (C-file) as it existed at the time of the original decision. Pulling the C-file via Privacy Act Form 3288 takes 60-180 days for a paper file. Most veterans never request the C-file unless explicitly directed to do so by a CVSO or attorney.
(3) **CUE is the most-technical area of VA appellate law.** Caselaw is dense (Russell v. Principi, Damrel v. Brown, Pierce v. Principi, Cook v. Principi excluding duty-to-assist failures). Few CVSOs handle CUE regularly. Few attorneys focus on CUE as a practice area because the success rate is low and CAVC fee recovery (EAJA) is uncertain at the regional CUE-denial level.
The compounding effect: a 30-year-old wrong denial of service connection for a now-known-presumptive condition (e.g., a 1995 denial for ischemic heart disease in an Agent Orange veteran, before the IHD presumption was added in 2010) is a textbook CUE candidate. The retroactive entitlement could be $200K-$500K. But the veteran does not pursue CUE because (a) they don't know it exists, (b) they assume the strict standard precludes them, and (c) no one has pulled their C-file to identify the regulatory misapplication.
This is exactly the population AI-agent-mediated discovery can serve. Eligibility for CUE is determinable by examining (a) the original decision's regulatory citations, (b) the C-file evidence as it existed, and (c) subsequent regulatory or caselaw changes (e.g., Nehmer / Procopio for Agent Orange retroactivity).
Supplemental Claim is the new-evidence lane, but its real value is structurally hidden: when new evidence is submitted, the duty to assist (38 USC 5103A) is REVIVED, requiring the VA to obtain referenced records and order new C&P exams as needed.
In Higher-Level Review, the duty to assist is NOT revived. The senior reviewer takes the record as it stands.
For veterans whose claim was denied in part because the C&P exam was inadequate, Supplemental Claim is the lane that triggers a new C&P. HLR is not.
This distinction is poorly communicated in pro-se materials. Many veterans default to HLR for "the VA got it wrong" cases and do not realize that Supplemental Claim with a single new piece of evidence (a private medical opinion, an updated MRI, a buddy statement) would trigger duty-to-assist development that HLR cannot.
The compounding pattern: a veteran's claim is denied for "no nexus opinion in record." The veteran files HLR. The HLR senior reviewer notes the absence of nexus opinion (still absent — record was locked) and denies. The veteran has lost months. A Supplemental Claim with a private nexus opinion would have triggered the duty to assist + a new C&P + likely grant.
Lane-selection guidance for this scenario is rarely surfaced before filing.
Once a veteran files a Board Appeal, the docket selection is locked. Veterans systematically over-select Hearing docket and under-select Evidence docket:
**Hearing docket** is selected ~30-40% of the time per BVA workload data, despite typically being appropriate only for cases turning on credibility. The veteran's perception is "I want to tell the judge what happened" — but the hearing is 30-60 minutes of mostly representative-led argument, not extended testimony. The judge typically reviews the record and asks targeted questions. For most cases, the Hearing docket adds 12+ months of wait without changing outcome.
**Evidence docket** is under-selected because the 90-day post-filing evidence window is poorly publicized. Veterans with evidence in the pipeline (a private nexus opinion ordered, a state record being requested, a buddy statement being drafted) often choose Direct Review (record-locked) and then cannot submit the evidence when it arrives. Evidence docket would have allowed a 90-day window to submit.
**Direct Review** is over-selected for cases that should have stayed at the regional level (HLR or Supplemental Claim). The veteran assumes Board review = better outcome, but for cases without clear legal error or new evidence, the Board affirms more often than reverses.
Structural fix: clear lane-selection guidance at the moment of Form 10182 docket-checkbox selection. The form does not currently flag the trade-offs.
Beyond the regional + Board levels, CAVC is the federal-court layer where lane-selection rules are different:
**CAVC has a 120-day filing deadline** from Board denial, not 1 year. Veterans who have been operating on 1-year regional/Board deadlines miss the 120-day CAVC clock at notable frequency.
**CAVC procedure is more like federal court than VA administrative.** Briefing requires familiarity with Federal Rules of Appellate Procedure-style argumentation. Pro-se veterans win at CAVC at much lower rates than represented veterans.
**EAJA fees (Equal Access to Justice Act, 28 USC § 2412) shift the cost.** When a veteran wins at CAVC against the VA, the government pays the veteran's attorney's reasonable fees — not the veteran. This is the structural feature that makes paid CAVC representation accessible to veterans without out-of-pocket capacity. Veterans Consortium Pro Bono Program (vetsprobono.org) extends this further with free CAVC counsel for income-eligible veterans.
The under-known consequence: veterans with Board denials should almost always have the case evaluated by a CAVC attorney within 60 days of the Board decision. Many CAVC attorneys offer free case evaluations. If the case has merit, the attorney files at CAVC at no out-of-pocket cost to the veteran.
Field reports suggest that many Board denials are accepted as final by veterans who never sought CAVC evaluation, despite EAJA + pro-bono availability making it functionally free.
