Transportation Insight turns live freight data into an AI decision-intelligence layer

Transportation Insight launched TI Infinity for enterprise and mid-market shippers, positioning the platform as a decision-intelligence layer over transportation data, market intelligence, spend, and performance information.

TI Infinity continuously analyzes those inputs for trends, risks, and opportunities, then lets users investigate causes in natural language, model alternatives, and receive a recommended next action. Its planned expansion connects transportation signals with broader enterprise data so margin and profitability effects can be evaluated alongside freight outcomes.

For logistics leaders, the change is from retrospective reporting to a live decision queue. The announcement does not establish a measured customer result, so the commercial test is whether the platform reduces time-to-decision while improving freight cost, service, and margin visibility.

NextSmartShip founder William Yu described a hybrid fulfillment model for direct-to-consumer brands that begins near production and shifts proven products into local warehouses in markets such as the United States and Europe.

The company combines an order-management and warehouse-management platform with distributed inventory, AI-enhanced order routing, and automated carrier selection across more than 400 shipping channels serving 220-plus countries. The design lets inventory placement change as demand becomes clearer rather than forcing a brand into one fixed network.

The model treats fulfillment as part of the customer experience and as a response to changed cross-border duty economics. NextSmartShip says it serves more than 2,000 brands; the operational question is whether allocation decisions reduce landed cost and delivery time without creating fragmented stock.

Brooklyn-based Ultra raised $62 million across a $50 million Series A and a $12 million seed round while deepening its relationship with Physical Intelligence. Ultra leases packing robots to warehouses rather than requiring customers to purchase the full system upfront.

Ultra installs and supports the hardware while Physical Intelligence supplies the robotics software, creating a division between the physical system and the model layer. The company says its robots have packed more than half a million orders, and customers pay an integration fee followed by a recurring hardware-and-software support charge.

That commercial structure changes how 3PLs can evaluate automation: monthly operating cost, integration time, uptime, and throughput matter more than a single capital-return calculation. The reported order volume is company-provided, so buyers still need site-specific evidence across packaging mix and exception rates.

The Defense Logistics Agency completed a multi-year rollout of a global logistics system across 24 sites. The deployment represents a completed enterprise systems transition rather than a single algorithmic pilot, with the operational objective of giving a large logistics network a common execution foundation.

A network-wide logistics system creates the data and process substrate needed for analytics, exception management, and future AI decision support. The important implementation challenge is harmonizing master data, roles, interfaces, and operating procedures across sites that previously may have used different practices.

For logistics organizations, system consolidation is a prerequisite for trustworthy optimization. A model cannot reliably compare inventory, fulfillment, or transportation performance when site definitions and transaction timing differ.

Finnish retail group Kesko selected Manhattan Active Warehouse to replace its legacy grocery-trade WMS. The initial program covers five distribution centers, a greenfield site planned for 2029-30, and terminal-hub operations.

The selected platform includes Manhattan Active Agents and Agent Foundry, with planned use of Wave Coordinator, Labour, and Warehouse Associate Agents. Kesko also expects to explore additional agents for inbound operations, allowing recommendations and execution controls to sit closer to live warehouse signals.

Kesko’s grocery trade recorded EUR 6.5 billion in 2025 net sales and EUR 418 million in comparable operating profit, making the project material at regional scale. The stated objective is to improve real-time interpretation and response without the disruption of repeated legacy-system upgrades.

The Supply Chain Management Review analysis describes two high-volume retailers using coordinated logistics networks to reduce cost, accelerate freight movement, and adapt inventory and operations to seasonal demand. One example, Outdoor Cap, is a Bentonville, Arkansas manufacturer whose customers emphasize ease of doing business, delivery speed, and complete orders.

The operating design connects import receiving, warehousing, kitting, transportation, and store replenishment, with selective direct fulfillment used where it avoids unnecessary movement through a central distribution center. The recommendations emphasize product-flow analysis rather than optimizing isolated transportation legs.

The network lesson is that speed without completeness or seasonal capacity can still damage the customer promise. The article recommends measuring cost, delivery performance, order completeness, and responsiveness together.

