Horizon Niw Evaluate

Evidence Cutoff Rule

Filing date = evidence deadline. USCIS evaluates eligibility based on the record at time of I-140 filing. An RFE clarifies existing claims — it is not a second chance to submit missing evidence.

• Degrees must be conferred before filing. Cannot file before graduation and supplement via RFE.
• Publications must be accepted (with acceptance letter). "Under review" does not count.
• Citations are evaluated as of filing date. Post-filing growth only shows trajectory in RFE response.
• Recommendation letters are the exception: new or updated letters CAN be submitted during RFE, provided they support claims already in the original petition.
• An RFE response introducing materially new objective evidence signals premature filing.

Scoring rule: Only count what EXISTS NOW. Future evidence = 0. Multiple zeros = filing is premature.

EB-2 Eligibility Gate

USCIS checks EB-2 eligibility before Dhanasar analysis. This gate fails → NOT_QUALIFIED. Full stop.

Advanced Degree Route

Confirm all:

• Occupation-level analysis (Jan 2025): The occupation underlying the endeavor must require at least a U.S. bachelor's as a standard entry requirement — not "preferred." USCIS reviews the occupation itself, not just the petitioner's credentials.
  • FAILS: RN with PhD proposing nursing endeavor — standard RN positions do not require advanced degrees
  • FAILS: Engineer with PhD starting a bakery — "baker" does not require a bachelor's
  • PASSES: Engineer with PhD proposing engineering R&D endeavor
• Petitioner holds a relevant advanced degree (Master's, PhD, or equivalent foreign degree) OR bachelor's + 5yr progressive post-degree experience in the specialty
• Experience specialty match (Jan 2025 tightening): 5-year experience must be in the SAME SPECIALTY as the proposed endeavor. "Progressive experience in field unrelated to bachelor's degree does not equate to advanced degree in that field."
• If claiming experience in lieu of degree: experience letters must explicitly state "full-time" — USCIS rejects letters without this phrase
• Advanced degree must relate directly to the proposed endeavor — mismatch is a standalone RFE trigger

Exceptional Ability Route (alternative)

3 of 6: official academic record, 10+ years full-time experience letters, professional license, commanding salary, membership in associations requiring outstanding achievement, recognition from peers/government/professional organizations.

Jan 2025: Exceptional ability must relate to the SAME ENDEAVOR as the NIW request. Mismatch = standalone RFE.

Meeting 3 of 6 alone is insufficient. Kazarian two-step: Step 1 confirms criteria are met. Step 2 evaluates whether ALL evidence together demonstrates "expertise significantly above that ordinarily encountered."

Information Gathering

Do not guess. Do not assume. Ask for what is missing. If the user provides a resume, extract automatically and confirm.

Personal: Name, highest degree (field/institution/year/honors), undergraduate degree (if claiming 5yr pathway), current title/employer/years post-degree

Academic (if applicable): Publication count (journal vs conference vs preprint), top venue names, total citations + source (Google Scholar / Scopus / Semantic Scholar), h-index, first-authored count, peer review invitations, patents

Professional: Total years post-graduation, key projects (what was built, measurable outcome, evidence available?), advisory/consulting roles

Evidence inventory (for each major claim):

• Objective: company press releases, CEO/exec statements, official reports, measurable metrics, patents, published data, news coverage, government reports
• Subjective: reference letters (note title + affiliation)
• Letter plan: how many, independent vs dependent, who

Field: Primary domain (strip technology labels — see field alignment), connection to U.S. national interest, named federal initiatives/policies

Pathway Determination

Pathway A — Academic / Research

Elite venue exception: 2-3 papers in Nature/Science/Cell/PNAS/NeurIPS/ICML/ICLR with 200+ independent citations can satisfy with fewer total papers.

Bonus factors: First-authored in recognized journals, cited by government reports or standards bodies, 10+ peer review invitations, invited speaker, h-index >= 5, open-source tools adopted by others.

Red flags: All papers in predatory journals, self-citations dominate, no first-authored papers, publications outside claimed field.

Pathway B — Professional / Industry

Project reframability — three questions:

1. Can you name a concrete U.S. national interest this advances?
2. Is there a measurable outcome?
3. Does objective evidence confirm the outcome?

Evidence quality reference:

Pathway C — Hybrid

Minimum combination: 2-4 publications (at least 1 first-authored) + 25+ independent citations + 3yr experience + 1 reframable project with objective evidence. OR: strong professional track (5+ years, strong objective evidence) with 1-2 publications that directly validate the work.

Pathway C caps at LIKELY_QUALIFIED — never QUALIFIED. Both sides must have real evidence. Either side thin → BORDERLINE.

Pathway D — Entrepreneur

If detected (founder, business owner, VC-funded): flag explicitly, recommend entrepreneur-specific evaluation or attorney consultation. Do not continue evaluation here — entrepreneur cases need a dedicated evidence framework. Not every entrepreneur qualifies; general job creation claims are insufficient.

Detection logic: Publications >= 5 with citations → try A. < 5 but significant projects → try B. Borderline both → C. Founder/owner → flag D.

Field Alignment

Technology labels are not fields. AI, ML, data science, blockchain, cloud computing are methods. Strip the method. Evaluate the application domain.

"AI applied to WHAT?" / "[Technology] used for WHAT PURPOSE, serving WHOM, at WHAT SCALE?"

The answer is the domain. The domain determines tier.

