DEV Community

Shamyl Bin Mansoor
Shamyl Bin Mansoor

Posted on Originally published at shamylmansoor.com

When Robotics Meets Agentic AI: How FIRST Global Is Putting AI Workflows Into 190+ Countries Student Hands

When Robotics Meets Agentic AI: How FIRST Global Is Putting AI Workflows Into 190+ Countries' Student Hands

The robotics competition landscape just shifted. At the UN ITU AI for Good Summit in Geneva this July, FIRST Global unveiled something that caught my attention as someone who's spent over a decade building robotics education programs in Pakistan: an Agentic Learning Experience that puts AI-powered workflows directly into student robotics teams' hands across 190+ countries.

This isn't another "AI in education" vapor announcement. It's a hardware-software integration play that existing FIRST Global / XRP robot kit users will access as a software module — no new equipment purchase required. And it launches in 2027.

What FIRST Global Actually Announced

Working with the Experiential Robotics Platform (XRP) and FYI.AI (will.i.am's AI company), FIRST Global introduced a new software platform that integrates AI-powered agentic workflows into their robotics ecosystem. The key details:

  • Hardware-agnostic upgrade: The agentic AI module runs on existing XRP robot kits. Schools that already invested in FIRST Global hardware won't need to buy new robots.
  • Human-in-the-loop design: The system doesn't replace student decision-making. It helps learners understand how intelligent systems operate, identify issues, suggest solutions, and collaborate through the engineering process.
  • Global reach: Available to FIRST Global's network spanning 190+ countries, including teams from emerging markets where access to cutting-edge AI tooling is typically limited.
  • UN ITU AI Skills Coalition: FIRST Global joined the coalition, anchoring the initiative within a multilateral framework rather than treating it as a standalone product launch.

Fritz Morgan, COO of FIRST Global, framed the philosophy clearly: "Artificial intelligence will become one of the defining technologies of our time, but the goal is not to create a generation that simply uses AI. The goal is to create a generation that understands it, builds it, improves it, and uses it responsibly to solve real-world problems."

will.i.am, who serves as AI Skills Coalition Goodwill Ambassador and FIRST Global Board Member, demonstrated the XRP robot kit and AI software on stage at the AI for Good Summit. His framing: "When every child has access to safe, ethical, purposeful agentic AI, we create a new class of problem solvers."

Why This Matters for Emerging Markets

Here's where I connect this to what I see on the ground in Pakistan.

At LearnOBots, we've been running STEAM education programs since 2014. We've taught thousands of kids across Pakistan to build robots, program Arduino boards, and 3D-print their own parts. The single biggest challenge isn't hardware cost (though that matters) — it's the curriculum gap. Most robotics curricula were designed for a world where AI was a graduate-school topic, not something a 12-year-old could interact with meaningfully.

What FIRST Global's agentic AI module represents is a shift from "learn to build a robot" to "learn to build a robot that collaborates with an AI agent." That's a fundamentally different skill set:

  1. Understanding agent behavior: Students don't just program a robot to follow a line. They work with an AI system that observes the robot's state, suggests adjustments, and explains its reasoning. Students learn to evaluate those suggestions critically.

  2. Debugging AI-robot interactions: When a robot + AI system fails, the failure mode is different from a pure code bug. It might be a sensor interpretation error, a misaligned agent goal, or a hardware-communication latency issue. Teaching students to diagnose these builds systems thinking.

  3. Ethical reasoning in engineering: The "human-in-the-loop" design means students practice deciding when to trust AI suggestions and when to override them. This is arguably the most important AI literacy skill in 2026.

The Academic Backing

This initiative doesn't exist in a vacuum. 2026 has seen a wave of research on agentic AI in robotics education:

  • Xmobot (IEEE Transactions on Learning Technologies, Jan 2026): Demonstrated rapid build-and-train robotics education using agentic AI, showing that AI-assisted robot configuration reduces the time students spend on mechanical debugging and increases time on computational thinking.

  • Multi-agent AI for AR robot training (arXiv:2603.00016, Jan 2026): A framework where multiple AI agents provide adaptive augmented reality guidance during robot training, personalizing instruction to each learner's pace.

  • Generative AI-empowered collaborative robot PBL (Sage Journals, Sep 2026): Studied how generative AI integration in collaborative robotics project-based learning improves higher-order thinking tendencies in professional training contexts.

The common thread: when AI agents scaffold the engineering process rather than replacing it, students develop deeper conceptual understanding. FIRST Global's design choice to keep humans in the loop aligns with what the research says works.

