Photometrics AI

Tagline: Where Light Meets Intelligence
Core proposition: Precision software for smarter street and roadway lighting
Founder: Ari Isaak, GISP | CTO: Bill Dollins | Contact: ari@photometrics.ai | 858-633-6447

What It Is

Photometrics AI is a software platform that optimizes public lighting performance through networked lighting controls. It determines the optimal operating parameters for each luminaire based on real-world geography, lighting standards, and configurable priorities, then sends those parameters to existing lighting control systems via API.

Software-only. No hardware. No retrofits. No field activity.

Application Scope

Photometrics AI works wherever outdoor lighting is present: city streets, suburban neighborhoods, rural roads, highways, parks, campuses, industrial areas. It is not limited to any population density or geographic classification. Benefits apply to the entire lighting system as an interconnected whole, not individual fixtures.

How It Works (4 Components)

1. •

   Target Lighting Layer (TLL) — GIS-based maps specifying desired illumination levels for distinct zones (crosswalks: 20 lux, streets: 7-11 lux, sidewalks: 2 lux, building footprints: 0 lux). Patented: US9894736B2.

2. •

   Optimization Engine — AI-accelerated system that tests thousands of lighting configurations in minutes. Generates in-house labeled training data for deep learning models specific to each luminaire type. Patent pending: 18/660,680.

3. •

   Dynamic Scheduling — Priority-based calendar resolving which lighting instruction runs based on time, events, or real-time triggers.

4. •

   NLC Integration — API communication with lighting control platforms (integrated with 2 major systems) to apply optimized parameters.

Current Photometric Approach

Photometrics AI currently calculates horizontal illuminance (light falling on the ground plane). This aligns with IES RP-8-14 methodology and is sufficient for most optimization work.

Future capabilities (not yet implemented):

• •Vertical illuminance (light on vertical surfaces, important for facial recognition)
• •Luminance (light reflected toward observers, used in newer standards)

If a project requires vertical illuminance or luminance calculations, this becomes a development priority and can be accomplished quickly.

Practical note: For most applications, horizontal illuminance serves as a good proxy. For example, 20 lux of vertical illuminance (the typical standard for pedestrian visibility) is close to 20 lux of horizontal illuminance at the same location. Not exact, but a reasonable approximation that still delivers value.

See the outdoor-lighting-standards skill for detailed discussion of photometric principles and regional standards.

Energy Savings Mechanism

Photometrics AI achieves significant energy savings through two mechanisms:

1. •

   Precision design optimization — Per-luminaire dimming based on actual geometry, overlapping beam spreads, and Target Lighting Layers. Replaces cookie-cutter "typical layouts" that ignore real-world conditions.

2. •

   Time-of-night dimming — As vehicle miles traveled and pedestrian activity decline overnight, lights dim further while maintaining safety standards.

For specific savings percentages and financial calculations, see the photometrics-ai-financials skill.

Operating Hours Definition

Street lights operate from the end of civil dusk (beginning of nautical dusk) to the end of nautical dawn (beginning of civil dawn). This period is called "dusk-to-dawn" or approximately 4,100-4,165 hours annually (averaging 11.4 hours per night).

Priority Hierarchy (Critical)

Photometrics AI never averages or compromises between competing priorities. Strict hierarchy—higher priorities are never degraded for lower ones:

1. •Dispatch — CAD system integration (fire, crash codes) triggers 500ft illumination
2. •Demand Side Management — Grid operator emergency triggers strategic dimming
3. •Transportation Safety — Crosswalks, bike lanes, high-injury locations get +2 lux dusk-8PM
4. •Crime Prevention — High drug/vehicle theft areas get increased lighting
5. •Special Events — MLB games, Halloween (most dangerous night—treat local roads as crosswalks)
6. •Migratory Birds — BirdCast integration dims low-speed/low-crime areas 2AM-sunrise on high-migration nights
7. •Default — On at dusk, midnight dimming (sidewalks/yards to 0 lux), off at dawn

National Energy Impact

With approximately 26-60 million streetlights in the US, Photometrics AI's optimization approach can deliver grid-scale energy savings equivalent to hundreds of thousands of homes.

For specific calculations and methodology, see the photometrics-ai-financials skill.

Current Traction

Deployed lights: 0 (as of January 2025)

Pipeline:

• •8,500 lights (Tennessee) — Contract with ESCO partner; ON HIATUS while ESCO negotiates with utility. Not deployed.
• •2,000 lights (Mississippi) — Awaiting city council approval. Not deployed.

Product status:

• •Core platform complete: optimization engine, scheduling, API integrations
• •Integrated with 2 major NLC systems
• •Software tested and operational

Final development in progress:

• •Training currently runs on 50 lights (for speed) — results not yet highly accurate
• •Building TLL adjustment tool to create variable training scenarios for better distribution
• •This is the last step before full production deployment

Do not claim:

• •"X streetlights in active deployment"
• •"Running in Memphis right now"
• •Any deployed/operational claims until projects go live

Patents & IP

• •US9894736B2 (granted 2018): Target Lighting Layers
• •18/660,680 (pending, issuance expected soon): AI training data methodology
• •Continuation filed on 18/660,680 — will result in 2 patents from this application

Total: 3 patents (1 granted, 1 issuing soon, 1 continuation pending)

• •Proprietary labeled training dataset—competitors would need years to replicate

Reference Files

For detailed information, read the appropriate reference file:

• •Value Proposition — Benefits messaging by stakeholder, ROI talking points, unquantifiable benefits
• •Technical — Architecture, Target Lighting Layers, optimization engine, training pipeline
• •Competitive — Market positioning, GRADIS comparison, why NLCs complement rather than compete
• •Go-to-Market — ICP, pricing, pilot structure, sales channels, utility programs

For financial analysis: See the photometrics-ai-financials skill for ROI calculations, unit economics, value methodology, and source documentation.

Source Documents

Do not load source PDFs for routine proposals—the summary data in reference files is sufficient.