Infrastructure for Fleet Rightsizing & Capacity Planning

Fleet Rightsizing requires current, findable documentation of fleet policies: lease vs. buy thresholds, vehicle replacement cycles, utilization benchmarks, and outsourcing decision criteria. These must be documented beyond tribal knowledge so the AI can apply consistent logic when recommending fleet size or vehicle mix. Auditors would verify that rightsizing criteria—e.g., 'retire asset when utilization falls below 65% for two consecutive quarters'—exist in an accessible wiki or policy repository, not in a fleet manager's notebook.

Capacity planning requires systematic capture of fleet utilization rates, loaded vs. empty miles, lease cost data, and seasonal demand patterns through defined workflow templates. ELD and telematics provide operational data automatically, but asset cost data and demand forecasts must be captured via structured processes—not ad-hoc spreadsheets. The system needs complete records with consistent fields (vehicle class, lane, utilization period) to generate valid rightsizing recommendations.

Fleet Rightsizing requires consistent schema across vehicle master records (class, age, capacity, cost basis), utilization records (loaded miles, idle time by lane), and demand records (volume by lane and season). All records must share defined fields so the AI can cross-reference utilization with costs and demand patterns to produce recommendations. An auditor would verify that all vehicle cost records include the same fields and that utilization data is tagged by vehicle type and geographic lane.

The AI capacity planning system must query telematics (utilization), fleet management (maintenance costs, asset age), TMS (lane demand, freight volumes), and financial systems (lease vs. purchase costs) via API. Without API access to these systems, the AI cannot assemble the multi-dimensional data required to recommend optimal fleet size. Access to most critical systems—ELD, telematics, fleet management—is achievable given their modern API ecosystems, even where dispatch software remains legacy.

Fleet capacity planning inputs—lease terms, vehicle replacement policies, cost benchmarks—are reviewed on a scheduled periodic basis (quarterly or annually) rather than event-triggered or real-time. This reflects the strategic planning cadence of rightsizing decisions, which are made seasonally rather than daily. An auditor would confirm that fleet cost and policy data is reviewed quarterly, even though operational telematics data updates continuously.

Fleet rightsizing currently relies on point-to-point integrations—telematics syncs to fleet management, TMS exports load data to spreadsheets, financial data lives in ERP. This is sufficient for periodic strategic planning cycles where data is assembled quarterly rather than in real-time. An auditor would observe that analysts manually pull cost data from ERP and combine it with telematics exports to build the AI's input dataset, with some direct system-to-system syncs for operational data.