Within the first few conversations on the first day of Manifest 2026, it was clear this wasn’t going to be another conference debating whether AI belongs in supply chain operations. That conversation has largely run its course. With more than 7,200 attendees and over 400 exhibitors, ‘accountability,’ and not hype, was front and center on the Las Vegas floor.
The mood felt different this year. Less “growth at all costs.” More “margin first.”
AI wasn’t positioned as a differentiator. It was treated as table stakes.
Jake Barr, CEO and Principal of BlueWorld Supply Chain Consulting, described the operating climate as “the new never normal,” and that sentiment lingered throughout the event. Supply chains are no longer bracing for disruption; they’re built inside it. DHL’s Insight 2030 research reflected that reality, with the majority of leaders expecting increased nearshoring investment, nearly three-quarters anticipating robotics to shape operations by 2030 despite lagging adoption, and over 60 percent expecting disruptive forces to materially impact their networks.
That kind of environment doesn’t reward experimentation. It rewards precision.
Security and resilience were impossible to ignore. Exhibitors showcased biometric trailer locks, hybrid cloud-edge architectures designed to reduce downtime, and cargo security systems aimed at combating theft that continues to surge. Robotics vendors emphasized flexibility and asset utilization rather than spectacle. Cold-chain drones and decoupled loading systems were presented not as futuristic gadgets but as practical tools to reduce truck turnaround time and maximize throughput.
The tone had shifted. Innovation was still present, but now evaluated through a margin lens.
The $600 Million Wake-Up Call
One of the more compelling moments came when Scotts Miracle-Gro shared how predictive AI helped reduce year-end inventory by $600 million. That figure landed because it was practical and represented disciplined decisions driven by better data.
As David Huskisson, Head of Enterprise Transformation at Scotts Miracle-Gro, explained, the math was simply too complex for humans to solve alone. That observation reframed the conversation. AI wasn’t about dashboards or incremental insights. It was about solving operational equations that materially affect capital.
The Real Conversations Weren’t On Stage
While the panels highlighted success stories, the more candid discussions were happening between booths.
On the floor, unified platforms promised to eliminate integration silos. AI systems were converting customs documents into digital filings. Dynamic rules engines were helping brokers operationalize pricing logic without code. Edge intelligence solutions were pushing decisions closer to docks and equipment, and autonomous trucks parked outside served as a reminder that the physical and digital supply chain is converging faster than many expected.
What stood out was that interest had shifted away from agentic AI as a concept toward deployable solutions. Prospects wanted to see working systems, with clarity on commercial models. Mid-market carriers and 3PLs, in particular, were looking for AI agents that could be implemented quickly and tied directly to measurable ROI.
Several leaders were candid about their own maturity gaps. AI and automation initiatives were often further behind than anticipated. Visibility into decision behavior was limited. Scaling efforts felt uneven. Some organizations were running on heavily customized WMS environments built years ago and were now facing modernization pressure. Others were in the middle of rethinking their warehousing stack entirely.
As Dave Yoder from Ryder noted, without a structured decision-making framework, agentic AI does not create value. That observation resonated because many teams are operating somewhere between pilot and production, a phase where enthusiasm often outpaces discipline.
And the challenge isn’t purely technical. Scaling AI also exposes talent gaps. Many software providers acknowledged leaning on nearshore and offshore partners to augment internal capabilities as they try to move from experimentation to operationalization.
That broader reality made the case for ArkOS feel less theoretical and more timely.
What We Brought Into That Tension
At Trigent, we focused on two things: operational AI Assistants designed for logistics workflows and the introduction of ArkOS.
The Logistics Assistants weren’t positioned as futuristic experiments. They handled carrier check calls, load confirmations, appointment scheduling, and exception management, the connective tissue of logistics operations. By the second day, operators were less interested in whether it worked and more interested in what a 15 to 25 percent reduction in repetitive coordination would mean for margin and staffing.
One executive joked that if the system handled check calls consistently, he might finally get his weekends back. Beneath the humor was a real calculation about capacity and cost.
Yet nearly every substantive conversation circled back to the same concern: how to deploy automation without introducing new fragility.
Why ArkOS Changed the Conversation
Building AI models is no longer the barrier it once was. Understanding how those models behave under production pressure is.
Pilots are controlled environments. Production is not. As volume grows, cost per transaction shifts, and latency behaves differently under sustained load. Decision logic that feels clean in isolation becomes harder to trace once distributed across systems. Infrastructure decisions, once embedded in hyperscale environments, become difficult to unwind.
That context was particularly relevant given what was visible across the show floor. Companies like Dispatch Science were introducing unified platforms that collapse execution and intelligence into a single layer. Descartes demonstrated AI reducing compliance false positives to fractions of a percent. Tabi Connect highlighted a rules engine that helped a $4B brokerage generate over $100 million in new revenue. Midmo showcased edge intelligence modules that bring decision-making closer to physical assets. Samsung SDS emphasized global AI expansion supported by cloud architecture.
The industry is embedding intelligence into every layer of operations.
As that happens, tolerance for opacity shrinks.
ArkOS was built for precisely that inflection point. It provides a client-owned workbench where AI workflows can be built locally, tested against real data, evaluated for cost per transaction and latency behavior, and refined with governance controls before promotion into production infrastructure.
Instead of discovering economic surprises after scale, teams can inspect execution behavior beforehand.
When we demonstrated how Logistics AI Assistants could operate within that validated framework, the focus moved from enthusiasm about automation to practical planning around responsible rollout.
Trust, transparency, and human change management are becoming differentiators in digital transformation. Without them, AI integration falls short. AI-driven decision making has to be explainable, governable, and aligned with operational reality.
A More Mature Phase
By the time the event wound down, what stayed with me wasn’t any single product announcement or demo. It was the change in posture across the room.
A couple of years ago, AI conversations felt exploratory. This year, they felt consequential. Leaders were no longer there to be impressed. They were there because they’re accountable for performance, and under pressure to deliver speed and visibility without sacrificing cost control or resilience.
This has shifted the focus from merely improving individual workflows with AI to embedding decision systems in operations that can be trusted to perform under pressure. That transition is where most of the complexity lives.
Our Logistics Assistants address a very tangible layer of operational friction. Trigent ArkOS addresses something more structural: how to validate and understand AI behavior before it becomes embedded infrastructure. That combination reflects where the industry seems to be heading: less fascination with capability, more focus on control.
If you weren’t able to stop by Booth 2314B, we’re happy to continue the discussion. These aren’t theoretical conversations anymore. They’re planning conversations.