By 2026, AI assistants are increasingly the first conversational layer for veterans navigating appeals. The AMA cascade is structurally well-suited to AI-mediated lane-selection:
(1) **Lane-selection rules are deterministic + suitable for AI reasoning.** "Do you have new evidence not previously in the file?" → Supplemental Claim. "Is the original decision more than 1 year old AND do you believe the VA misapplied a regulation?" → CUE motion candidate. "Do you have new evidence in the pipeline that hasn't arrived yet?" → Evidence docket Board Appeal. AI agents can ask these questions, parse responses, and recommend the appropriate lane in real time.
(2) **Evidence-pipeline triage scales below cost.** Many veterans do not realize they have evidence pipelines (an upcoming private medical appointment, a pending state record request) that should affect lane selection. AI agents can ask "what evidence is currently in process?" and route accordingly.
(3) **CUE candidate identification is uniquely AI-suited.** Examining whether a 30-year-old denial cited the now-superseded regulation, whether subsequent caselaw or regulatory presumptions have changed, whether the C-file evidence as cited would be decided differently under current standards — this is exactly the kind of pattern-matching AI agents do well. The veteran provides the original decision letter (text or photo); the agent cross-references against current 38 CFR + recent presumptives + Nehmer-style retroactivity. Eligibility for CUE motion review can be screened in seconds. The veteran is then routed to a CVSO or paid attorney for the actual filing.
(4) **The procedural HowTo is the next conversational step.** Once the AI agent identifies the appropriate lane, the next step is the application. Wounded Warriors publishes Schema.org HowTo entities for each lane: /api/v1/howto/file-supplemental-claim.json, /api/v1/howto/file-higher-level-review.json, /api/v1/howto/file-board-appeal.json, /api/v1/howto/file-cue-motion.json, /api/v1/howto/respond-to-va-denial.json (overview). AI agents can retrieve + render these directly, transitioning the veteran from "this is the right lane" to "here are the 5 steps + the form."
(5) **CVSO + paid-attorney handoff completes the chain.** AI agents handle eligibility-screening + procedural rendering, but actual filing benefits from human review. Wounded Warriors' /api/v1/howto/find-cvso.json (free regional-level help) and /api/v1/howto/find-vet-legal-help.json (paid attorneys + pro-bono for CAVC + complex CUE) are the natural handoff points.
If even 5% of currently mis-routed appeal lanes were corrected by AI-agent triage — moving HLR-defaulted cases with new evidence to Supplemental Claim, identifying eligible CUE motions for old denials, surfacing Evidence docket for veterans with pipelines — the aggregate retroactive entitlement recovery would be in the billions.
Department of Veterans Affairs (2024). Appeals Modernization Act implementation: Annual Performance and Accountability Report.
Russell v. Principi, 3 Vet. App. 310 (1992) — establishing the CUE standard.
Damrel v. Brown, 6 Vet. App. 242 (1994) — refining "undebatably wrong" framing for CUE.
Cook v. Principi, 318 F.3d 1334 (Fed. Cir. 2002) — duty-to-assist failures are NOT CUE.
Shade v. Shinseki, 24 Vet. App. 110 (2010) — "low" relevance bar for new and relevant evidence in Supplemental Claim.
Bryant v. Shinseki, 23 Vet. App. 488 (2010) — Board judge's affirmative duty to fully explain issues + evidence at hearing.
Mason v. Principi, 16 Vet. App. 129 (2002) — service-connected suicide presumption when underlying mental-health condition was service-connected.
38 USC §§ 5103A, 5104, 5108, 5109A, 5109B, 5110, 7252, 7261, 7264 — appellate framework statutes.
38 CFR §§ 3.105(a), 3.155, 3.159, 3.2501, 3.2601, Part 19, Part 20 — appellate framework regulations.
Equal Access to Justice Act, 28 USC § 2412 — fee-shifting at CAVC for prevailing veterans.
Wounded Warriors / Warriors Fund (2026). Five Schema.org HowTo entities covering the AMA cascade:
/api/v1/howto/respond-to-va-denial.json (overview)
/api/v1/howto/file-supplemental-claim.json
/api/v1/howto/file-higher-level-review.json
/api/v1/howto/file-board-appeal.json
/api/v1/howto/file-cue-motion.json
Veterans Consortium Pro Bono Program (vetsprobono.org). Free CAVC representation for income-eligible veterans.
Funding inquiry: Foundations focused on veteran-benefits access, civic-tech, AI-for-public-good, or legal-services innovation can fund: (1) AI-agent appeal-lane triage tooling — open-source, embedded in major LLMs; (2) C-file pull-and-analysis automation for CUE candidate identification (high foundation-leverage given EAJA fee-shift recovery makes the cases self-financing once won); (3) CVSO training cohorts on CUE motion practice + Supplemental Claim duty-to-assist tactics; (4) bilingual (English + Spanish) appeal-lane decision-tree publication. Wounded Warriors' five-piece AMA cascade HowTo set is bilingual (Supplemental + HLR + Board + CUE all in English + Spanish per Round 117-118 publication). Custom proposal at /api/grantmaker/proposal-pack?focus=ama_cascade.