Build the baseline. Detect waste. Reduce demand. Protect Roanoke.
The pilot does not pretend Roanoke already has perfect conservation intelligence. It starts by creating the minimum viable water-use dataset needed to make AI useful in the real world.
AI cannot optimize what a community has not measured.
Conservation begins with a trustworthy baseline.Large users are debated in gallons per day. Thousands of smaller outdoor and evaporative uses are often invisible because they are not separated by purpose at the property level.
The point is not to pit one class of water user against another. The point is to build a shared baseline so Roanoke can talk about conservation with evidence instead of assumptions.
In conservation rules, “irrigation” is more than watering a lawn. Delivery method, end use, runoff, and metering determine how a use is tracked or restricted.
These distinctions matter during drought restrictions because they separate high-loss, automated systems from low-volume or supervised watering.
Permanently installed underground pipes connected to pop-up spray heads, rotors, or impact sprinklers. Heavily regulated because clock timers can run without human attention.
Portable oscillating, rotary, or impact sprinklers attached to a standard garden hose. In enforcement, this is often treated like an automatic system.
Low-volume tubes, emitters, bubblers, or drip lines that apply water to the root zone. Often treated more favorably because wind and evaporation losses are lower.
A handheld hose with a positive shut-off nozzle, or a watering can. Many drought policies separate hand-watering from unsupervised irrigation.
The pilot should not simply ask “did irrigation happen?” It should separate turf, plantings, agriculture, and other consumptive uses.
Mowed lawns are usually discretionary outdoor demand and are commonly first in line for restriction during shortage stages.
Trees, shrubs, flowerbeds, and groundcovers have different community value and replacement cost than turf.
Commercial crops, pasture, orchards, and farm systems are governed differently from residential watering and should not be collapsed into lawn use.
The first phase is deliberately practical: collect clean inputs, classify end uses, normalize them, detect outliers, and recommend the highest-confidence fixes first.
Meter reads, bills, rainfall, property details, and leak-check results.
Understand normal use by property type, season, occupancy, and outdoor watering behavior.
Spot likely leaks, irrigation waste after rainfall, unusually high use, and abnormal night flow.
Prioritize the highest-impact fixes first: toilet leak checks, irrigation skips, and targeted repairs.
Measure before/after change and convert savings into public-facing conservation progress.
This concept preview uses demo data only. It is designed to show the type of public reporting the pilot can produce once local measurements exist.
Data shown is illustrative for concept demonstration. Not real operational data.
The first deliverable is not a grand claim. It is a working baseline, a ranked action list, and a repeatable conservation data model.
Map what exists, what is missing, and which properties can participate.
Enroll homes, businesses, and community sites with clear privacy expectations.
Collect four weeks of reads, rainfall, property profiles, and leak checks.
Send targeted recommendations and measure before/after changes.
The zip includes vector logo assets, optimized real-photo backgrounds, a social/poster graphic, Open Graph artwork, and the deployable static page.
Use the SVG logo for production layouts. Use the PNG poster and OG image for sharing. The site uses the real Roanoke photo as a CSS background with overlays, not a generated city scene.
Conservation is a public-facing Clearframe initiative to show what AI can do outside of a chatbot: create evidence, rank action, and make infrastructure decisions clearer.
