IntBot José Arrives as Airport Humanoid Concierge

José the humanoid at San José International Airport greets travelers in 50 languages and actually knows where to send them.

San José’s Mineta International Airport has welcomed a new multilingual concierge, a humanoid named José built by IntBot Inc. Stationed in Terminal B, the robot is designed to answer questions, provide real-time flight information, and guide passengers through a busy terminal, signaling a growing push to blend hospitality with automated assistance in high-traffic public spaces. Engineering documentation shows José relies on real-world perception and contextual reasoning to interpret traveler requests and translate information on the fly. The device was also on hand at NVIDIA’s GTC 2026, where it demonstrated greeting visitors and staffing a help desk, albeit in a disciplined, wired-for-continuous-ops setup.

From a usability standpoint, José’s most striking trait is scope: it’s intended to operate in a crowded, noisy environment and respond in more than 50 languages. The airport’s management framed the deployment as a concrete step in using emerging technologies to improve daily services while reinforcing San Jose’s tech-forward identity. The technical specifications reveal a focus on interaction quality and situational awareness rather than brute lifting power or rapid, gymnastic dexterity.

Yet certain specifics remain murky. DOF counts (degrees of freedom) and payload capacity for José have not been disclosed publicly, making it hard to gauge how many independent joints or servos power its hands, torso, and locomotion—and what weight it can safely handle. The same goes for power logistics and runtime: the GTC demonstration described José as wired for continuous operation, which implies a tethered power or data setup during an event, but there’s no confirmed battery specification for untethered airport duty. In other words, how long José can roam the terminal on its own charge remains unknown, a critical factor for a device designed to work in a continuous, customer-facing role.

Technology readiness, in practical terms, sits near field-ready for José. Deploying in an operational public space marks a higher TRL (Technology Readiness Level) than many lab demos, with real-time flight information and multilingual dialogue tested in a live environment. That said, the airport setting also exposes the robot to real-world edge cases—flight delays, gate changes, late arrivals, and multilingual pings that may require rapid data refresh and network resilience. The result is a mixed picture: a convincing system demonstration in situ, but still contingent on back-end data pipelines and ongoing maintenance to stay accurate.

In terms of improvements, José points toward a broader trend: automata that can operate in human spaces without tripping over social nuance. The previous generation of public-facing robots often traded conversational breadth for reliability, or conversely, boasted strong scripts with weak perception. José’s blend of “real-world perception and contextual reasoning” suggests progress in parsing intent from natural dialogue, then marrying it to live information. The practical outcome is a concierge that can pivot from a simple directions prompt to a flight-status check in multiple languages, in real time.

Industry context helps frame José’s arrival. Across the robotics space, a parallel thread is the push to fuse humanoid platforms with foundation-model AI to drive smarter, more autonomous behavior. Agile Robots, for instance, disclosed plans to deploy Google DeepMind’s Gemini Robotics models on its humanoid solutions to lift performance in industrial settings, aiming to scale safe, autonomous collaboration with humans. That collaboration, while not a field deployment in public spaces, underscores a broader trajectory: base hardware capable of cooperative work with advanced AI models, moved from lab benches toward real-world operation. The Agile ONE family—alongside its arms and payload-appropriate variants—illustrates the diversity of capability sought in this generation of humanoids, even as exact DOF/payload specs remain case-by-case.

What to watch next, practically speaking:

Runtime and autonomy: expect follow-up disclosures on battery packs, charging regimes, and how long José can operate between charges in a terminal environment.

Reliability in busy spaces: look for metrics on speech recognition accuracy in noisy airports and error-handling when live data feeds lag or update slowly.

Data governance and safety: public deployments demand robust privacy and safety controls as the robot handles traveler information and navigates crowds.

AI-augmentation pathways: the Agile Robots–DeepMind collaboration signals the industry’s interest in embedding foundation-model reasoning into humanoids, potentially improving decision-making under uncertainty.

The takeaway is clear: José is a tangible step from demonstration to operational service, but critical technical details—DOF, payload, untethered power, and long-run reliability—will determine whether this model becomes a repeatable, scalable template for airports worldwide.

Sources

IntBot humanoid robot greets visitors to San Jose Airport

Agile Robots to deploy Google DeepMind foundation models on its humanoid