--- title: CoPlaid Route Optimisation description: Constraint-based route optimisation for field service fleets — time windows, crew sizes, shift boundaries and real drive times, solved together. url: https://route.coplaid.ai/ site: CoPlaid audience: Operations teams and the implementation partners who deploy for them updated: 2026-08-29 --- # Every constraint you actually have, solved at once. A field service day is not a travelling salesman problem. It is time windows, service durations, one- and two-person crews, day and night shifts, vehicle capacity and a depot the van has to get back to. CoPlaid solves those together — over real road-network drive times — and hands your crews a route they can run. Runs from 10 to 1,000 jobs. Horizons of 1 to 14 days. Day and night shifts. ## The problem: ordering stops is the easy part Any mapping tool will sequence a list you have already decided. What makes field service scheduling hard is everything that decides which jobs belong on the list in the first place — and those constraints interact. Relax one and another breaks. CoPlaid models the whole thing as a vehicle routing problem with time windows, then searches for the assignment that minimises total travel while satisfying every constraint simultaneously: - **Time windows** — when each site will actually accept a visit, not merely when a van could get there. - **Crew size** — jobs needing two people are recognised and their duration adjusted, so estimates and assignments stay honest. - **Shift boundaries** — separate day and night windows, including shifts that run past midnight, each with its own fleet size and speed profile. - **Vehicle capacity** — a hard ceiling on minutes per route, plus break time and travel buffers, so a plan survives contact with a real day. - **Depot return** — routes begin and end where dispatch says they must, and the return leg is part of the cost, not an afterthought. - **Service frequency** — daily, weekly and several-times-a-week patterns are expanded across the horizon, each occurrence scheduled in its own right. ## The engine: what happens between upload and a route on a map An optimisation run is a pipeline, not a single call. Each stage narrows the problem so the solver spends its time where it changes the answer. ### Stage 1 — Geographic clustering Before any optimising, jobs are grouped by spatial proximity using clustering that adapts to fleet size and job density. This cuts the search space dramatically and keeps routes geographically sensible rather than mathematically clever. ### Stage 2 — Real drive times Travel between every pair of locations is calculated over the actual road network through professional routing services — not straight-line distance. Large job sets are batched, requested in parallel and reassembled into one distance matrix, with frequent pairs cached. ### Stage 3 — Capacity-aware assignment Jobs are pre-assigned to vehicles from their cluster, crew requirement, shift preference and remaining capacity. This gives the solver a feasible starting point instead of asking it to find one. ### Stage 4 — Constraint solving The core solver works the vehicle routing problem with time windows, iteratively improving assignments and minimising total travel while every constraint holds. Search strategies escape local optima, typically converging within five to ten minutes. ### Stage 5 — Load balancing A second pass redistributes jobs across vehicles towards roughly 80% utilisation, moving work only where it improves balance without breaking a time or capacity constraint. No crushed routes beside half-empty ones. ### Stage 6 — Results you can act on Sequenced routes per day, travel time and distance per route, workload and utilisation summaries, an explicit unscheduled-job report, an interactive map per route, and CSV for field deployment. ## NetSuite console If your operation runs on NetSuite, the whole route optimisation console can live there. A SuiteScript bundle renders it inside NetSuite's own UI shell — header, global search and navigation menu all stay put — so planners work without a second login or a second tab. What planners get in there: - **Route planning** — the configuration profiles that drive every run. - **Schedule** — dates and recurring cadences for optimisation runs. - **Optimize** — kick off a run and watch it progress. - **Run history** — every past run, its routes and its coverage rate. How it is kept safe: - Console keys are scoped to one product and one access mode — read-write or read-only. - A NetSuite host allowlist controls which pages may embed the console. - Sessions are short-lived and refresh by reloading the Suitelet. Install outline: generate a console key in CoPlaid (Account → API Keys → New, type Console, scope Route Optimisation), upload the SuiteScript to the File Cabinet, create the Suitelet script record, add a free-form script parameter holding the console key, then deploy it, set the audience roles and add it to the navigation menu or a dashboard portlet. About 15–20 minutes. A matching console covers Messaging, installed the same way with its own key. ## The workflow These steps run in order — each depends on the output of the one before it. 1. **Configure.** Select or create a profile: shift windows, vehicle counts per shift, break times, capacity limits, travel buffers and speed profiles. 2. **Load the work.** Upload a CSV or read jobs from the system of record — addresses, durations, crew requirements, time windows and frequencies. 3. **Schedule the run.** Choose specific dates or a recurring cadence across a horizon of one to fourteen days, and start it. 4. **Optimise in the background.** The request is acknowledged immediately and processed behind the scenes; addresses are geocoded, frequencies expanded, drive times built and routes solved. 5. **Balance and finalise.** Work is redistributed towards the utilisation target, and anything that could not be placed is reported rather than dropped. 6. **Review on a map.** Managers open each route in sequence, sanity-check it against local knowledge, and adjust where a person knows better than the model. 7. **Deploy to the field.** Approved routes are exported as CSV or handed to mobile workforce tools, so crews follow the schedule through the shift. ## Results - **50–70% less time spent planning** — hours of manual routing become minutes of run time. - **20–30% less travel time** — crews finish earlier or fit more work into the same shift. - **~80% target vehicle utilisation** — held by the balancing pass across the whole fleet. - **10–1,000 jobs per run** — from a single day to a multi-week horizon, without degrading. Ranges reflect results seen across field service deployments and depend on fleet size, job density and how routes were planned beforehand. We will size the expected gain against your own job data before you commit to anything. ## Frequently asked questions **How is this different from a mapping tool that orders stops?** Ordering stops is one constraint. A real schedule also has to respect when each site will accept a visit, how long the work takes, whether it needs one person or two, which shift it belongs to, and when the vehicle must be back. CoPlaid solves those together rather than sequencing a list you have already decided. **Do you use straight-line distance?** No. Travel times come from professional routing services over the actual road network, so estimates reflect the roads a driver will really take. Straight-line distance produces schedules that fall apart in the field. **How large a problem can it handle?** Runs range from around ten jobs to a thousand, across a single day or a multi-week horizon. Large job sets are clustered geographically and their distance requests batched and processed in parallel, so scale does not degrade the run. **What happens to jobs that cannot be scheduled?** They are reported as unscheduled with the run, rather than quietly dropped or forced into a route that breaks a constraint. A coverage rate is shown for every run so you know immediately whether the plan is complete. **Can a planner override the result?** Yes. Every route can be reviewed on an interactive map before it is released, and adjusted where local knowledge beats the model. The optimiser produces a plan for a person to approve, not an instruction. **How does this reach the crews and customers?** Routes export as CSV or pass to mobile workforce tools. Where the customer needs telling — an arrival window, a delay, a confirmation — CoPlaid Messaging sends it and brings the reply back against the job. ## Contact Partnerships: · Support: · Partner portal: Developer resources: · OpenAPI: Related: [CoPlaid platform](https://coplaid.ai/) · [CoPlaid Messaging](https://msg.coplaid.ai/)