When AI Stops Thinking and Starts Doing

gpcarracedoUncategorized

AI

When AI Stops Thinking and Starts Doing

By Gabriel Pastrana·November 29, 2025·4 min read·Issue #17

AI is no longer a spectator sport. From humanoids learning through teleoperation to warehouse systems wiring themselves for agility, we’re crossing the threshold from intelligence to agency. Machines aren’t just reasoning — they’re acting, adapting, and reshaping how logistics and operations flow.

This week, we look at what happens when agents, robots, and humans become skill partners — not replacements.

🤖 Agentic AI and the Rise of the Actionable Machine

The next wave of automation won’t be about smarter chatbots — it’s about machines that decide and act autonomously within operational systems. According to McKinsey, “agentic AI” marks the shift from passive to proactive systems, capable of initiating workflows, negotiating with other agents, and adapting to live feedback.

For logistics and e-commerce, this means a new class of digital and robotic agents that can plan, coordinate, and correct — without waiting for human commands. Think of a warehouse where bots negotiate picking priorities based on delivery windows, or an AI system that reroutes stock before forecasts shift.

The concept of skill partnerships, explored in McKinsey Global Institute’s latest report, reframes this shift as collaboration, not replacement. The future warehouse isn’t human-free — it’s human-leveraged. Teams of engineers, planners, and bots share a continuous feedback loop, with humans setting intent and agents executing the details.

The industrial implications are immense. A recent $600 million raise by Physical Intelligence shows investors’ appetite for robot “foundation models” — AI systems trained to generalize motion, planning, and adaptation across use cases. Meanwhile, Flexion secured $50 million to build AI systems powering humanoids via “sim-to-real” transfer, bridging simulation learning with real-world execution.

In short, agentic systems are blurring the boundary between code and motion. The winners will be those who design for continuous co-adaptation: human intent guiding machine initiative.

🧠 Supporting Insights

🦿 Humanoids Learn Through Control, Not Autonomy

1X is redefining humanoid development by prioritizing teleoperation over full autonomy for its NEO robot. Instead of building from end-to-end AI, 1X uses human-guided control to train movement libraries. This approach mirrors the “foundation model” logic of robotics — learn from human expertise, then scale.

Is it a pioneer or just another gimmick?

🧩 Plug-in Tech Over Platform Overhauls

A DC Velocity survey shows most logistics firms now favor modular, plug-in tech tools instead of full software rewrites. The message is clear: agility beats overhaul. In warehouses, that means layering AI-driven analytics or route optimization engines on top of legacy systems — not tearing them down.

🧠 Miniaturization Drives Smarter Mobility

Miniaturization is making robots smaller, faster, and more autonomous. The Robot Report highlights how advances in sensing and compute density are reducing form factors while increasing intelligence. Expect leaner robots in last-mile and intralogistics — capable of navigating tighter spaces and more dynamic environments.

🏗️ DHL’s Strategic Expansion in the Gulf

DHL is investing $150 million in a Saudi Arabian mega-warehouse, alongside a €130 million expansion by DHL Supply Chain at the Special Integrated Logistics Zone. The move underscores the region’s emergence as a logistics hub connecting Asia, Europe, and Africa — and DHL’s long-term bet on nearshoring resilience.

🧮 The Tech-Savvy Leader’s Playbook

McKinsey urges executives to become “tech translators” — leaders who understand the strategic language of AI and automation. The takeaway: being tech-savvy isn’t coding; it’s fluency in what technology enables. Expect this skill to define the next decade of operations leadership.

⚡️ Snippets

  • Agile Robots acquires thyssenkrupp Automation Engineering, expanding into high-precision manufacturing automation.
  • The surgical robotics market is projected to double by 2029, driven by minimally invasive systems and flexible automation.
  • Advantech integrates compute with D3’s embedded sensing tech, merging edge AI with navigation for mobile robots.
  • Fabletics cuts full-store inventory checks down to 18 minutes, illustrating the next retail automation wave.
  • Sweetgreen sells its Spyce robotics kitchen unit to Wonder for $186.4M — a sign of shifting food-tech economics.
  • Parallax Worlds raises $4M to build hyper-realistic digital twins for robot stress-testing.
  • DHL reports 84% of online retailers expected a Black Friday surge, signaling renewed e-commerce resilience.
  • Starship Technologies partners with Uber Eats to deploy sidewalk delivery robots in U.S. cities.
  • Walmart tests “dynamic delivery windows” — promising real-time order flexibility down to minutes.

Smart Automation

A four-minute weekly newsletter on automation, AI, intralogistics, supply chain, and operational excellence.

Subscribe free