YMX Logistics and Outrider announced a five-year commercial partnership to expand self-driving, zero-emission yard trucks at enterprise customer sites. YMX will finance, operate, and support the fleets, while Outrider supplies the autonomous yard system and remote technical support.

Outrider’s physical-AI system handles trailer spotting, mixed-traffic navigation, hitching and unhitching, precise dock positioning, brake-line connection, trailer inventory tracking, and EV charge monitoring. The system integrates with yard, warehouse, and transportation-management systems, while YMX offers capacity-based pricing.

The arrangement packages site planning, deployment, financing, and operations into one service relationship for food and beverage, retail, consumer packaged goods, and automotive customers. It is a planned deployment model rather than a reported productivity result, so yard safety and dwell baselines remain essential.

IFS introduced Fulfillment Flow from IFS Softeon, a warehouse capability intended to help teams understand what is happening, why it is happening, and what should happen next. The product is designed to move AI from isolated tools into the operating flow of fulfillment.

Digital workers focus on inbound, wave management, labor, automation, slotting and replenishment, and outbound. They share a common view of facility conditions, live operations, adjacent processes, site goals, associates, and automation with a central operational agent that evaluates cross-area consequences.

The architecture addresses a planning problem common to complex facilities: a local optimization can worsen receiving, manufacturing feeds, or outbound cutoffs. IFS describes the capability as a new product introduction and does not publish a customer KPI result.

Logistics Viewpoints framed its 2027 agenda around eight executive decisions spanning transportation, warehousing, fulfillment, planning, automation, AI, data, resilience, trade, and market structure. The argument is that dependencies among these domains are becoming a source of both risk and competitive advantage.

The agenda asks leaders which processes should operate as one system, where intelligence belongs, how much autonomy is appropriate, and what evidence is needed before scaling a decision. It treats interfaces, state, decision rights, and failure modes as design objects rather than leaving coordination to sequential human handoffs.

For network planners, the practical consequence is a shift from buying isolated optimization tools to designing a coherent response path. A transportation exception should be able to trigger inventory analysis, warehouse replanning, customer prioritization, and appointment changes without losing time between systems.

Transfix and E3 Group integrated driver authentication and fraud-prevention software into the Transfix TMS. Brokers can authenticate a driver before releasing the pickup number, and Direct Traffic Solutions is among the first brokers using the live integration.

E3 scores 12 to 14 driver risk factors, while Transfix carries the full signal set into the dispatch workflow instead of reducing it to a basic pass/fail field. The result is a single-screen review that avoids re-keying and separate portal checks during load assignment.

The partners report zero stolen loads to date and a fee-refund guarantee, but that claim is not an independently audited fleet-wide outcome. The operational value is the placement of the control at the moment a fraudulent pickup can be prevented.

Loadsmart AI launched freight agents that execute repetitive transportation work inside the systems enterprise shippers already use. The rollout is designed to begin with one high-volume workflow, customer-defined rules, and guardrails rather than a full platform migration.

The agents can collect and file documents, retrieve carrier status, rebook dock appointments, retender failed loads, update TMS records, and process claims. Loadsmart connects through API, EDI, MCP, or existing system interfaces, and its freight operators close exceptions that the agent cannot resolve; the company reports roughly 80% autonomous resolution for the tasks it touches.

The onboarding implication is that customers can introduce automation at the workflow boundary they already understand. Because the remaining 20% is handled by Loadsmart’s operators rather than returned to the shipper team, the buyer must define escalation ownership, audit records, and the conditions for expanding scope.

Logistics Reply introduced the LEA AI Agent Authority Model for warehouse execution on its LEA Reply platform. The platform combines warehouse management, planning and optimization, automation and robotics integration, orchestration, and AI capabilities.

The authority model is designed to define what agents may observe, recommend, or execute within the warehouse environment. It builds on execution software and data intelligence rather than treating an agent as a free-standing conversational layer.

For 3PLs, authority boundaries are part of customer trust and operational safety. A recommendation affecting inventory, labor, or equipment may require a different approval path from a low-risk status explanation.

