The Automated Dispatcher vs. The Empowered Planner: Why Agentic AI Is the New “Best Buddy” for Manufacturing Executives

George Ninikas, SVP of Sales and Accounts and Supply Chain Planning – Americas for ORTEC, examines why the future of manufacturing artificial intelligence lies in supporting experienced planners with faster recommendations, continuous monitoring, and intelligent scenario analysis – not replacing human expertise.

Manufacturing executives are not short on enthusiasm for artificial intelligence (AI), but they are short on certainty about how to deploy it where it counts.  

A recent poll of supply chain and logistics professionals found that only 29 percent are currently deploying AI to support real, operational decisions. The largest group, 37 percent, describe themselves as actively exploring how AI might help, testing pilots and trying to separate genuine capability from marketing language. 

That gap matters for plant and operations leaders balancing tight labour markets, capacity constraints, and customer commitments all at once. It suggests the industry has moved well past asking whether AI belongs in manufacturing operations and is now wrestling with a harder question: what role should it actually play once it arrives?  

The poll also pointed to where practitioners expect the most value to materialise. 34 percent identified dynamic routing and real-time replanning as the application most likely to deliver meaningful impact in the near term, ahead of other commonly discussed use cases. Operations leaders do not want AI to simply generate a static production plan and walk away; they want a system that can respond as conditions change throughout the shift – the way the plant floor actually operates. 

MOVING PAST THE REPLACEMENT NARRATIVE  

Much of the public conversation about AI in manufacturing still centres on automation replacing human judgment and the algorithm quietly taking over the scheduler’s desk. That framing, whilst attention-grabbing, does not match what is happening inside most manufacturing planning organisations.  

A more accurate view is also more practical. Agentic AI, meaning AI systems capable of taking multi-step actions on a planner’s behalf rather than simply producing a single output, is increasingly being designed as a support system. It sits alongside the human decision maker, not in place of them. 

This distinction is not trivial. A planner managing production scheduling and outbound logistics is not solving one isolated problem. They are juggling machine capacity, workforce availability, customer delivery windows, fluctuating demand, supplier disruptions, and last-minute changes, often simultaneously.  

In one recent example, a team was using 14 planners for six hours a day to engineer delivery territories, assigning dedicated zones of delivery by driver and day of the week. Decades of work have gone into giving planners powerful optimisation engines to handle that complexity mathematically. What has been missing is an easier way to interact with those engines without needing a data science background to operate them. 

SOLVING THE PARAMETER PLAGUE  

Anyone who has worked inside an advanced planning system knows the feeling of staring at a dense configuration screen, full of buttons and fields that control behaviour in ways that are not always obvious. This condition might be called the parameter plague, and it is one of the quieter barriers to AI adoption in manufacturing operations. Powerful optimisation logic is only useful if the people who need it can actually access it. 

Agentic AI offers a practical answer to this problem. Rather than requiring a planner to manually adjust dozens of settings to test a scenario, a natural language interface lets that same planner simply describe what they want to explore. Asking a system to show what happens if a machine goes down, or to compare two production schedules, or to flag which orders are most exposed to a supplier delay, becomes a conversation rather than a technical exercise. The underlying optimisation math has not changed; what has changed is who can reach it and how quickly. 

This is also where the idea of AI as a digital co-worker becomes useful. A good colleague does not take over someone’s job, but handles the repetitive, time-consuming groundwork, surfaces relevant information at the right moment, and offers a recommendation whilst still leaving the final call to the person with context and accountability.  

Applied to manufacturing planning, that means an agentic system can monitor plant and logistics operations continuously, flag a disruption the moment it happens, propose a reassignment, and explain its reasoning, whilst the planner retains the authority to accept, adjust, or override the suggestion. 

FROM PLAN CREATION TO PLAN ANALYSIS  

This shift is changing what the planning role actually looks like day to day. Historically, planners spent much of their time building production schedules from scratch, often under significant time pressure. As agentic systems take on more of that initial construction work, the planner’s role is shifting toward analysis and judgment: reviewing automated proposals, adjusting for context what the system may not fully capture, and deciding how much autonomy to grant for routine, low-risk decisions versus how much to reserve for direct human approval. 

That balance of autonomy is not fixed. Manufacturers are increasingly able to dial it up or down based on comfort level and track record. A single recurring task, such as reassigning a production run when a machine or team member is unavailable, might begin as a fully human-reviewed decision and gradually become automated once the system has demonstrated consistent, reliable judgment in that specific scenario. 

WHAT THIS MEANS FOR MANUFACTURING EXECUTIVES  

The poll data points to an industry in transition rather than one that has arrived. With the largest segment of practitioners still in an exploratory phase, and the clearest near-term demand centred on dynamic, real-time decision support rather than full automation, the opportunity for manufacturing executives is to treat agentic AI as an extension of their planning teams, and not as a replacement for them.  

The technology’s value lies less in removing people from the loop and more in giving them faster access to better information, fewer manual steps, and more time for the strategic decisions that genuinely require human judgment. That is a more grounded story than utopian automation or job displacement, and it is increasingly the one playing out on plant floors today.