Logtech Essentials: From Procurement to Payment
Special Report vendor, said during the panel.“ But AI agents have the ability to do those things.”
The consensus in that discussion was that agentic AI will allow users to build ad hoc TMS-like systems, leapfrogging legacy software. AI agents could even interact with and augment legacy systems without replacing them.
“ We don’ t want to go in and uproot a broker’ s TMS or whatever they’ ve built over the years,” Stewart said.“ Our approach is to sit at the top of that layer for you, and then we can automate as much as you allow us to automate.” Trouble communicating may be on the TMS side.“ You have these awesome old-school TMSs that are very helpful to multi-billion-dollar companies, but they just can’ t communicate with LLMs( language learning models) used in AI,” Nathan said.“ The TMSs are going to need to figure out how to connect with LLMs and then help their customers take action.”
Reengineering processes
The thought is that in the not-so-long run investment in agentic AI will compel companies to reengineer processes rather than simply build a replacement for existing processes— a technological band-aid. But don’ t expect an overnight transformation.
“ Getting a demo [ product ] is very, very easy, but getting to reliability is another order of magnitude,” Buckingham said.
AI agents could interact with and augment legacy systems without replacing them. Shutterstock. com
Large freight brokers are already well underway on their AI journey.
C. H. Robinson Worldwide, for example, the largest US freight broker with $ 11.6 billion in North American surface transportation revenue, introduced its Agentic Supply Chain Solutions platform last year as part of its Lean AI operating model.
Agents are popping up in virtually every software platform with which shippers interact.
C. H. Robinson has said AI agents help unlock hidden value in unstructured data such as emails, texts and phone calls. That data used to be invisible to the 3PL’ s systems. AI identifies optimal appointment times based on transit-time data from the millions of shipments across 300,000 shipping lanes C. H. Robinson handles each year.
As AI develops quickly and agents become embedded in more systems, smaller brokers are learning to implement technology to better replicate the results larger 3PLs are gaining at greater scale.
“ There’ s some segment of their business that they probably carve out to say,‘ Okay, this might be a case for a managed transportation service,’” Stewart said. As smaller brokers adopt agentic AI,“ I think they start rethinking even their internal strategy to‘ maybe we can actually go do this.’”
email: bill. cassidy @ spglobal. com www. joc. com November 2, 2026 | Journal of Commerce 37