California's 483 cities are drowning in plan review backlogs
4–12 week backlogs
Multi-month delays on civil plan review block housing delivery and public infrastructure. Every week of backlog costs developers tens of thousands in carrying costs.
$150–$300/hr outsourced cost
Small cities outsource plan review at premium consulting rates — with zero institutional knowledge retained after each engagement.
Silver Tsunami
Median age of California public works PEs exceeds 50. Decades of institutional engineering knowledge is retiring with no capture mechanism.
As-built acceptance gap
Cities permanently accept non-conforming infrastructure into public ownership — creating multi-million dollar long-term liability with no automated check.
"I did not discover this problem because of AI — I discovered AI because of this problem." — Koosun Kim, PE, 23-year Public Works Director
The solution
A PE reasoning engine — not a chatbot
Civilinfra.ai reads civil engineering plans exactly as a licensed PE reads them — sheet by sheet, discipline by discipline, rule by rule — and generates structured compliance comments with dual citations traceable to both city standard numbers and master rule IDs.
📋
Permit review
Automated plan check before construction. Comments cite city standard + master rule.
🏗️
As-built acceptance
Verify completed infrastructure before city accepts permanent ownership. Unique to Civilinfra.ai.
U.S. Provisional App. #64/069,095 — Filed May 19, 2026
1
Universal Rule Mapping & Citation Engine
Maps any city's local standard numbers to 515 master rule IDs. Dual-citation output per comment.
2
Sheet-Title Routing Engine
Loads only discipline-specific rules per sheet — 90%+ token reduction, 60–90 sec review.
3
Three-Layer Standards Hierarchy Engine
Special Provisions → City Standards → CA Title 22/DPH baseline. Automated conflict resolution.
4
Master Template + City Delta Architecture
515-rule baseline; each new city maps only 10–15% delta. Onboarding: $5–10K vs. $50–100K.
5
Universal Gap Detection & Advisory Engine
Flags missing city standards; applies master values as advisory across all 17 disciplines.
6
Confidence-Gated Extraction Engine
Mathematical confidence scoring on every OCR parameter. Flags uncertain extractions for PE review.
7
3D Model Generation & As-Built Verification
2D parameters → 3D digital twins. Cross-utility spatial conflict detection and as-built deviation flagging.
Market opportunity
$2B+ national TAM — two revenue sessions per project
$75K
SJV pilot ARR per city
$12M+
California beachhead TAM
$500M+
California all-government TAM
$2B+
National TAM — 90,837 US local govts
Go-to-market sequence
3 SJV pilot cities → 40 SJV agencies ($1M ARR) → 483 CA municipalities ($12M+) → national expansion
SJV Pilots (Year 1)$75–90K ARR
SJV Region (Year 2)$1M ARR
California (Year 3–4)$12M+ ARR
National (Year 5+)$2B+ TAM
Two-sided market
Agency adoption drives firm demand. Firm adoption drives agency awareness. Same 515-rule dataset serves both sides.
Public agencies — $15–30K/yr per city
Engineering firms — $5–50K/yr per firm
Special districts — $15–25K/yr per district
Source: 2022 US Census of Governments
Competitive landscape
We occupy the regulatory compliance layer no one else has built
Capability
Civilinfra.ai
Bluebeam Max
Generic AI
Jurisdiction-specific rule mapping
✓
✗
✗
Dual-citation output (city std + master rule)
✓
✗
✗
Three-layer hierarchy resolution
✓
✗
✗
As-built acceptance verification
✓
✗
✗
3D cross-utility spatial conflict detection
✓
✗
✗
PE-validated confidence gating
✓
✗
✗
May 2026 priority patent date
✓
✗
✗
Bluebeam Max (using Anthropic Claude) launched in early 2026 — validating the market. But Bluebeam targets private contractors. We own the public agency regulatory compliance layer they cannot reach.
