How the Iran conflict stress-tested AI supply chain planning platforms

The most useful evidence from the Iran conflict is not a quote about resilience or a slide about autonomous planning. It is a usage spike. During the crisis, Kinaxis said oil-and-gas customers increased scenario-planning activity by 1,400%, and that daily planning activity across industries rose 306% on June 22, 2026.[1] For anyone evaluating AI supply chain disruption planning readiness during the Iran conflict, that matters because it is dated operational telemetry rather than another forecast about what AI might eventually do.

It does not prove Kinaxis solved the disruption. It does not prove competing platforms failed. It proves something narrower and more valuable: when the news cycle became operationally relevant, enterprises reached for scenario-planning capacity in a visible, measurable way. Planning teams were not waiting for a quarterly S&OP cycle to absorb the shock. They were testing assumptions, refreshing plans, and asking systems to answer questions that spreadsheets usually inherit when the official planning layer is too slow.

What the Kinaxis spike actually measures

A 1,400% increase in oil-and-gas scenario planning is a behavioral signal. It means planners were asking more “what if” questions inside the platform: what if energy flows are constrained, what if lead times stretch, what if a port or lane becomes less reliable, what if capacity needs to move, what if inventory policy has to change faster than the normal planning calendar allows. The 306% daily planning surge is broader, but it points in the same direction: more users, more plan refreshes, more pressure on the planning system to represent a changing operating reality.[1]

That distinction matters. Scenario generation is often discussed as a feature demonstration: enter disruption, generate options, compare outcomes. Crisis usage is different. The planners are not exploring possibilities because it is interesting. They are trying to decide whether to expedite, reallocate, buffer, substitute, rebalance, or tell commercial teams that a promise date is no longer safe.

Kinaxis deserves credit for publishing the kind of time-stamped platform activity buyers rarely see. Most AI planning claims are expressed as capabilities, customer anecdotes, or analyst positioning. Usage during a live disruption is closer to the evidence planning owners need: when uncertainty rose, did the system become part of the operating response, or did it sit beside the real work?

The caveat is equally important. The data is first-party and vendor-published. It is not an independently audited benchmark, and it is not a head-to-head comparison against o9, Blue Yonder, RELEX, Anaplan, or any other planning platform. The absence of comparable public data from those vendors should be read as a disclosure gap, not as evidence that their customers did less planning or handled the conflict worse.

Still, disclosure changes the buyer conversation. A vendor that can show when customers used scenario planning, in which industries, and by how much is giving procurement and transformation teams a better starting point than a generic claim about AI-enabled resilience. The next question is whether that activity translated into executable decisions.

The hard part begins after the scenario is generated

GEP’s contemporaneous analysis put pressure on the part of the story that platform marketing often skips: orchestration across the systems where actual constraints live. GEP argued that the war was stress-testing global supply chains while fragmented supplier, logistics, and contract data kept AI planning tools from acting with the speed that geopolitical disruption requires.[2]

That is the familiar gap in a more urgent setting. A planning model can show that an alternate supplier would reduce exposure. The contract system may still show minimum commitments with the incumbent. The logistics team may know that the alternate lane is technically available but commercially unattractive. Procurement may have supplier risk information that never made it into the planning model. Inventory may be visible at the regional level but not clean enough by site, batch, or allocation rule to support the modeled move.

The result is not a simple failure of AI. It is a failure of connection. The system can make uncertainty legible without making the response executable. In a war room, that difference shows up quickly: the dashboard says one thing, the buyer is waiting on a supplier confirmation, the logistics lead is asking whether a lane is still protected, finance wants to know the margin impact, and someone exports the scenario into a spreadsheet because the decision path crosses too many disconnected systems.

This is where the Kinaxis numbers become more interesting, not less. A surge in planning activity shows that demand for scenario capability was real. GEP’s critique shows why scenario volume cannot be treated as proof of autonomous response. The useful evaluation question is not whether the platform can produce more scenarios under stress. It is whether the organization can move from scenario to approved action without rebuilding the plan manually across procurement, transportation, contracts, and inventory.

Many companies entered the crisis already short on agility

The Iran conflict did not hit a clean slate. In January 2026, before the full supply-chain effects of the war had materialized, a RELEX survey of 514 supply chain leaders found that 86% had been impacted by trade policy changes, 51% were raising prices, and only 18% were restructuring supply chains.[3] That is not an Iran-war outcome study. It is a pre-war baseline for the operating posture many companies carried into the disruption.

The gap between “impacted” and “restructuring” is the part planning buyers should sit with. Price increases can be executed through commercial processes. Structural changes require supplier qualification, network redesign, inventory policy changes, and contractual flexibility. A planning platform can support those moves, but it cannot invent governance or data readiness at the moment a crisis starts.

This also explains why scenario planning can spike without producing equally fast structural adaptation. A planner can model a region shift today. The organization may need weeks or longer to approve the supplier, secure capacity, update contracts, validate quality, and reroute logistics. The model compresses the analysis cycle; it does not erase the operating work.