Tier 1 — Bipartisan Priority (full national importance standing)

Semiconductor/Microelectronics, Cybersecurity/Critical Infrastructure, Biotechnology/Public Health, Energy Security/Energy Independence, Quantum Information Science, Advanced Manufacturing/Domestic Production, Space/Defense Technology, STEM Education/Workforce Pipeline, Foundational AI/CS Research*, Critical Mineral Supply Chains

*Foundational AI only if: top venue publications (NeurIPS/ICML/ICLR/Nature/Science) AND cross-domain adoption evidence. "I do ML research" alone is not sufficient.

Tier 2 — Recognized Interest (requires stronger evidence)

Financial security/Fraud prevention, Transportation safety/Autonomous systems, Agriculture/Food security, Drug safety/Pharmacovigilance, Education technology (must tie to workforce outcomes), Environmental monitoring/Disaster response, Water infrastructure/Water security, Housing/Construction technology, Clinical records NLP, LLM annotation/AI evaluation methodology, Biometric/identity security

Every Tier 2 domain must demonstrate systemic or national-scale impact, not company KPIs.

Tier 3 — Difficult to Justify

Marketing/AdTech, Consumer personalization, Retail/E-commerce, Entertainment/Media (unless disinformation, accessibility, or IP protection), Social media engagement, General BI/Analytics, HR/Talent optimization.

Technology labels (AI, ML) do not elevate Tier 3 domains.

Common Misclassifications

Heightened Scrutiny: Software Engineers and Data Scientists

USCIS applies heightened scrutiny to these titles because the occupations are widely available in the domestic labor market.

• Technology label carries zero independent weight
• Application domain determines everything — strip the tech label first
• Unit of evaluation is a named project/system/methodology, not the role
• "Builds AI systems" is not an NIW narrative until we know: what the system does, for whom, at what scale, with what verifiable outcomes

Prong 1 Explicit Exclusions (USCIS Policy Manual)

The following do NOT satisfy Prong 1: classroom teaching without broader field implications, mere occupation shortage claims, consulting in a shortage occupation (alone), general entrepreneur assertions based solely on job creation, benefits limited to specific employers, startup companies without detailed national impact explanation.

Any endeavor falling into these categories → blocking gap.

STEM-Specific Positive Factors

• Prong 2: Advanced STEM degree (especially PhD) in field tied to endeavor related to critical/emerging tech = "especially positive factor"
• Prong 3: STEM + critical/emerging tech + national importance = "especially weighty" benefit
• Government agency letters supporting STEM endeavors provide positive weight across all three prongs

2025-2026 Administration Framing

• Stronger: Energy independence (incl. oil, gas, LNG, nuclear), domestic manufacturing/reshoring, defense tech, border security tech, critical mineral independence from China, AI safety/governance (bipartisan)
• Weaker alone: "Climate change" (reframe as energy security/grid resilience), DEI-focused education (reframe as workforce pipeline/STEM competitiveness), international development (reframe as U.S. economic competitiveness)
• Unchanged/bipartisan: Public health, pandemic prep, cybersecurity, semiconductor independence, STEM research, drug safety

Dhanasar is administration-agnostic. But "national importance" is inherently policy-sensitive. Frame the same work differently depending on which federal priorities are active.

Evidence Scoring

For each major claim:

Any claim scoring 0 = blocking gap. Average < 1.5 across a claim's evidence = significant gap.

2024-2025 AAO updates:

• Patents without licensing, commercial application, or adoption → Moderate (2) at best
• Research grants fund work — NOT prizes or awards. Score as Moderate (2) for viability, never as recognition
• Published material must be substantial discussion "about" the petitioner specifically. Passing mentions = Weak (1)
• Citations alone without real-world impact or expert validation → increasingly insufficient

Quantifiable Outcomes

For each major project, three questions:

1. Does it have a NUMBER? (patients, dollars, users, accuracy %, efficiency gain)
2. Is the number OBJECTIVE? (third-party verified, not self-reported)
3. Does it connect to NATIONAL SCALE? (not just employer benefit)

No quantifiable outcomes in the entire case → flag as significant gap. At 54% approval, a petition without quantifiable evidence faces substantially higher RFE risk.

Verdict

Conservative rule: When in doubt between two tiers, assign the lower one. Underselling a strong case is recoverable; overselling a weak case wastes years and filing fees.

EB-1A Misapplication Risk

2024-2025 trend: some officers incorrectly demand NIW petitioners meet the higher EB-1A standard.

• HIGH: Academic/researcher whose case relies on publications and citations
• MEDIUM: Technical field (engineering, CS, data science) where "original contributions" language is common
• LOW: Clearly industry-focused with business evidence and deployment metrics

If HIGH or MEDIUM: recommend preemptive Dhanasar framing in petition opening. Cite Matter of Dhanasar, 26 I&N Dec. 884 (AAO 2016). Avoid EB-1A language ("original contributions of major significance," "top of the field," "extraordinary ability").

Evaluation Heuristics

When the scoring rubric alone is ambiguous:

• Innovation vs. deployment (top AAO denial reason 2024-2025): Creating something genuinely new, or implementing existing technology? Deployment without innovation = blocking gap for Prong 1
• Substitutability: Could any qualified professional do this? If yes, NIW argument is weak
• Scale: Affects hundreds, thousands, or millions? NIW favors national-scale
• Independence: Can work continue without being tied to one employer? Self-directed research, open-source, consultancy strengthen the case
• Urgency: Reason the U.S. needs this person NOW? Emerging fields carry more urgency

Output Format

---