THE PROBLEM

The buyer carries the integration & the provider competes all or nothing

Freight is still bought on the carrier’s terms. Whatever the channel, the carrier sets it and the buyer conforms. A small shipper types the shipment into a carrier’s portal or app. A large shipper does the same through EDI and APIs, mapping into each carrier’s schema. Either way the buyer sends what the carrier asks, in the carrier’s format, once for every carrier it wants to compare, then lines up the quotes by hand and re-keys the booking into the winner’s system. All of the adapting sits on the buyer’s side.

The supply side is the mirror image. Carriers compete all or nothing. A provider wins the whole shipment or earns nothing, even on the legs where it was the strongest option. And the clean end-to-end win is getting rarer as supply chains spread across more providers, so the inland, terminal, and visibility revenue a losing carrier could have supplied goes uncaptured.

A different arrangement is now in reach. The buyer states intent once, its agent calls every carrier’s quotation agent in parallel, and ranked offers come back ready to book with no re-keying. Control moves off the carrier’s channel and onto the buyer’s agent, and competition shifts from owning the channel toward agent quality and execution reliability.

The inversion. Today the buyer adapts to each carrier’s channel, whether portal, app, EDI, or API. Tomorrow the carriers’ agents adapt to one buyer intent.

WHY NOW

Why the shift is happening now

The idea itself is not new, but the market around it has changed. As supply chains become more fragmented, the traditional all-or-nothing carrier model is already starting to weaken, and buyers increasingly expect outcomes to be assembled from the providers best suited to each leg rather than from a single provider chosen for the whole shipment. AI agents make that shift practical because they give the buyer a digital actor that can express intent once, compare options across providers, and move from quote to action without forcing the buyer to adapt to every carrier’s channel. The risk of waiting is also growing, because every quarter a provider remains unprepared for agent-led buying adds to its onboarding debt while faster competitors begin to shape the new operating defaults.

THE INCENTIVES

Both sides have a reason to move

Buyers are the ones that will pull this shift into the market, because an agent gives a shipper capabilities that are hard to achieve through today’s carrier-led channels. Instead of receiving quotes in different formats and reconciling them manually, the buyer gets offers that are structured, comparable and rankable from the start. Instead of being limited to one provider’s end-to-end offer, it can assemble stronger multi-leg options across providers and instead of carrying the integration burden for every carrier relationship, it can express intent once and let the agent handle the interaction. For a procurement team, this means every ranked offer comes with the provenance needed to understand why it was selected, and decisions that previously took days can move to execution in seconds.

Agents unbundle the shipment

A carrier has an obvious reason to refuse. Exposing a quotation agent puts its price next to its rivals’ on every shipment. But the all-or-nothing model is the thing already at risk, and agent-mediation changes the unit of competition from the shipment to the leg. The carrier that loses the ocean leg can still win the inland, the terminal, and the visibility and get paid for each. The revenue the old model forfeited becomes capturable.

The unit of competition moves from the shipment to the leg. Legs shown are illustrative.

This is unbundling and it decides who gains. When value is no longer hidden inside a single end-to-end shipment price, it becomes easier to see which provider is genuinely strongest across each part of the journey. That creates new opportunity for differentiated carriers that may not win the full shipment today but can still capture the parts where they perform best. At the same time, it creates pressure for integrators that have relied on bundled pricing to balance stronger and weaker parts of their network. The caveat is that not every part of the value chain can be separated cleanly, because some value only exists when the pieces work together and may not survive if each service is treated in isolation.

WHAT IT LOOKS LIKE

One intent, parallel carriers, ranked offers

For a small shipper, the change feels like moving everything onto one simple screen. The shipper describes the shipment in plain language, the agent reaches relevant ocean and inland providers at the same time and the responses come back as ranked options that can be compared against the buyer’s priorities. Those options may include a single provider’s route or a combination of providers across different parts of the journey, but the buyer no longer has to gather, normalize, and compare each offer manually. Large shippers follow the same pattern at enterprise scale, using a comparison agent grounded in their contracts, preferred lanes, service history, and past performance. The impact also extends beyond quotation, because booking, shipping instructions, milestones, and invoicing can move through agent-to-agent interactions, reducing the manual work of re-keying information, chasing status updates, and reconciling invoices by hand.

THE CONTROL POINT

Who owns the buyer agent

This shift depends on the buyer having an agent of its own, because that agent becomes the new point of control in the buying process. The answer will look different by buyer size. The largest shippers are likely to build and operate their own agents, while most mid-market and long-tail shippers will need a trusted surface they can plug into rather than building that capability themselves. That second group represents a large part of the market, and it creates the opening for a neutral integrator to bring shippers and carriers together through a common agent-enabled marketplace.

