AI could take on much of a logistics workflow, but exceptions and oversight still need people. Lean Solutions Group CTO Alfonso Quijano’s advice is to map the work first, then choose how much autonomy each task deserves.
Watch Desk analysis
What happened
In a FreightWaves Sponsored Insights article published on 27 October, Quijano argues that AI is best suited to repetitive work, while staff handle exceptions, supply judgement and help improve the system. He points to document-heavy tasks such as invoice intake, checking discrepancies against bills of lading and routing emails.
Quijano says AI typically handles 70 to 80 per cent of a process in Lean’s experience, leaving 20 to 30 per cent for people to manage as exceptions. He also cites 65 to 70 hours returned to operators each month in document-based automation, while noting that some of that time goes back into training and feedback.
Why it matters
That is a more useful starting point than asking whether a company should “adopt AI” as if every task needs the same answer. Mapping where exceptions occur can help organisations decide which work suits straightforward automation, where an agent might be appropriate, and where a person should make the call.
The figures are Quijano’s account of Lean’s experience, not independently measured results for every logistics operation. Still, the distinction between automating routine work and keeping people responsible for unusual cases gives operators a practical framework to test.
Our read
The sensible lesson is not to bolt an agent onto every process and hope the paperwork sorts itself out. Start by mapping the workflow, identify where errors or exceptions matter, and pick the simplest tool that handles the routine work. Keep trained people in the loop where judgement changes the outcome.
What to watch
- Whether logistics firms publish results showing how automation affects accuracy and workload.
- How often human exception decisions lead to changes in operating procedures.
- Whether companies match the level of AI autonomy to the actual complexity of each task.
Discussion spark: Should companies design AI workflows around human handling of exceptions from the outset, or try to automate more first and add oversight where it proves necessary?
Sources and evidence
- Source update (27 October 2026, 16:13 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.