Lineage connected Kargo AI camera towers directly to its WMS at a temperature-controlled warehouse in Decatur, Alabama. The facility supports adjacent poultry plants and handles more than 500 pallets per day.

The camera system creates receipts during ordinary forklift movement, linking material identification and lot information to warehouse transactions without adding a separate scanning step. The project automated more than 64,000 receipts over six months, according to the companies.

Lineage and Kargo emphasize case-level traceability and food safety as well as labor savings. The broader operational test is whether automated receiving improves dock-to-stock time and reconciliation without weakening controls for damaged, partial, or ambiguous pallets.

Transporeon described an agent-ready logistics platform that brings procurement, planning, dock, yard, and settlement workflows into a connected operating model. The platform is aimed at replacing spreadsheet-heavy tendering and fragmented supplier communication.

NextGen Freight Procurement for Road uses lane and pricing data to prepare AI-assisted RFQs, while Living Purchase Orders let supplier changes flow into transport and yard plans before a truck departs. The design links planning inputs to downstream execution rather than freezing the plan at tender time.

For inbound networks, the ability to update a plan before arrival can reduce avoidable dock conflicts and manual rework. The source describes platform capability rather than a controlled KPI result, so users need lane-level measures for appointment, dwell, and tender quality.

DHL Supply Chain announced a five-year alliance with Robust.AI to introduce Carter collaborative robots into its Latin American operations. The initial Mexican retail rollout includes 15 units, with a longer-term plan for hundreds across the Americas.

The robots will be integrated with DHL’s WMS by 2026 and deployed through a phased scaling plan. DHL already uses multiple autonomous mobile robot and robotic systems, so the new agreement adds another capability to an existing multi-vendor automation environment.

The operational challenge is not simply moving robots into a facility; it is maintaining consistent task, WMS, safety, and exception behavior across customer sites. The announcement does not report a Mexican productivity result yet.

DHL Supply Chain announced a broader deployment of SVT Robotics’ SOFTBOT multi-agent orchestration platform across its global warehouse network. The platform is live at 30 sites, with plans to exceed 100 sites over the next three years.

DHL says more than 8,000 collaborative robots are already active in its operations, while SOFTBOT can deploy robotics integrations up to 12 times faster than traditional custom coding. The platform is designed to support modular automation that can be reconfigured as customer profiles, volumes, and technology change.

The business implication is a move from one-off robotics projects toward a reusable control layer. Faster integration will matter only if it preserves uptime, WMS consistency, safety controls, and site-specific customer SLAs.

Zalando announced plans to deploy up to 50 Nomagic robots across fulfillment centers in Rotterdam, Stockholm, Paris, and Giessen. The rollout follows a testing phase and focuses on automated handling of shoeboxes, including two-piece and unsealed boxes.

The robots use AI-driven pick-and-place behavior for a product class with variable packaging and presentation. The multi-site plan gives Zalando a way to compare performance across different facility layouts and operating conditions rather than relying on one demonstration site.

For ecommerce fulfillment, shoe handling sits directly on the path from inventory to customer promise. The source does not provide a broad productivity result, so the relevant measures are pick success, exception rate, damage, and pack-station continuity.

Barrett Distribution Centers expanded its partnership with UNIT AI and plans to deploy UNIT’s Networked Physical AI Platform across its operations beginning in 2027. The program is designed for a national 3PL network serving apparel, footwear, consumer products, and other high-touch industries.

UNIT’s platform connects distributed inventory placement, fulfillment orchestration, network-wide visibility, and decentralized returns processing. Its Warehouse-as-a-Service model is intended to let Barrett introduce automation incrementally without the capital burden and implementation timeline associated with a traditional full-site deployment.

The proposal moves warehouse automation from facility-level efficiency toward coordinated network response. Because deployment is planned, Barrett must prove that inventory placement, fulfillment latency, and returns recovery improve without making customer-specific exceptions harder to manage.

SAP published its Logistics Assistant as part of an autonomous supply-chain portfolio spanning warehousing, fulfillment, and distribution. The assistant is intended to detect disruptions, plan responses, and coordinate execution involving labor, inventory, dispatch, and freight decisions.

