Student transportation intelligence · Edge AI
HopSkipDrive's RouteWise AI™ — the world's first student transportation intelligence platform — runs multi-LLM scenario modeling at scale for school districts like Denver Public Schools, Littleton, Colorado Springs D11, Ohio County. Your DNS shows OpenAI, Anthropic, and Perplexity all in production. Cloudflare's developer platform is built for exactly this shape of workload.
RouteWise AI proven outcomes
Increased efficiency
High-utilization bus routes (30+ students)
Capital savings
Identified savings over 10 years
On-time arrival
From 85% to 99%
Source: HopSkipDrive RouteWise AI public case studies — Denver Public Schools, Littleton, Colorado Springs D11, Ohio County
RouteWise AI workflow
Each layer of the RouteWise AI engine has a clean Cloudflare equivalent. The result is faster scenarios, lower inference costs, and per-district isolation without managing per-tenant infrastructure.
Your DNS shows OpenAI, Anthropic, AND Perplexity. RouteWise AI spans all three. Each scenario today hits a provider directly — no caching, no observability layer.
CF mapping
AI Gateway + Workers AI
One observability + caching + rate-limit layer across all three providers. Identical scenario prompts cache. 30-50% inference cost cut without code changes.
"This would have taken our team nights and weekends" — RouteWise AI runs hundreds of scenarios fast. Today: single-region AWS. Workers: 330+ cities.
CF mapping
Workers + Hyperdrive
Edge compute next to the requesting district. Hyperdrive accelerates AWS RDS round-trips. 5ms cold starts vs. AWS Lambda ~400ms.
10,000+ schools, many districts. DPS ≠ Littleton ≠ Colorado Springs. Each needs its own scenario state, branding, policy rules.
CF mapping
Workers for Platforms
Dispatch namespaces per district — isolated edge compute without you running per-tenant infrastructure. Replaces the shared-AWS multi-tenancy pattern.
RouteWise AI learns from millions of miles across 10,000+ schools. Cross-district pattern matching is a vector similarity problem.
CF mapping
R2 + Vectorize
R2 stores scenario corpus + training data with zero egress. Vectorize indexes route embeddings — "find districts with similar bell-time problems" runs in <50ms at edge.
AI Gateway cost model
RouteWise AI's batch scenario modeling involves repeated prompts with overlapping content. AI Gateway caches identical/near-identical requests across providers.
Your inputs
Monthly figures, RouteWise-AI-scale defaults.
Scenario modeling, route optimization, district Q&A.
Includes context, prompt, response.
Blended across OpenAI / Anthropic / Perplexity.
Typical: 30-50% in batch-scenario workloads.
Monthly
≈ $192,000/year
Monthly
≈ $124,800/year
Estimated annual inference savings
35% lower LLM inference cost from caching alone
Excludes the additional value from observability (debugging multi-provider workflows), rate-limit enforcement (cost ceiling per district), and model-portability (switching providers without code changes).
End-to-end mapping
App Sec is already live. The expansion layers everything else on top.
AI Gateway + Workers AI
OpenAI + Anthropic + Perplexity all in production. One observability + caching + rate-limit layer. Model-portability without code changes.
Net-new — biggest play
Workers for Platforms
Dispatch namespaces give DPS, Littleton, Colorado Springs D11, Ohio County, etc. each their own isolated edge runtime. 5ms cold starts.
Replaces shared-AWS pattern
R2
Scenario inputs, AI outputs, route data, ride logs, training corpus. S3-compatible, zero egress. Drop-in for boto3.
Replaces AWS S3 + Netlify Large Media
Vectorize
"Find districts with similar bell-time problems" / cross-district routing pattern matching. <50ms similarity queries at the edge.
Augments: existing AI
Hyperdrive + D1
Hyperdrive pools + caches AWS RDS for edge Workers. D1 for per-district config + scenario state at edge.
Augments: AWS RDS
Pages
www.hopskipdrive.com is on Netlify today (hsd-website.netlify.app). Pages = git-driven deploy, branch previews, integrated with Workers / R2 / AI Gateway.
Replaces Netlify
Current — live
Already in production. WAF + Bot Management protect the api / drive / admin / rideiq endpoints.
Renewal 12/17/2026
Cloudflare One
CareDriver / parent / school admin identity. Access + ZTNA + CASB across Salesforce, HubSpot, Zendesk, Atlassian, Figma, Greenhouse, Workspace.
Net-new at renewal
Data Localization Suite
Paubox in your TXT = you handle PHI (IEP, foster care, McKinney-Vento). DLS + Regional Services pin sensitive data by region/state without separate deploys.
Augments App Sec
Business case
Inference savings, edge performance, and a 6-11× expansion path.
Inference Cost Cut
AI Gateway caching across OpenAI + Anthropic + Perplexity. Batch scenario workflows are caching-friendly by nature.
Edge Compute Cold Start
Workers vs. AWS Lambda ~400ms. Scenario modeling that "took nights and weekends" runs in seconds.
ARR Expansion at Renewal
From App Sec ($80K) → App Sec + Dev Platform + CF One ($480-880K) at the December 17, 2026 renewal.
Next step
60-90 min architecture review with your engineering + product team. We'll plug RouteWise AI's actual scenario volume into the calculator.