Agentic AI is transforming traditional robots into smart, adaptable machines that can think, learn, and act on their own. Experts point out that, unlike older robots that followed fixed instructions, these new systems can handle unexpected changes, work safely with humans, and improve over time.
Previously, robots could only follow fixed instructions fed by humans. They worked on predefined inputs and could not “think by themselves” or “act on their own”.
Puneet Tiwari, Chief Technology Officer, ARAPL Ltd, explained that traditional robots and Autonomous Mobile Robots (AMRs) relied on static maps, rule‐based planners, and off line programming pipelines. While they excelled as long as the real-world environment matched their programmed model exactly, even small deviations -- like a spill partially covering floor markers, or a temporarily blocked aisle, would often force the system into a predefined “fail” state, requiring human operators to re-map, re-program, or manually intervene.
Industrial robotics
“There has been increased adoption of industrial robotics across sectors for diverse applications. Traditional robots are being deployed because of their reliability and precision. However, they could not adapt to the changes in the dynamic work floor, posing safety concerns. The integration of Agentic AI allows robots to make autonomous decisions, plan, and adapt to dynamic workspace actions through learning in real time. This allows robots to perform more versatile tasks and operate effectively beyond controlled settings, requiring minimum human intervention,” noted Prateek Jain, co-founder & COO, of Addverb.
He added that with, Agentic AI, robots can perceive their environment, make real-time decisions, and adapt to new situations, allowing them to handle complex tasks while human workers focus on strategic operations. This allows them to work alongside human workers and other robots more seamlessly, reducing downtime and operational bottlenecks.
For instance, Addverb’s robots like AMRs and Robotic Sorters can now autonomously plan optimal routes, adapt to changing warehouse conditions, as well as effectively collaborate with other robots or human workers.
Similarly, in Amazon’s fulfilment centres, intelligent robots navigate complex warehouse layouts, avoid obstacles, prioritise tasks, and adapt routes without requiring direct human intervention. These robots do not just follow set paths but choose the most efficient one in real-time, operating without human intervention.
Anurag Gupta, co-founder of STEMROBO Technologies Pvt. Ltd, said, “Like humans, with Agentic AI, robots can sense changes and decide what to do next. If something blocks their way or if the task changes, they don’t stop or crash but instead think through and take action. They use sensors, smart software, and learning tools to adjust.”
Problem solvers
For example, a robot in a school lab can find a new path if something is in the way, or help another robot finish a task. The intelligence lies not just in detecting change, but in responding with purpose. “Agentic AI transforms robots from passive machines into proactive problem solvers capable of operating in uncertain and evolving conditions,” he added.
With Agentic AI, robots can also learn from experience, just like humans do. If a robot encounters a challenge, figures out a better way to complete a task, or identifies a more efficient path, it can retain that knowledge and apply it in future scenarios. This knowledge can also be shared, with one robot transferring its learning to others, creating a collaborative learning ecosystem. This peer-to-peer learning significantly reduces the burden on human programmers and trainers.
“Modern fleets employ federated and hierarchical learning, where they learn from their own experiences and share them with the group. Each robot logs interactions, which are aggregated centrally, turned into improved instructions, and broadcast back to the fleet, often overnight or even edge-to-edge.
This “collective experience” paradigm has been demonstrated in large-scale reinforcement learning deployments, where new facility layouts are mastered in days instead of weeks, and every unit benefits from each peer’s learnings,” said Tiwari.
With Agentic AI, robots are no longer just tools -- they are fast becoming intelligent collaborators that learn, adapt, and evolve alongside human beings.