Causal AI TMS Case Study: 31% Margin Improvement at Colombian 3PL
See exactly how one Colombian third-party logistics provider moved from predictive to causal AI and transformed their profitability. Complete with before/after metrics and implementation timeline.
Company: Logística del Valle (pseudonym), a Medellín-based 3PL specializing in retail, FMCG, and light manufacturing.
Fleet: 187 owned vehicles + 340 contracted carriers
Challenge: Despite solid predictive AI tools, margins had stagnated at 8.4% for three years while competitors with newer technology were pulling ahead.
The Causal AI Initiative (Q3 2025 – Q2 2026)
The company partnered with a specialized causal AI logistics vendor to build causal models on top of their existing cloud TMS.
Key Discoveries That Predictive AI Missed
-
Empty Miles Root Cause: Predictive models blamed “lack of backhaul demand.” Causal analysis revealed the primary cause was rigid customer time windows that could be safely relaxed for 63% of shipments without impacting satisfaction.
-
Driver Turnover: Predictive models correlated turnover with salary. Causal models proved the dominant factor was unpredictability of shift endings. Implementing causal recommendations reduced this variability dramatically.
-
Fuel Consumption: The causal graph showed that aggressive routing algorithms were increasing idling time in traffic — a factor predictive models had completely missed.
Results After 9 Months
- Net Margin: 8.4% → 11.0% (31% relative improvement)
- Empty Miles: Reduced by 19%
- Driver Retention: Improved by 26%
- Fuel per km: Down 14%
- On-time delivery: Increased from 91.2% to 94.7%
Total annualised benefit: COP $4.87 billion
Implementation cost: COP $1.41 billion (including technology, change management, and training)
ROI: 245% in first 12 months
Implementation Timeline
- Months 1-3: Data foundation and causal discovery workshops
- Months 4-6: Model building and validation with local data
- Months 7-8: Pilot on Medellín–Bogotá corridor
- Months 9-12: Full rollout with continuous causal model retraining
Critical Success Factors
- Strong executive sponsorship from CEO and COO
- Close collaboration between operations, technology, and commercial teams
- Willingness to change customer contracts based on causal insights
- Investment in change management — particularly helping planners understand new AI recommendations
Lessons for Other Colombian Logistics Companies
Causal AI delivers the highest value when applied to complex systems with many interacting variables — exactly the reality of Colombian road freight.
The competitive advantage comes not from having the models, but from having the organizational courage to act on counter-intuitive causal recommendations.
Ready to write your own success story?
Our team offers the same causal AI diagnostic workshop used by Logística del Valle. We will identify your three highest-ROI causal use cases and provide a detailed business case within three weeks.
Apply for Causal AI Diagnostic Workshop
This case study was prepared by James Thornton, Principal Consultant at the firm’s Bogotá office. All financial figures have been validated by the client’s CFO.
