At Home Delivery World 2026, HERE Technologies showcased AI-powered "last meter" guidance that moves delivery optimization beyond routing a truck to a curb. The system focuses on what happens after the vehicle stops-identifying the right parking spot, the fastest walking path to a building entrance, and the correct access point inside large complexes-helping logistics teams squeeze wasted seconds out of every stop as fulfillment windows tighten.
Moving beyond rooftop navigation
Traditional routing ends at a street address, but that is rarely where the real delivery work finishes. In dense urban areas, apartment blocks, hospitals, and campuses, the final handoff can burn minutes that cascade across a driver's entire shift. HERE's new capability collects positioning and sensor data from handheld devices and in-vehicle systems to learn how deliveries actually get completed on the ground, rather than relying solely on planned routes.
Bart Coppelmans, senior director of product management at HERE Technologies, described the feedback loop. "What we basically do is deploy a client-side on that device and then it's automatically in the background collecting the trace," he said. Over repeated deliveries, the platform distinguishes between traffic stops, parking locations, walking paths, and entrances, then surfaces practical recommendations without locking drivers into rigid instructions. "We basically give different options of how they want to configure it for their customers," Coppelmans added. "This is really tied to their operations and how much flexibility they want to give the drivers or not."
Why seconds matter in modern delivery networks
The savings per stop look small-30 seconds trimmed at one address doesn't feel dramatic. But across a route of 80 or 100 stops, that margin compounds quickly. "If it saves 30 seconds of delivery, maybe it doesn't sound that much," Coppelmans said. "But at the end of the day, maybe they've saved half an hour and now they can make another five deliveries." Pilot programs in the U.S. and Europe are now measuring whether AI-driven last-meter guidance can lift daily delivery counts without redesigning the underlying network or adding vehicles.
The core insight is not about forcing drivers into a tighter script. It's about capturing real-world execution patterns that planning software has always been blind to-and feeding that intelligence back into the routing algorithm so the next driver gets a smarter starting point.
AI still struggles with geospatial reasoning
Even as large language models advance, they consistently fail at the spatial logic that logistics depends on. Coppelmans pointed to hallucinations and unreliable outputs when models are asked to reason about truck restrictions, mandatory parking, or delivery sequences. "What we're seeing with AI … is that they don't understand geospatial," he said. "And they really also hallucinate in certain complex queries."
To bridge that gap, HERE built a location reasoning layer that grounds AI agents in verified mapping data and operational rules. The goal is to give logistics systems a more reliable foundation before they generate routing decisions, especially as the industry experiments with agentic AI that must act autonomously in the real world.
From generative AI to physical AI
The conversation in 2026 is pivoting from generative tools toward what some call physical AI-systems that operate in warehouses, on sidewalks, and inside delivery vans. HERE is already examining how its geospatial grounding could support curbside robotics and autonomous last-mile vehicles. "We're monitoring the effect of robotics on curbside robotics deliveries," Coppelmans said, noting that mapping precision and real-time location feedback will determine how effectively those machines can work in the same tight spaces that challenge human drivers today.
Why this matters for management
Last-meter inefficiency is a cost that compounds silently across every route, every depot, and every region. A 30-second improvement per stop isn't a technology demo-it's a measurable productivity lever that can add dozens of deliveries per driver each week without adding headcount or changing fleet size. For management teams evaluating where to apply AI next, the lesson is clear: the fastest returns often come from making existing execution data smarter, not from ripping out the routing engine. The AI Learning Path for Supply Chain Managers provides practical frameworks for connecting geospatial intelligence, driver feedback loops, and operational planning so that gains at the curb translate to the bottom line.
Your membership also unlocks:
