Like many of you, I returned to the Automate show this year. Unlike many of you, I was returning for the first time in five years — and, boy howdy, what a difference half a decade makes! 

Fun fact: The recurring theme at Automate 2019 was that the then biannual event was saying a permanent goodbye to Chicago. At the time, the argument went, it made more sense for the largest robotics, vision, motion, and advanced automation show to be held in the country’s automotive manufacturing epicenter: Detroit. 

Clearly, plans changed. All things considered, we’re all lucky that the COVID-19 pandemic didn’t permanently turn Automate into a virtual online event. 

Five years ago, Automate attracted just over 20,000 attendees to its show floor, which featured 500 exhibitors. To underscore how quickly the automation industry has bounced back from (or perhaps was fueled by) the global pandemic, Automate 2024 attendance set a new record, with 42,895 registrants viewing over 840 exhibits.  

A disproportionately high number of the exhibitors I spoke to this year said they had reached their target show totals for new leads before Day 1 had closed.  

My half-decade away from machine vision and automation highlighted how quickly the industry has evolved since I had last walked Automate’s floor. 

AMRs redefine “dynamic navigation” 

In 2019, autonomous mobile robots were just gaining traction thanks to their ability to digress from a fixed path and dynamically navigate industrial environments. But “dynamic” then meant an AMR could steer itself around a fully loaded pallet that had been randomly planted in its assigned path. Somewhere along the way, dynamic navigation has come to mean the ability to actively detect, avoid, or even interact with moving objects, like people or other bots.  

Further underscoring how quickly AMR adoption is rolling forward, Neura and other exhibitors like Omron, Zebra Technologies, and LexxPluss were framing deployments in terms of scalability by highlighting fleet management solutions.  

Don’t call them collaborative robots 

Cobot Automate 2024Five years ago, cobots were still new enough that everyone felt compelled to keep spelling out “collaborative robots,” and that was sufficient to distinguish them from industrial robots that were larger, harder to program, and comparatively dangerous to operate alongside human workers.  

It didn’t take long at this year’s show to realize that cobots’ “ease of use” pitch had faded a bit from overuse. Though there were cobots on the 2019 show floor with payloads reaching as high as 15 kg, their number has grown and the ceiling has clearly risen. In fact, with 20- and 30-kg payload cobots on the floor, 15 kg seems more like the median payload range now.  

Additionally, the cobots I saw at Automate 2024 had become more physically flexible — in the case of Kassow Robots’ 7-axis machines — and yet engineered to perform particular tasks from palletizing to screwdriving. In fact, there were so many application-specific demos on the show floor that one colleague quipped Automate had effectively become a job fair for cobots. 

 AI’s black box problem 

Deep learning was all the buzz at Automate 2019. At the time, it seemed revolutionary for a machine vision system to classify objects within an image or to inspect the components of a car door while simultaneously confirming correct assembly. All users needed to do was input thousands of images to train a neural network to recognize “good” vs “bad” results, and the computer would do the rest. So cool. 

Since then, the launch of ChatGPT and large language learning models (LLMs) have made headlines. While LLMs don’t have significant impact on machine vision or automation, they did refocus the collective imagination on how AI writ large might streamline programming, help machines teach themselves, and generally create more intelligent and adaptable automation. 

Having missed five years of that rapid evolution, I could only sense rather than quantify the impact of AI on the exhibitions at Automate 2024 — aside from seeing more of them trying to associate themselves with NVIDIA. In some ways, the topic of AI seemed to engender the same fundamental question in 2024 as deep learning did in 2019: Sure, it will clearly enable more adaptable and functional system designs, but where to start? 

I won’t name names, but some of the claims around the AI-powered robotic systems seemed “powered” more by enhanced sensor input than computational intelligence. Granted, AI might have played a role in fusing sensor data or interpreting input coming from dynamic environments. As I said, it was easier to sense the impact of AI than to isolate the value it was rendering from inside some black box. It was easier, in some ways, to grasp the immediate impact of deep learning. 

Maybe that will change by the time Automate convenes in Detroit next year. Considering all that transpired since Automate 2019 — within and beyond this industry — I hesitate to speculate what might unfold in the next 12 months. But one thing is for certain: I look forward to rejoining the exhibitors and crowds at future shows.