Case Study • Predictive Pavement Lifecycle
$1.4M avoided by locking one quarter's schedule a week early
- Client
- PavingX
- Service
- AI Solutions|Data Analytics
Overview
A field report on PaveIQ, the predictive lifecycle-management layer Sphere built for PavingX — a national pavement management and paving contractor — over its 2,400+-property national-account portfolio: forecasting which lot fails when, and pricing the cost of waiting before anyone has to find out the hard way.
The Situation
PaveIQ watches 2,400+ properties across five national accounts — WarehouseCo, PharmacyCo, HomeCo, AutoCo, and HardwareCo — built on the same 52 years and 38,600 archived jobs behind PavingX's other AI systems. Portfolio-wide average PCI sits at 71.4, up 1.8 points year over year; 142 properties, 5.9% of the portfolio, are already in the poor band, and another 184 are forecast to cross into it within 18 months without intervention. This morning the model re-scored 214 properties overnight and is already sitting on $62k in avoided cost from optimized timing today alone.
Deciding when a lot needs treatment
Reactive — noticed on a site visit, or when it's already failing.
62 WarehouseCo lots due for crackseal
Scheduled whenever capacity allows, sometimes past the window.
22 licensed Alliance partner firms
Each runs its own judgment, with no visibility into who's actually using the recommendations.
Our Solution
PaveIQ forecasts remaining useful life per property, prices the cost of deferral, and surfaces the highest-value treatment decision before the window closes.
1. Degradation Forecasting
The model doesn't just score condition today — it forecasts remaining useful life, validated against 1,240 historical treatment interventions, accurate within ±6 months. Heavy-traffic lots degrade fastest: 2.6 years average RUL for AutoCo dealer and WarehouseCo warehouse lots against 5.1 years for light-retail storefronts. WarehouseCo's #418 location just dropped to a 7-month RUL on the latest heavy fuel-center truck-traffic data — the recommendation is to accelerate treatment to Q1, before the failure mode advances past a simple crackseal into something more expensive.
2. Treatment Timing Optimizer
218 lots are due for treatment this quarter — 96 crackseal, 88 sealcoat, 34 mill & overlay — and 156 are already auto-scheduled at an 84% AI-suggestion acceptance rate from branch managers. The single highest-confidence recommendation on the board: 62 WarehouseCo lots due for Q3 crackseal. Deferring them past the quarter doesn't just delay the job — it lets the failure mode cascade from a crackseal into a full mill & overlay, raising the five-year cost by $1.4M. The system's recommendation is simple and time-boxed: lock the schedule this week, while crews are already allocated.
Why This Matters to PavingX's Capital Plan
Capital gets spent before failure, not after
Treatment timing is priced against the 5-year cost of deferral, so a $1.4M cascade gets caught while it's still a decision, not a bill.
One model, five national accounts, one standard
WarehouseCo, PharmacyCo, HomeCo, AutoCo, and HardwareCo are scored on the same PCI and RUL methodology, so a renewal conversation has a defensible trend line behind it.
Heavy-traffic risk is priced differently than light retail
An AutoCo dealer lot and a PharmacyCo storefront don't degrade at the same rate, and the capital plan reflects that instead of averaging it away.
The Alliance partner network scales the model, not just the software
Licensing PaveIQ to independent regional contractors extends PavingX's predictive edge past its own 40+ branches, with adoption tracked so underperforming partners get caught early.
A multi-year capital plan that survives a budget review
92% plan adherence and a modeled $14.8M in avoided cost turn “trust us” into a number a client's facilities team can defend upstream.
Results
That's the pattern across the network: catch the lot before it fails, price the cost of waiting, and let a branch manager decide with the number in front of them instead of a gut feeling. Zoom out to the full five-year capital plan — $52M committed network-wide, $18.6M of it WarehouseCo's alone — and the model is tracking $14.8M in projected avoided cost from optimized timing versus a reactive, run-to-fail approach, at 92% plan adherence. The same predictive engine is now licensed to 22 of PavingX's 60+ independent Alliance partner firms, with adoption tracked per partner so the ones falling behind — three firms currently under 70% — get a retraining session instead of falling further out of sync.