The page describes specialized assistants for outbound warehouse task bundling and queue assignment, with related planning assistants for creating, adjusting, and reviewing plans. The product is positioned as an orchestration layer around existing supply-chain operations rather than a single forecasting model.

For fulfillment teams, the useful test is whether the assistant converts disruption signals into approved work assignments without hiding tradeoffs. SAP does not publish a customer-specific KPI in the page, so buyers should validate queue latency, override rate, and service outcomes.

PULPO WMS released a merchant portal and activity-based billing for fulfillment providers. The portal covers dispatch performance, inventory and expiry data, demand analytics, sales orders, returns, and invoices, while merchants can ask questions against a specific order or case.

The billing engine prices more than 60 services, including receiving, storage, picking, packing, returns, and value-added work, with charges accruing as tasks are completed. Demand tools predict SKU stock-out dates, recommend reorder quantities, organize picking rules, flag blocked tasks, and recommend cartonization and put-wall flows.

The release is live for customers in more than 20 countries and was rolled out in August 2026. Its value is operational and commercial at once: fewer spreadsheet reconciliations, more transparent merchant service, and better visibility into where a 3PL actually consumes labor.

Kodiak AI and Charger USA began hauling refrigerated and dry freight for consumer packaged goods and food-and-beverage customers on the 435-mile Dallas-Laredo lane. The companies completed their first delivery on September 8 and are operating initially with a safety driver.

The Kodiak Driver supplies the autonomous driving system, while Charger contributes its TMS, managed transportation, and carrier operations. The partners intend to move to driverless operations after Kodiak completes its route-specific safety case, and Kodiak says its long-haul readiness measure reached 96% at the end of September.

The corridor reaches Laredo, the highest-volume commercial land port in the United States, so predictable transit and extended operating hours could matter for replenishment flows. The current deployment is supervised, and the driverless transition remains contingent on evidence rather than the announcement alone.

Geotab and AtoB introduced the Geotab Fleet Card for eligible U.S. fleets, combining a payment platform and fuel network with Geotab vehicle data. The product is designed to control a fuel purchase before it is completed rather than relying only on post-transaction alerts.

The card can remain locked until the driver unlocks it through the Geotab Drive app while parked at the merchant. Authorization checks the vehicle’s proximity, merchant GPS coordinates, and fuel level, and can decline a transaction when the vehicle is not at the pump or the tank cannot hold the requested amount.

Geotab also reports price differences of up to 77 cents per gallon in an analysis of 419,000 diesel fills across 660 fleets. The savings and fraud estimate are company analyses, so fleet managers need local transaction and route baselines.

Modern Distribution Management examined AI route optimization as distributors face rising diesel, labor, material, and operating costs. Building-supply and food-and-beverage distributors face especially difficult delivery patterns because vehicle type, unloading equipment, jobsite timing, and customer requirements vary by stop.

The operational problem is not simply finding the shortest route. A useful optimizer must account for vehicle and equipment constraints, short lead times, tight delivery windows, jobsite availability, and the cost of miles between distribution centers and customer locations.

The article frames route optimization as a margin-protection tool that must preserve reliable delivery. The relevant test is whether fewer miles reduce cost without increasing failed deliveries, late jobsites, driver overtime, or customer-service exceptions.

Kooner Fleet Management Solutions and Global Partners announced a referral partnership connecting mobile fleet maintenance with commercial fueling services. The offering is designed to perform complementary work where vehicles are already located rather than sending them on separate service trips.

The model links bulk fueling, preventive maintenance, inspections, and on-site repairs through coordinated referrals, with the objective of using fleet downtime more productively. It is an operating-service integration rather than an AI product, but it creates a structured data opportunity around service timing, vehicle availability, and maintenance completion.

For outbound fleets, avoiding off-site trips can protect route capacity and reduce nonproductive miles. The announcement does not quantify the effect, so operators should measure vehicle availability, maintenance cycle time, and missed-route exposure by depot.

Walmart’s marketplace logistics strategy includes seller protections, returns, fulfillment, multi-channel inventory, parcel transportation, pricing tools, and advertising. The combined offer is intended to give sellers a single operating model for customer orders and post-purchase work.

