The Robot Is Ready. Are We? Mark Kennedy on Getting the Doing Done

Mark Kennedy introduced himself by explaining that he is trying to keep our infrastructure working so we don't end up paying $50 for a tomato.

That's a pretty good way to explain why autonomy matters.

Mark works on off-road autonomy in North America for SICK. His world includes the sensors that help machines understand where they are, what's around them, and whether they're doing what they're supposed to do. We talked about agriculture, construction, safety, AI, and the gap between a technology demonstration and something people actually trust enough to use.

We also talked about his band opening for Alabama.

As one does when discussing industrial sensors.

The field did not read your project plan

Mark walked through a useful way to think about different operating environments. Inside a building, there can be a tremendous amount happening, but the walls generally stay where you left them. The floor doesn't usually decide to become a ditch.

Outside, the situation changes.

A construction site gets dug up. Vegetation grows. An obstacle appears. A deer walks into the area. The machine has to make sense of what's happening around it while still doing its job.

Mark describes SICK's role as providing the senses for those machines. That distinction matters. A sensor is part of the system that helps a vehicle perceive its surroundings. It isn't, by itself, the whole autonomous operation.

SICK's own mobile machine overview describes this work across agriculture, mining, forestry, and municipal vehicles, using technologies including LiDAR, radar, cameras, and ultrasonic sensing.

But the part I kept coming back to was why anyone needs this in the first place. Mark described growers struggling to find enough people to get the work done. In the situations he sees, the conversation isn't simply about replacing somebody. It's about work that isn't getting done because the people aren't available.

That's a very different starting point from buying a robot because the demo looked cool.

My lawn already taught me this lesson

I brought up my own experience with an autonomous mower. I used one. Now I pay a person to mow because, in my situation, that ended up making more financial sense.

I love technology. That does not mean technology automatically wins every argument.

Mark's response was essentially: don't make the solution more technologically complicated than the problem requires.

That is a lesson we could probably print on the outside of a few product boxes.

What are you trying to accomplish? What does the job cost today? What would have to improve to justify changing it?

Those questions should come before we start comparing sensors, platforms, or how impressive something looks driving across a trade show floor.

The machine isn't the only thing that needs training

One of Mark's central arguments was that, for many of the applications he's working on, technology is no longer the primary bottleneck. He sees confidence, education, and adoption as major obstacles.

That is his assessment of the market, and it led directly into something that surprised me at SICK's Robotics Xchange Summit.

I had been invited to speak about agentic workflows. Going in, I assumed an audience working with robotics and autonomous vehicles would already be familiar with the software agents available on the operational side of the business.

Then I started talking to people.

Some of the people building these machines weren't familiar with what was available for coordinating work, handling information, or connecting business processes. I had to adjust my assumptions.

That wasn't a criticism of the audience. It was a correction to my own thinking.

Being deeply knowledgeable about one part of technology doesn't mean you have time to follow everything happening in every other part. I spend a lot of time working with agents, and the digital twin demonstrations at that same event changed how I thought about what was possible.

We were all there to learn something.

If we want adoption, we have to make room for that. We can't hand somebody a tool, tell them it's intuitive, and count that as change management.

At some point, the AI has to leave the laptop

Mark used a phrase I liked: “laptop experiments.”

That's a familiar stage. You try a tool. It gives you an interesting answer. Maybe it saves some time. Then you close the window, and the rest of the business continues exactly as it did before.

The next step is connecting that capability to work that actually needs to happen.

For example, what if information from a machine could help the service operation prepare earlier? What if the right parts could be identified before someone was standing at a counter trying to describe the problem? What if operational planning could use information that already exists but never reaches the person who needs it?

Those are possibilities to design for, not a claim that every machine already does them.

And yes, fixing a spreadsheet is useful. I spend plenty of time in that world. But connecting what a machine knows to what a business does next is where this conversation gets particularly interesting to me.

Try to break it before the customer does

Our discussion of digital twins came back to a very practical question: what could go wrong?

Mark sees a strong opportunity for AI to help teams explore more failure scenarios before a product goes into the field. A simulation can give engineers a place to investigate conditions and combinations they might not have thought to test.

The value isn't declaring the model correct and going home. It's finding another question worth asking, another edge case worth investigating, or another design decision worth revisiting.

We both came back to human judgment and physical reality. Simulated results still need scrutiny. Someone still has to decide what they mean for the actual machine.

For me, the demonstrations at the summit made this much more concrete. I left thinking about the information those environments could provide and how it might connect to the operational work I do with customers.

That's why getting out of your normal corner of the industry matters. Sometimes you don't know which question to ask until somebody shows you something you haven't seen.

People still need to get home

The safety part of this conversation matters to both of us.

We talked about our involvement with AEM and the value of bringing people together across equipment, technology, and product safety. Mark emphasized the need for shared expectations and standards as the industry develops.

For me, the point of that work is very simple. The person operating the equipment, working beside it, or walking through the site needs to get home to their family.

That's what I appreciate about the product safety community. The work has technical and legal dimensions, but the reason for doing it is human.

The networking matters, too. Mark talked about building a group of people you can call when you need help. I told him that I love being the person somebody calls to ask, “Who should I talk to about this?”

Most of those calls aren't business for me. They're a chance to connect somebody with the right person.

That relationship has value long before there's a purchase order attached to it.

And then we got to the guitars

Mark describes himself as a “technical Forrest Gump,” which becomes easier to understand as he tells his stories.

He grew up around music, returned to guitar with renewed enthusiasm in high school, and, as he tells it, ended up in a band whose first performance was opening for Alabama in front of thousands of people.

My public speaking nerves suddenly seem very reasonable.

His interest in guitars and amplifiers also helped lead him into electrical engineering. He wanted to understand how the things he loved actually worked.

We talked about my kids, music, and making silly songs with AI. My daughter isn't a fan of that particular activity. My boys and I have had fun with it.

Mark made room for both sides. There's something special about people sitting together with a guitar. There is also value in tools that let someone explore creativity when they don't have the time or ability to master an instrument.

I appreciate that answer. Curiosity doesn't have to arrive through the same door for everybody.

Start with the job

Near the end, Mark brought us back to the question that ran through the whole conversation: what are you trying to do?

Before the sensor. Before the agent. Before the autonomous vehicle.

What does a good outcome look like, and what would help you get there?

Sometimes the answer will involve sophisticated technology. