Factory settings: Human plus humanoid

How robots that look like us are reshaping the workplace.

Once confined to science fiction, humanoid robots are stepping onto factory floors – not to replace workers, but to work alongside them. With the convergence of AI, robotics, spatial computing, and digital twins, enterprises now face a profound shift: automation with arms, legs, and reasoning skills. These human-shaped machines can adapt to existing environments, learn new tasks, and scale operations without disruption.

But the real breakthrough isn’t just technical, it’s collaborative. Humans, humanoids, and agentic AI systems are about to become one team. So how do we do it?

More flexibility, less disruption, better scaling.

Until recently, human-shaped robots – complete with a head, arms, and legs – were only speculative. But the first production models are now a reality from companies such as California-based Figure AI, which is building a factory expected to manufacture 12,000 humanoid robots per year. These will be game-changers for enterprises across multiple sectors – including manufacturing, life sciences, automotive, aerospace, defense, energy, utilities, and consumer products.

Robots and automation are nothing new on the factory floor, but legacy deployments involved purposebuilt machines and dedicated assembly lines. When manufacturing changes were required, a company would have to specify new robot designs and rebuild factories – which translated into significant financial investment and production disruptions.

Humanoid robots address these drawbacks by making the robot as adaptable as a human worker – capable of mimicking human gestures thanks to the rapid development of sensors and other hardware. What’s more, they are no longer just machines: humanoids are autonomous and adaptable physical reasoning agents equipped with cognitive capabilities. Artificial intelligence merges with robotics to take a physical form, making this the next big thing in AI and leading to the ultimate stage of automation on the shop floor.

Humanoids can be deployed to automate brown field operations without rebuilding, and humanoids can be easily integrated into existing industrial operations to perform undesirable, dangerous, labor-intensive, or repetitive tasks. Organizations can start small with a few humanoids working on focused activities, and then simply add more robots to scale up over time.

And when production needs change, companies will no longer have to retool. Instead, they can retrain both human employees and humanoid robots to undertake new operations in flexible workspaces.

Investigating the pathways to true convergence

That said, there’s work to be done to unlock the full potential of humanoids. These robots consist of an incredible package of sensors, controllers, motors, processors, and other hardware. At Capgemini’s dedicated AI Robotics and Experiences Lab, we’re exploring how to apply our expertise in agentic AI and LLM, computer vision, digital twins, data analytics, robotics, and sector-specific industrial processes to humanoid robots. Our goal is to help shape the technical convergence of human-centered, digital physical interactions between humans, systems of robots, AI-agents.

The hub of this convergence is a new virtual space in which most of the digital-human interactions will happen. In this space, unified and pre-trained data models, AI agents, edge AI, and digital twins merge with an interaction layer that leverages technologies such as real-time 3D and spatial computing. Data comes from an aggregation layer linking various data sources – including IS/IT/OT systems, sensors, and machines.

Agentic AI and physical AI then enable autonomous digital-physical interactions, with virtual AI-agents being able to act in the real world by interacting with AI agents housed in robots.

The challenge is how to best leverage technology enablers to unlock the productivity gains of humanoid robots with minimal operational disruption. To address this, Capgemini is adding agentic AI-enabled decision-making capabilities to a network of interconnected digital twins that replicate a physical industrial environment. These enhancements apply several key concepts in robotics and AI research:

· Vision language action models, which enable agents to understand and execute natural language instructions in a visual environment

· Reinforcement learning and simulation to real transfer learning, which enable the sharing of knowledge from various data sources, such as videos or motion capture, to train in virtual environments and then translate into real-world scenarios

· Teletraining, which enables a human to remotely control the agent to demonstrate desired actions

· System 1 and system 2 reasoning models, which are dual-process architectures that mimic a human brain’s ability to both generate fast, reactive responses and engage in slower, deliberate planning.

Collaboration in action

On the factory floor of the very near future, Capgemini envisions humans, AI-robots, and multi-agent systems working as a team to enhance operational efficiency, precision, and safety. Each participant will make important contributions to this, leveraging their unique abilities.

Human operators will bring their expertise and adaptability to bear upon the production process. Using spatial computing, they’ll supply guidance and oversight. They’ll also handle complex tasks that require fine motor skills and decision-making. And because humans can quickly identify and address unexpected issues, they’ll provide a level of problem solving expertise that enhances automated systems.

Humanoid robots will ensure precise handling and placement of parts, reducing the risk of damage and ensuring consistency. They can also perform repetitive tasks quickly and accurately, increasing throughput and reducing cycle times. And by taking over potentially hazardous tasks, robots will enhance workplace safety by reducing the risk of injury to human operators.

Agentic systems will continuously monitor the operational processes and environment to ensure each step is executed correctly. They will identify deviations, alert the team, and suggest corrective actions to help maintain workflow integrity and prevent errors. They can also make real-time decisions about whether operations require additional action or reinforced quality inspection process. Eventually, robots driven by virtual agents will be able to perform some of these actions.

Hybrid workforces will address pressing challenges

Industrial organizations today face pressing global challenges, including the need to contain costs, attract and retain talent through better employee experiences, and improve sustainability. Emerging solutions that build upon ongoing digital transformations – such as a properly designed and executed strategy to enhance workforces with humanoid robots – are essential to address these pressures. Companies that move first stand to gain the greatest rewards.

Start innovating now –

Establish your foundation. Leading companies are already undertaking complete digital transformations of their operations. This is essential for deploying humanoid robots – so make completing these transformations a priority.

Don’t delay. Humanoid robots are here today, and will soon be deployed in real-world environments – across multiple industrial sectors – to provide early adopters with significant competitive advantages. Enterprises can’t afford to wait.

Support, don’t replace, humans. The best outcomes will be achieved if humanoids are deployed as part of teams that also include human workers and multi-agent systems. Now is the time to determine how each team member’s strengths can best be leveraged.

Start innovating now

Establish your foundation
Leading companies are already undertaking complete digital transformations of their operations. This is essential for deploying humanoid robots – so make completing these transformations a priority.

Don’t delay
Humanoid robots are here today and will soon be deployed in real-world environments – across multiple industrial sectors – to provide early adopters with significant competitive advantages.

Support, don’t replace, humans
The best outcomes will be achieved if humanoids are deployed as part of teams that also include human workers and multi-agent systems. Now is the time to determine how each team member’s strengths can best be leveraged

Interesting read? Capgemini’s Innovation publication, Data-powered Innovation Review – Wave 10 features more such captivating innovation articles with contributions from leading experts from Capgemini. Explore the transformative potential of generative AI, data platforms, and sustainability-driven tech. Find all previous Waves here.  Find all previous Waves here.