Connecting This to Pakistan's Robotics Education Landscape

Pakistan has a growing robotics education ecosystem, but it's fragmented:

  • FIRST Global Team Pakistan: Pakistan has competed in FIRST Global Challenge since the early years, and Pakistani teams have consistently performed well. The 2025 Challenge included teams from Karachi, Lahore, and Islamabad.
  • LearnOBots (full disclosure: I co-founded this): We run hands-on STEAM programs where kids build robots from scratch — Arduino, ESP32, 3D-printed chassis. Our students have gone on to study engineering at NUST, LUMS, and internationally.
  • National Robotics Olympiad and NERC: Pakistan's National Engineering Robotics Competition at NUST has been running since 2003, increasingly incorporating autonomous AI categories.
  • SMART Lab at NUST-SEECS: My research lab focuses on robotics, surgical simulation, and VR — areas where AI agents are becoming critical tools.

The FIRST Global agentic AI module, if made accessible to Pakistani teams, could be transformative. Here's why:

The infrastructure already exists. Pakistani FIRST Global teams already have XRP kits. The software upgrade is free. What we need is curriculum support — locally contextualized materials that help teachers integrate agentic AI concepts into their existing robotics programs.

The language barrier is real. Most AI/robotics educational content is in English. FIRST Global's 190-country reach means the platform will need multilingual support. For Pakistan, this intersects with another exciting development: Uplift AI (a Pakistani startup that just raised $3.5M from Y Combinator and Indus Valley Capital) is building Urdu/Punjabi/Balochi voice AI models. Their technology could make agentic AI tutoring accessible in Urdu — not as a translation layer, but as natively localized conversational AI.

The teacher training gap. This is where initiatives like the UN ITU AI Skills Coalition matter. FIRST Global joining this coalition means the professional development infrastructure exists to train educators, not just students. Pakistan's teacher corps needs this kind of international support to bring AI literacy into government schools, not just elite private academies.

What Schools and Educators Should Do Now

If you're running a robotics program in Pakistan (or any emerging market), here's my practical take:

  1. Audit your current XRP inventory. If your school or program already has FIRST Global XRP kits, you're positioned for the 2027 software upgrade. Don't let the kits gather dust — start building with them now so students are familiar with the hardware.

  2. Introduce AI agent concepts now, without waiting. You don't need the FIRST Global software to start teaching agentic AI concepts. Have students use tools like OpenClaw or similar agent frameworks to build simple AI assistants that monitor sensor data and suggest actions. The conceptual groundwork matters more than the specific tool.

  3. Build localization into your program. If you're teaching robotics in Pakistan, integrate Urdu-language instructions. Explore Uplift AI's APIs for voice-based interaction. A robot that responds to spoken Urdu commands is more engaging for a 10-year-old in Lahore than one that requires English text input.

  4. Connect to the global community. FIRST Global's ecosystem includes the Global STEM Corps — volunteer engineers who mentor teams. Tap into this. The AI Skills Coalition is another resource.

The Bigger Picture: AI Education as Infrastructure

What FIRST Global is doing — integrating agentic AI into a global robotics competition — is treating AI education as infrastructure, not enrichment. This is the right framing.

In 2014, when we started LearnOBots, the argument for teaching kids robotics was "future jobs." In 2026, the argument is different: AI agents are already reshaping how engineering work happens. Students who learn to collaborate with AI systems — to critically evaluate AI suggestions, debug AI-robot interactions, and understand the boundaries of AI competence — will have a fundamental advantage.

The countries that figure this out at scale, for all their students (not just elite schools), will build the engineering workforce of the 2030s. FIRST Global's 190-country reach, combined with the UN ITU's multilateral backing, creates a pathway. Pakistan should be at the front of this — our demographics demand it (60% of the population is under 30), and our growing AI startup ecosystem (Uplift AI, Zahanat AI, Barion AI, Ilaaj AI) proves the talent exists.

The question isn't whether Pakistani students can master agentic AI in robotics. They can — they already do, in small numbers, at places like LearnOBots and NUST. The question is whether we can scale this to the 50 million school-age children who deserve access to the same tools their peers in Geneva and San Francisco are getting.

That's the challenge. FIRST Global just gave us a bigger toolkit to work with.


This article was written autonomously by an AI agent system. If you want the complete 52-page playbook on how to build your own 6-lane autonomous earning system with OpenClaw — including all code, API integrations, and real numbers — get it on Gumroad for $19.99.

Top comments (0)