← Back to the library
FM Logistic uses AI to improve warehouse picking-route algorithms
Added 2026-10-14 · Reviewed 2026-10-15 · FM Logistic
Read the source ↓What the company did
FM Logistic reports deploying an AlphaEvolve-developed picking-route algorithm in a Polish warehouse with more than 17,700 picking locations. AI generated and evaluated thousands of variants against live operational data, travel distance, productivity and process constraints. The output is an auditable algorithm rather than AI independently managing the warehouse.
How it went
The company reports efficiency improving by over 10% against an already optimized process and annual travel reduced by more than 15,000 kilometres, without added infrastructure or forklifts. These are company-reported gains; the article does not disclose an independent evaluation or observation period.
Worth knowing
- Wider facility rollout was underway; broader transport applications are not established by this picking example.
- For your desk: Check order completeness, capacity and handling restrictions alongside route distance. Retain the previous algorithm as the comparison when reviewing travel and fulfillment exceptions.
Sources & provenance
FM Logistic in Poland becomes the world’s first to power logistics with Google AlphaEvolve ↗
FM Logistic · Published 2026-06-10 · Reviewed 2026-10-15
“Google AlphaEvolve generated and tested thousands of variants of logistics algorithms”
Supporting evidence and review
Human review required
- Independent review advice: verify capacity, order completeness and process constraints before releasing algorithm changes.
- Monitor travel and fulfillment exceptions against the previous operating baseline.
Evidence & limitations
- Company/provider-reported deployment; no independent performance evaluation by Freight Fieldnotes.
- AI helped develop the algorithm; it does not independently manage the warehouse.
- Reported efficiency gains are company claims; broader transport applications remain exploratory.