← BACK TO HOME

AI Product Strategy

Freight Spend Optimization

AI and machine-learning optimization that balances lowest-cost outcomes with real-world business rules.

Status: In DeliveryTimeline: 12+ monthsRole: Product LeadershipTeam: Cross-functional

Deliver an intelligent optimization workflow that combines machine-learning-assisted freight decisioning with configurable rule-based constraints, so teams can reduce spend without sacrificing execution requirements.

Freight Spend Optimization introduced an AI/ML-powered constraint builder that helps shippers move beyond pure lowest-cost awards. Teams can define rule-based goals, lane filters, and participant filters, then run optimization scenarios that balance cost, service, and operational realities.

Customers were relying on manual bid analysis and static spreadsheets. Pure lowest-cost allocation often ignored important constraints (like carrier strategy, incumbent rules, and lane-level requirements), which created inconsistent award outcomes and missed savings.

My Contribution

  • •Defined product strategy for AI/ML-assisted freight optimization
  • •Led cross-functional design of the constraint builder and scenario workflow
  • •Translated customer procurement rules into scalable product primitives (goal, scope, lane filters, participant filters)
  • •Aligned product, operations, and commercial teams on rollout and adoption

Process

Decision Model Framing

Mapped where lowest-cost optimization worked and where business constraints required rule-based overrides.

Constraint System Design

Built a structured model around goals, lane scope, participant scope, and filter logic to support real-world award strategies.

Scenario Validation

Used iterative testing to catch infeasible scenarios early and improve recommendation trust with customer teams.

Key Features

AI/ML Optimization Engine

Runs cost-aware award recommendations across lanes while accounting for capacity and bid behavior.

Rule-Based Constraint Builder

Lets users enforce business requirements even when they differ from pure lowest-cost outcomes.

Scope + Filter Controls

Supports lane and participant targeting as a whole, individually, or by grouped segments for precise optimization.

Challenges & Solutions

Challenge

Customers needed to optimize for cost while still honoring nuanced procurement and network constraints.

Solution

Combined machine-learning optimization with configurable rule-based goals and fallback logic.

Outcome

Enabled practical, explainable award decisions instead of one-dimensional cost-only outputs.

Results

  • •Improved confidence in freight award decisions by making optimization logic transparent and configurable.
  • •Unlocked significant spend-reduction opportunities while preserving carrier and lane strategy requirements.

Frontend

Next.jsReact

Backend

Optimization servicesML-assisted decisioningScenario evaluation pipelines

Tooling

FigmaAnalytics dashboardsCustomer pilot frameworks

Another Project

If you liked this case study, check out this related project next.

Product LeadershipProcureOS Pro→