For federal agencies
From a weeks-long screening memo to a determination in minutes
ClearPath NEPA is a working application that screens a proposed action against the real federal Categorical Exclusion catalog and produces a cited, defensible determination. A human reviewer commands every decision, and the whole system runs on CEQ's NEPA & Permitting Data Standard.
Solution abstract
The most common NEPA decision, finally digital-first
Categorical Exclusions are the highest-volume decision in federal NEPA, and still the most manual. ClearPath NEPA runs the whole path in minutes: intake, an AI Categorical Exclusion match with citation, a human decision, a machine-readable determination, and a live data-standard API. It screens against all 2,105 exclusions in the real federal catalog, on the federal data standard, and a human approves or overrides every AI output.
- Real federal CEs screened
- 2,105
- Per case, not weeks
- Minutes
- Human-approved decisions
- 100%
- Est. administrative effort saved
- ~80%
- Functional requirements met
- 10 of 10
The approach · What you save
Start where the volume is: Categorical Exclusions
Federal environmental review begins with a single screening question: does a proposed action qualify for a Categorical Exclusion? CEs are the most common NEPA action by far. Yet reviewers still search agency CE catalogs by hand, draft a memo, and reconstruct the administrative record after the fact. The work runs for weeks, and the result varies from one reviewer to the next.
ClearPath attacks that exact step. A proponent describes an action in plain language; ClearPath runs an AI-assisted completeness check, screens against the real CE catalog, and returns the applicable exclusions, each with its legal citation and a confidence score, alongside a lightweight extraordinary-circumstances and GIS screen. Because CEs are the highest-volume pathway, every case made faster compounds across an agency's entire portfolio.
The before and after: a weeks-long manual memo becomes a digital-first determination in minutes, with an estimated ~80% less administrative effort (modeled against a ~45-day manual baseline).
Built for the people who use it
Four real audiences, each with a purpose-built view
ClearPath is built around the people who actually do the work, and you can switch between their views live in the demonstration. The AI never decides for any of them. Every recommendation is labeled pending review until a person acts. The interface meets Section 508 and WCAG 2.1 AA: semantic HTML, keyboard navigation, visible focus, and AA contrast throughout.
Project proponents
A guided, plain-language intake. Submit now and refine later; receive plain-language completeness feedback instead of a flat rejection.
Agency reviewers
AI recommendations sit beside approve and override controls. Overriding a flag requires a written rationale saved to the record. The reviewer is always in control.
Agency leadership
Portfolio status with at-risk cases surfaced first, plus a projected completion date and days saved on every case.
The public
Consistent digital records, documented rationale, and a machine-readable administrative record make the workflow transparent and auditable.
What you save
Speed, consistency, and a record built as the work happens
ClearPath leads with an estimated ~80% reduction in administrative effort on CE screening and documentation. The Leadership dashboard reports the days saved on each case against a stated, adjustable 45-day manual baseline. The assumption sits in the open, not buried. Every action is screened against the same 2,105-CE catalog with a cited basis, which narrows the variation between reviewers. And because the determination and its administrative record are produced together, the record is litigation- and FOIA-ready the moment the determination issues, never reconstructed later.
The leadership portfolio: projected completion and days saved per case, against a stated baseline, with at-risk reviews surfaced first.
Deployable today
A working application, not a concept, mockup, or wireframe
The demonstration is the live system. It runs on real federal data (the 2,105-CE CE Explorer catalog) and on the NEPA & Permitting Data Standard, with all thirteen entities, the standard's own field names, and provenance on every record. It runs with and without a live AI model: when the model is unavailable it falls back to a deterministic result that looks identical on screen, so your reviewers’ workflow never stalls on a model outage.
ClearPath pairs the application with a concrete federal implementation path, detailed in the Architecture Brief: an AWS GovCloud landing zone, a FedRAMP Moderate baseline, NIST 800-53 control inheritance and agency ATO, Login.gov or agency SSO with PIV / CAC, a swap to a FedRAMP-authorized model endpoint, and import of agency CE catalogs through the standard's format with provenance preserved.