The marketplace, however, is the destination rather than the first move. It cannot work until there are agents on both sides worth connecting, so the immediate priority is more foundational: help buyers and carriers build their agents, make carrier capabilities agent-addressable, and enable those agents to discover and interact with each other. That foundation has value even before a marketplace exists. A carrier benefits from making its services callable by buyer agents, and a shipper benefits from having an agent that can reach the market more easily.
This is also where IBM Consulting is starting, with current Travel and Transportation proposals focused on building agents on both sides and enabling discovery between them before committing investment to the marketplace layer itself.

The marketplace becomes the next bet once that foundation is in place, and its success depends on one critical condition - trust. A surface that ranks carriers against each other is only credible if it is operated by a party that has no stake in which carrier wins. A carrier would struggle to run it because rivals would not want to expose their agents inside a competitor’s environment. A forwarder would face the same issue because it is also competing for the freight, and a shipper-owned surface would create the mirror concern from the supply side. Neutrality therefore becomes the qualification. The party best positioned to operate the marketplace later is the one that first helps both sides build their agents, connect to each other, and establish trust, without forcing the full marketplace investment before the foundation is ready.

The destination. A neutral marketplace can rank carriers head to head because it competes with neither side. It is reached only after both sides have agents and can discover each other, and big buyers can still run their own agents on the same rails.

BUILDING IT

Build the foundation first

This foundation will not appear on its own, and the order of work matters. Before any marketplace can exist, both sides need agents that are useful in practice: carriers need to make their capabilities addressable, buyer agents need enough context to reason beyond basic booking data, and there needs to be a discovery layer so the two sides can find and interact with each other. The raw material is already there. Carriers already have APIs and EDI gateways, models can turn a plain-language shipment request into structured parameters, and emerging agent protocols allow agents to discover and call one another. The near-term task is to connect these pieces properly by making existing carrier surfaces agent-addressable, giving agents trusted knowledge to reason from when the transactional data is not enough, and enabling discovery between buyers and providers.

On the supply side, carriers need a way to make their existing operations agent-ready without replacing the systems they already depend on. IBM Enterprise Advantage brings these capabilities together in one platform: understanding current processes and identifying where agentic automation can create value, grounding agents in enterprise knowledge, and connecting them securely to pricing, routing, transport management, EDI, compliance, and other operational systems. This creates a governed boundary through which agents can discover capabilities, interact with one another, and act across the carrier’s technology estate while keeping execution auditable.

The harder part is grounding. Agents may work well when a shipment follows the happy path, but become unreliable when decisions depend on rules buried in tariff schedules, dangerous-goods classifications, customs requirements, or exception-handling procedures. IBM Enterprise Advantage provides the enterprise context needed for agents to reason over that knowledge while retaining the evidence behind their decisions. For example, if an agent incorrectly classifies 10,000 car parts under a single HS code and risks triggering a customs hold, it can query the enterprise knowledge layer, identify the correct multi-line classification, understand the rule behind each line, and correct the filing before the shipment is delayed. That same provenance is also essential to any future marketplace, because rankings and execution decisions that cannot be audited will not be trusted.

The marketplace should therefore come later, by design. First, federation makes carrier capabilities reachable, grounding gives agents reliable context for decisions, and discovery allows buyers and providers to find each other. A full exchange needs additional layers such as identity, registry management, and ranking integrity, but those investments make sense only once there are working agents on both sides worth connecting. ContextForge becomes the seed of that future exchange by creating the governed agent gateway first. Taken together, the foundation is clear: make the supply side addressable, make the demand side discoverable, and put grounding and provenance underneath. Once that foundation is real, the marketplace earns its place, and the neutral party that helped build trust on both sides is best positioned to stand in the middle.

The buyer states an outcome. Everything downstream, from quote to invoice to a customs correction at the border, becomes the agents’ problem to solve.

The question on the table

Tomorrow’s freight will increasingly be booked through agents talking to agents, with buying decisions broken down across the parts of the journey where each provider is strongest. For logistics providers, the question will no longer be only whether they can win the full shipment, but where they are genuinely competitive and whether their agent is present when those opportunities are being evaluated. For the companies building the rails, the immediate priority is to get agents working on both sides and make it easy for them to discover and interact with each other. The neutral middle, where freight can eventually be compared, ranked, and booked across providers, is the prize that comes later, and it will belong to the party that earned trust from both sides by helping build the foundation first.