The value of robots must ultimately be realized in real-world work.
The robotics industry is at a critical juncture, moving from technological breakthroughs toward deployment at scale. Advances in artificial intelligence and robotics continue to expand what is possible, bringing an ever-growing range of exciting innovations. Yet as we move from technological progress to practical adoption, a more fundamental question remains: robots must demonstrate not only that they can complete a task once, but that they can perform real jobs consistently, reliably, and over the long term.
A single success demonstrates what may be possible. Sustained work demonstrates how reliable that capability is. Technology demonstrations showcase breakthroughs; real-world applications test results. Basic research, technology demonstrations, and standardized testing each have irreplaceable value. When robots enter real workplaces, they must also withstand sustained tasks, complex operating conditions, and customers' actual requirements.
Real-world applications are where working capabilities are tested. Sustained operation is how quality is assessed. Customer value is a key measure of the industry's achievements. Regardless of country, region, or technical approach, credible capabilities, lasting reliability, and verifiable value in use should be shared goals across the industry.
We call for real-world work to become a common benchmark for evaluating robot capability, intelligence, quality, and value.
To that end, XROGO puts forward the following initiative to robotics companies, AI and model developers, customers, research institutions, and industry partners around the world:
1. Honest specifications. Competition on merit.
Specifications can highlight advantages. Real capabilities must stand up to verification.
Build trust with accurate specifications and compete on real capabilities. Do not inflate specifications or exaggerate capabilities. Do not substitute a longer list of specifications for better quality. When publishing key metrics, clearly explain the test conditions, scope of applicability, and capability limits. Distinguish peak from sustained performance, and experimental results from performance delivered to customers. Encourage objective, reproducible verification of product capabilities, so that different technical approaches can compete fairly on a credible basis.
2. Test models in real-world applications.
A demonstration can show potential. Real tasks must test intelligence over time.
Respect basic research, encourage model innovation, and turn technological progress into the ability to solve real problems. When evaluating intelligent models for robotics, consider research and standardized test results alongside effectiveness, generalization, and deployment performance in real tasks. Assess the working capability of the complete system objectively, taking into account the robot hardware, control system, and operating environment. Do not substitute model size or leaderboard rankings for results in use, or a single demonstration for sustained working capability. Let applications raise the questions, let models improve through practice, and let intelligence serve real work.
3. Test quality through sustained operation.
One success does not establish long-term reliability. Passing factory acceptance does not establish quality across the full lifecycle.
Make safety a prerequisite for deployment, and put reliability to the test through sustained operation. Pay attention to hardware durability, temperature rise, retention of accuracy, and consistency across production batches. Verify the stability of models, software, and hardware working together. Assessments of service life and reliability should specify operating conditions, the sample scope, definitions of failure, statistical methods, and supporting evidence. Report conforming output, downtime caused by failures, human intervention, and maintenance effort accurately. Do not use selected clips to represent long-term performance. Maintain necessary safety risk assessments and compliance requirements. Combine standardized testing, validation in real work, and long-term operating data. Performance in the field must not replace necessary safety testing and certification.
4. Measure value by customer returns.
Robots should be worth using over the long term, not just worth trying.
Customer returns should be reflected in measurable improvements in efficiency, quality, safety, and business performance. Fully account for equipment, deployment, operation, human support, and maintenance costs. Evaluate actual results against clearly defined tasks and quality requirements. Distinguish projected returns from results already achieved. Explain the conditions under which results apply and the problems that remain unresolved. Help customers understand their investment and see the improvements, so that technological breakthroughs ultimately translate into sustainable value at work.
From a shared initiative to real action
We call on the industry to raise its standards of evaluation, and we are ready to be held to those same standards ourselves.
As robots begin to take on real work, the way we evaluate them should more closely reflect the way we evaluate a job: not just whether a task can be completed, but the quality of the result, performance over time, and the value delivered.
As a robotic labor platform serving real-world workplaces, XROGO is ready to start with itself, letting real work test the robot capabilities and service quality we provide. We will continue to advance capability validation in real settings, long-term quality evaluation, and the measurement of customer value. We invite industry partners worldwide to explore objective, transparent, and reproducible evaluation methods together.
We respect different technical approaches, intellectual property, trade secrets, and customers' data rights. We welcome companies at different stages of development to participate in ways that suit them: sharing practical experience, improving evaluation methods, and advancing the industry. XROGO seeks to initiate this effort and put it into practice, not to have one company draw conclusions on behalf of the entire industry.
The next phase of robotics must keep pushing the limits of capability while also improving the reliability of real work. Every technological breakthrough deserves respect, and every advance that stands the test of sustained work deserves recognition.
Today, XROGO launches this initiative from China and extends an open invitation to robotics companies, AI companies, research institutions, customers, and industry partners around the world: start with real tasks; build trust with accurate specifications; validate intelligence through real applications; test quality through sustained operation; and measure value by customer returns. Together, let us move the robotics industry from technological breakthroughs toward deployment at scale.
Let every technological breakthrough have the opportunity to take on a real job. Let every advance in the industry be put to the test through sustained work.
Put robots to work. Let real-world applications test intelligence.
XROGO | October 2026