- Live golden path on real federal data and the data standard
- Deterministic fallback: identical UI when the model is offline
- Documented FedRAMP / GovCloud / NIST 800-53 implementation path
Works across agencies
Interoperability by construction
ClearPath is built on the data standard directly, not adapted to it after the fact. Entity and field names match the standard, and every record carries the six provenance fields that make secure transfer between agencies possible: record_owner_agency, data_source_agency, data_source_system, data_record_version, last_updated, and retrieved_timestamp. The same record reads through a documented OpenAPI endpoint, and the screening engine is configured from data: selecting another agency's CE set re-runs retrieval against that catalog with no redeploy. The same engine serves Forest Service, BLM, NPS, FWS, and others.
The same case, retrievable as data-standard JSON with provenance on every entity, through the documented OpenAPI surface.
Service delivery standards & roadmap
Two standards today, a clear path to more
ClearPath delivers against all four service delivery standards and leads with the two that matter most for this pathway:
- Workflow Automation. A process_model and decision_element engine derived from CE regulations automates the completeness check, CE matching, and the extraordinary-circumstances screen.
- Digital-First Documents. The determination is born as structured data (JSON) and rendered to PDF on demand, never scanned or retyped.
The same standards-native foundation extends naturally. Named here, not built for this CE-focused demonstration, the roadmap adds:
Environmental Assessments (EA)
Extend the same case, event store, and document engine to the EA workflow and its findings.
Environmental Impact Statements (EIS)
Scale the canonical model and administrative record to the full EIS lifecycle.
Public comment management
The comment entity is already modeled; automated compilation and analysis is a short extension (FR #8).
Application data sharing
Provenance and the OpenAPI surface already support cross-agency write-back (FR #2).
Additional permitting workflows
The standard's entities generalize beyond NEPA to other agency permitting processes.
Who stands behind it
A federal-focused firm with a path from demo to pilot
Nuvitek LLC (founded 2012, Tysons, VA; SBA 8(a); Inc. 5000) works in exactly the disciplines this work requires: case management, workflow automation, paper-to-digital transformation, and FedRAMP cloud modernization across AWS and Salesforce. For this solution, Nuvitek fields a small, senior team: full-stack and cloud / DevSecOps engineers who build inside FedRAMP and NIST 800-53 guardrails, paired with NEPA and permitting policy advisors who ground the decision logic in statute and agency practice. The team is amplified by agentic AI coding, the same modern practice the product itself embodies.
Nuvitek's FedRAMP-aligned cloud practice and active federal contract vehicles (GSA, GWACs, MACs, BPAs, and IDIQs) let a selected agency move from a live demonstration to an authorized pilot without standing up a new vendor relationship from scratch.
Why it matters
What it adds up to for your agency
Every section above answers one of the five questions an agency should ask before it buys. Nothing here exists without a reason, and every claim has a touchpoint in the running application.
- What you save
- CEs are the highest-volume NEPA pathway. Weeks become minutes, with an estimated ~80% less administrative effort (modeled against a ~45-day manual baseline) and days saved on every case.
- Built for the people who use it
- Proponents, reviewers, leadership, and the public each get a purpose-built view, and the reviewer is always in control.
- Deployable today
- A working application on real federal data and the data standard, with a deterministic fallback and a documented FedRAMP / GovCloud path.
- Works across agencies
- Built natively on the data standard, with provenance on every record, per-agency configuration, and a single documented OpenAPI contract.
- Who stands behind it
- Nuvitek, a federal-focused firm since 2012, has the disciplines and the active contract vehicles to move from demonstration to authorized pilot.
Next step
Request a pilot / Book a briefing
See the live golden path end to end, or talk through an agency pilot. Tell us a little about your review workload and we'll follow up.