In 2011, a grocery retailer in Greece was already running autonomous routing. Orders arrived by phone, fax and email, and five or six times a day the system optimized them and assigned them straight to vehicles already on the road. No planner intervened, and no AI was involved. So why is nobody operating that system today? This conversation with a host and ORTEC's George Ninikas answers that, and sets out what autonomous routing takes now.
What you will take back to your operation
A practical, vendor honest view of what autonomous routing means, and the framework to get there.
A concrete four step framework
The spine of the session, mapped so you can apply it to where your own operation sits today.
What autonomous really means
Not robots with nobody at the wheel. A practical definition where planners are pulled in only when a decision needs them.
Inside the self learning layer
Stop times learned per customer, travel times learned per driver, and driver familiarity handled as soft boundaries.
Optimization or AI
A straight answer to the question the host keeps asking, and what has to be true before you let go of the wheel.
What the session covers
Five chapters, from a system that worked in 2011 to the self learning layer that makes autonomy hold.
Autonomous routing was solved in 2011, then it stalled
A grocery retailer running same day delivery from its own stores, with time based triggers firing the optimizer five or six times a day and orders assigned straight to vehicles already on the road. No planner, no AI, and it worked. The one thing missing is the reason it could not keep pace with the business.
What autonomous routing actually means
Not an operation run by robots with nobody at the wheel. George shares a more practical definition, where complexity is removed and planners are pulled in only when a decision genuinely needs them, so one team can cover a far larger network without adding headcount.
The four steps from where you are to autonomous
The four essential steps every business takes to become autonomous, from getting daily rules and tribal knowledge out of people's heads and into data, through to a plan that recalibrates itself after every run. This is the spine of the session, and the framework to take back to your own operation.
The secret to the self learning layer
Where the compounding value sits. Stop times learned per customer rather than a flat 25 minutes for every site. Travel times learned per driver, because the same leg takes one driver 30 minutes and another 35. Driver familiarity handled as soft boundaries that expand as drivers earn them, all updating quietly before the next planning round.
Optimization or AI, and when you let go of the wheel
Is this optimization or is this AI, and why should the buyer care? Answered directly, along with what has to be true before autonomy is real: an optimizer with data it can trust, drivers with a plan they can trust, and OTIF well above where most of the industry sits today.
In conversation with a host
George Ninikas
Senior VP Sales & Accounts Americas, ORTEC
George is the SVP of Sales and Accounts for the Americas at ORTEC, leading growth initiatives, strengthening client relationships, and driving sales excellence across North and South America. Since joining ORTEC in 2010, George has advanced through a series of increasingly senior roles thanks to his ability to translate complex technologies into client-focused strategies that deliver measurable value. Previously serving as Industry Director for Retail & Wholesale, he shaped market strategy and leveraged advanced analytics to improve efficiency, profitability, and customer experience for major clients. His appointment as SVP reflects his exceptional leadership, strategic vision, and proven track record in building and sustaining high-performance teams. In addition to his corporate achievements, George holds a Ph.D. in Operations Research with a focus on vehicle routing problems and has university-level teaching experience, combining academic rigor with practical business impact.
View LinkedIn profile