A shared inventory pool can support orders from multiple channels, while returns and seller-protection processes feed back into how inventory is managed and how merchants evaluate fulfillment performance. The operating design therefore connects customer-facing policy to warehouse disposition and replenishment decisions.

The approach can reduce fragmentation, but it also requires consistent return status, inventory ownership, and recovery rules across channels. The holiday period will test whether the platform maintains promise accuracy and recovery value when return volume rises.

BuyWander announced a $21 million Series A led by Madrona Venture Group and Inspired Capital for its returned-goods resale operation. The Seattle-area company runs seven-day auctions for returned and overstock merchandise from retailers including Amazon, Target, Walmart, and Home Depot.

The operating bottleneck is intake: workers identify mixed items, record condition, estimate value, choose a disposition route, and create listings. BuyWander described an AI-enabled workflow that uses photos and product information to identify merchandise, pull retail valuations, and generate listing content, while keeping the warehouse worker in the loop.

BuyWander reported 400% year-over-year growth, nearly 500,000 unique items sold per month across eight fulfillment locations, and merchandise clearing at an average 75% below retail. Those are company-reported figures, but they show why structured intake can become a throughput and recovery constraint in regional liquidation facilities.

Zalando began using Sereact’s AI-powered robotic systems at CEVA Logistics facilities in Greven, Germany, and Świebodzin, Poland. The live deployment addresses fashion returns, where items can arrive disordered, damaged, or in configurations that conventional fixed-program robots do not expect.

Sereact’s Cortex physical-AI platform lets dual-arm robots identify, grasp, and sort returned fashion items without individual training for every product or a predetermined SKU list. The robotics-as-a-service model shifts the deployment toward supervised stations, exception management, and quality control.

Zalando reported that about 50% of items ordered across its markets are returned, while more than 96% of returned items are classified as ideal condition and about 98% can ultimately be offered again. The robot program is an initial phase, so the key proof points are handling rate, salvage accuracy, and reduced manual touchpoints at CEVA sites.

Fleetio expanded its AI Service Advisor for fleet maintenance, supporting more than 8,500 fleets and millions of assets, according to the company. The tool can prioritize fault codes, draft work orders, review repair costs, and close eligible issues after service.

The advisor uses connected vehicle and maintenance data to interpret faults and recommend work within fleet-defined approval limits. During a six-month open beta, Fleetio reported that assets returned to service an average of 2.5 hours sooner per repair, while fleet managers retained control over which low-risk outsourced repairs could advance automatically.

The reported result is an average across the beta rather than a universal guarantee. Fleets need to separate faster administrative processing from true reliability improvement and check whether automated approvals fit vehicle criticality, warranty, and safety rules.

IndexBox reported continued growth in the GPS tracking-device market, driven by fleet telematics and broader IoT adoption. The development reflects expanding availability of location, utilization, and vehicle-condition data across commercial transportation operations.

The technology layer can supply route positions, geofences, mileage, dwell, and asset status to transportation-management, maintenance, and safety applications. More data does not automatically create better decisions; identifiers, sampling frequency, connectivity, and ownership rules determine whether the signals can be trusted.

For logistics operators, the strategic question is whether telematics data is converted into a common event model that supports dispatch, maintenance, compliance, and customer visibility. The market forecast itself is not an operational KPI, so buyers should measure decision latency and data quality locally.

Inbound Logistics examined configurable warehouse-management systems designed for operations with changing products, customers, workflows, and service requirements. The focus is on adapting execution rules without turning each new exception into a custom software project.

Configurable WMS capabilities typically expose rules for receiving, putaway, picking, replenishment, labor, and shipping so operators can change behavior through controlled parameters. The operational discipline is to pair configuration with auditability, role permissions, test environments, and performance measurement rather than allowing ungoverned local edits.

For 3PLs, variability is a commercial fact because different accounts impose different packaging, labeling, cutoff, and billing rules. The continuous-improvement opportunity is to distinguish a reusable rule from a one-customer exception and measure the impact on throughput and error rates.