Traction & IP
Built in research phase — zero dilution to date
IP portfolio
Patent
7 inventions — App. #64/069,095 — May 19, 2026
Copyright
17 datasets — Cases 1-15166194041 & 1-15166194107
Trade secrets
515 rule values & city mapping tables — never published
PCT pending
International filing — 150+ countries — May 2027
Milestones achieved
✓ 515 PE-curated rules across 17 disciplines — complete
✓ Live functional demo — civilinfra.ai
✓ SAM.gov registered — UEI PWETR87JB9E1
✓ NSF SBIR Project Pitch submitted — #00113625
✓ FBN registered — Santa Clara County
✓ Mercury Bank account — active
→ StartX & Alchemist applications submitted
→ Delaware C-Corp formation — upon trigger event
ADA Title II WCAG 2.1 compliance deadline — April 26, 2027 (agencies 50K+ pop). Civilinfra.ai's structured output is inherently compliant. Bluebeam Revu is not. This creates a forced migration event directly into our beachhead market.
Business model
Annual SaaS — two sessions per project, two customer segments
Agency mode
Small city (<50K) — $15,000/yr
Mid-size city (50K–200K) — $25,000/yr
Large city (>200K) — $30K + $500–1K/set
County public works — $20–30K/yr
Special district — $15–25K/yr
Engineering firm mode
Small firm (<10 eng) — $5,000/yr
Mid firm (10–50 eng) — $10,000/yr
Large firm (>50 eng) — $15–50K/yr
1 revision cycle saved = more than annual subscription cost. One agency subscription drives 5–10 firm subscriptions in that city's orbit.
80%+
Plan check overhead reduction
2×
Revenue per project (permit + as-built)
85–90%
Rule inheritance — new city onboarding
Team
Domain authority that cannot be hired, replicated, or accelerated
KK
Koosun Kim, PE
Founder & Chief Engineer | Civil PE California #C64769 | Stanford MS CEE
✓ 23+ years California municipal public works leadership
✓ Director of Public Works — City of Mountain House (2025–present)
✓ City Engineer — City of Tracy (2022–2025)
✓ Led California's first Diverging Diamond Interchange — ASCE + APWA Outstanding Project 2021
✓ Secured $41.35M federal grant — US DOT Rural Surface Transportation Program, January 2025
✓ $100M+ in total federal grants across career
✓ AI optimization research 1997–1999 — Hanyang University — predates ChatGPT by 25 years
The 27-Year Arc: In 1999, the problem was visible. At Stanford in 2001, the technology was not ready — so the founder went inside the industry rather than pursue a PhD. 23 years building the domain expertise the technology would eventually need. In 2024, the LLM inflection point arrived. This is not a pivot to AI. It is the completion of the arc.
"I did not discover this problem because of AI — I discovered AI because of this problem. In 1999 the problem was visible. In 2001 the technology was not ready — so I went inside the industry to build what no algorithm could: 23 years of PE-validated domain expertise. In 2024 the technology finally caught up. This is the completion of a 27-year arc."
Seeking: Technical co-founder
Full-stack + computer vision + 3D rendering background. Complements PE domain authority with production engineering depth. Joining post-NSF award.
The ask
Seed round — $2M at $12–14M post-money cap
Use of funds
Technical co-founder + eng team
50%
Platform development & cloud infra
25%
PE validation & pilot city onboarding
15%
Legal, IP prosecution, operations
10%
Milestones this funding achieves
✓ 3 pilot city contracts signed (SJV beachhead)
✓ Full-stack platform production-ready
✓ 50+ PE-validated projects — moat established
✓ NSF SBIR Phase I award ($275K non-dilutive)
✓ $1M ARR — SJV regional footprint
✓ Series A ready — 483 CA city expansion
Non-dilutive pipeline
NSF SBIR Phase I ($275K) + DOT SBIR ($200K) + EPA SBIR ($200K) + EDA Tech Hub ($500K) = up to $1.175M non-dilutive alongside seed round
Contact
Koosun Kim, PE — info@civilinfra.ai — (650) 318-1844 — civilinfra.ai
Confidential — Research Phase — Patent Pending #64/069,095