Route planning capabilities
Route optimization decides how work is assigned and in what order it is visited. Flowtlas's planning engine combines those decisions with the durations, working periods and resource rules supplied for the operation. On the Pro and Enterprise plans, an AI planning agent guides the search, while the planning engine keeps enforcing every rule. Contact us to confirm generation availability and integrations for your workspace.
The operational problem
The nearest available van may lack the equipment for a job. A customer may open after the team's first work period. Several nearby visits may take longer to complete than a full shift. The planning model needs these details as well as coordinates.
Product capability
Provide the task locations and duration at each stop, any visit windows, the available tours, and relevant skills or capacity. Identify work that is required and decide whether the request may return optional work unassigned. Use consistent units and realistic service durations.
The planning engine searches for eligible assignments and stop sequences using road distances, time windows and supplied resource rules. It can repair unassigned work and try moves between tours within a bounded search budget. A complete result means the active work was covered, not that the plan is mathematically optimal. For supported static service visits, optional workload balancing runs after a complete plan is found. It compares workloads against available time while retaining a fixed cap on extra distance.
Check the tasks covered, their assigned tours, stop order, arrival/departure estimates, driving and service totals, and any unassigned work. Confirm that the inputs reflect the operation; a planning engine cannot correct missing business requirements simply by optimizing harder.
Practical workflow
Supply locations, service durations, visit windows and resource requirements.
Specify vehicles, assigned people, start locations and dated working periods.
Include consistent capacity units, eligibility and stop order where needed.
When generation is enabled, inspect assignments and estimates, including any work left unassigned.
Operational benefits
In practice
On the Pro and Enterprise plans, an AI planning agent works alongside the planning engine during each run. At fixed checkpoints, after the problem is prepared and after assignment, routing and repair, it reviews a bounded summary of the plan: task locations and restrictions, tour workloads, route metrics, bottlenecks and the reasons work is still unassigned.
From that review it suggests which tours suit which work, starting assignments and stop orders for the search, and specific work worth another repair attempt. The planning engine weighs these suggestions only among assignments that are already allowed.
The planning engine stays in charge. The AI cannot add work, change vehicles or people, relax time windows, capacity, skills or commitments, or edit the returned routes. Its share of the search time is capped, and if it fails or does not respond in time, planning continues without it. Each run records which suggestions were accepted or rejected.
Questions
Route planning describes the work, resources and rules that define a usable route. Optimization searches for assignments and visit sequences within those rules. Accurate inputs matter in both steps; a shorter journey is not useful if a required job cannot be completed.
Work may lack an eligible tour or fail to fit capacity, time or route constraints. When the request permits partial work, the result can report unassigned tasks and reason categories. Review those reasons together with the task and resource data.
No. It means the bounded search did not find a solution. Review the inputs, constraints and search budget. This differs from an infeasible result based on an established hard contradiction.
Yes, on the Pro and Enterprise plans. An AI planning agent reviews each run at fixed checkpoints and suggests assignments, starting routes and repairs. The planning engine applies only suggestions that keep every hard rule, so the result still respects time windows, capacity, skills and commitments.
No. Complete describes coverage of active tasks in the returned result. It does not prove global optimality. Compare the result using your actual constraints and a reproducible baseline before making any efficiency claim.
Work with operational context