The responses leaders described were exactly the ones platforms struggle to orchestrate

Zero100’s April 2026 roundtable adds a useful, limited window into what leaders were actually doing. In a group of roughly 20 to 30 supply chain leaders concentrated in consumer goods, food and beverage, QSR, packaging, and logistics, 90% said they had moved orders to different regions, 60% said they had increased critical-material inventories, and 88% said the conflict would accelerate regionalization.[4]

Those findings should not be generalized across every sector. They do not tell us how semiconductor, automotive, chemical, or energy supply chains behaved as a whole. But they are directionally useful because the actions named in the roundtable are the same ones that test a planning platform’s practical depth.

Each response crosses functional boundaries. Moving orders is not only a planning calculation. It touches procurement authority, logistics feasibility, customer allocation, and finance approval. Raising inventories is not only a safety-stock adjustment. It consumes cash and warehouse capacity, and it may create excess if demand shifts again. Regionalization is not a scenario; it is a multi-year operating choice unless the company already built the optionality.

That is why the phrase “AI planning platform” can mislead during geopolitical disruption. The planning layer may be intelligent, but the decision is only as fast as the slowest dependency it must validate. If supplier risk, freight availability, contract exposure, and inventory positions remain outside the planning workflow, the platform becomes a very good place to see the problem before the organization solves it somewhere else.

A useful stress test has multiple clocks

The better scenario-planning frameworks do not treat a geopolitical shock as one event with one answer. Logistics Viewpoints framed Iran-related scenario analysis around different durations, including a short 7-day conflict and a prolonged 4-plus-week conflict, with different recovery implications for global manufacturing.[5] That is the right instinct because the response to a short shock is often an allocation and expediting problem, while the response to a longer one starts to look like sourcing, inventory, and network redesign.

AIMMS describes geopolitical disruption planning as a sequence that includes identifying scenarios, modeling full-network consequences, evaluating response options, and building pre-authorized playbooks.[6] The last phrase is easy to underestimate. Pre-authorization is what separates a modeled option from a decision that can survive a 7 a.m. executive call. If every alternate move requires fresh approval from procurement, finance, legal, logistics, and sales, the AI system may be fast while the company remains slow.

This is also where RFx language needs to become more operational. Asking a vendor whether it supports geopolitical scenario planning is too soft. Most serious platforms can say yes. The better questions are about which data objects the scenario can actually use, which constraints are live rather than manually refreshed, which approvals can be embedded, and which actions can be pushed back into execution systems without a side process.

What buyers should take from the Iran conflict

The Iran conflict did not produce a public platform ranking. It produced one unusually concrete disclosure from Kinaxis, a contemporaneous warning from GEP about fragmented execution data, and supporting signals that leaders were already moving orders, raising inventories, and thinking harder about regionalization. That is enough to change the evaluation standard, but not enough to declare winners.

For a renewal or selection team, the first implication is to ask every vendor for crisis-period telemetry. Not customer logos. Not a maturity model. Ask whether the vendor can show dated changes in scenario runs, planning frequency, exception volumes, user activity, decision latency, or execution handoffs during a named disruption. Kinaxis has made that kind of disclosure part of the public record for this event; others may have similar internal evidence, but buyers should not have to infer it from positioning decks.

The second implication is to test the handoff. A scripted demo should not end when the system produces three scenarios. It should continue until one option becomes an approved action: alternate supplier selected, logistics constraint checked, contract exposure visible, inventory impact calculated, customer allocation updated, and the execution system receiving the decision. If that path breaks into emails and offline reconciliation, the platform may still be valuable, but it is not orchestrating the response.

The third implication is to separate planning intelligence from organizational readiness. RELEX’s January survey suggests many companies were already feeling trade-policy pressure before they had made structural changes.[3] Zero100’s roundtable suggests some leaders were actively shifting orders and increasing buffers as the conflict risk intensified.[4] A platform can accelerate analysis for both groups, but the company that already has alternate suppliers, approved playbooks, and clean constraint data will get more out of the same AI capability than the company that has to build those foundations during the event.

Supply Chain Brain’s broader read on the crisis argued that the Iran conflict exposed AI supply chain blind spots, especially where sophisticated planning claims met fragile real-world infrastructure and data dependencies.[7] That framing is useful as long as it does not become a blanket dismissal. The better lesson is more specific: AI scenario planning became more necessary during the conflict, and its limits became more visible at the same time.

That is a more useful conclusion than either triumph or failure. The demand signal was real. The orchestration gap was real. The next serious platform conversation should start where the Iran conflict left planning teams exposed: not at “Can you generate a scenario?” but at “Can this organization act on the scenario before the workaround takes over?”

References

1. How the Iran conflict triggered a 1,400% supply chain scenario planning surge — Kinaxis
2. U.S.-Israel-Iran War Is Stress-Testing Global Supply Chains (And Here's What You Must Do) — GEP
3. RELEX: Companies are investing in resilience rather than stability — The SCXchange, January 2026
4. Four Potential Iran War Supply Chain Scenarios – and What to Do Now — Zero100, April 2026
5. Supply Chain Scenario Analysis: Short vs. Prolonged U.S.–Iran Conflict — Logistics Viewpoints, March 4, 2026
6. Building Geopolitical Disruption Scenarios for Your Supply Chain — AIMMS
7. The Iran Crisis Is Exposing AI's Supply Chain Blind Spots — Supply Chain Brain, 2026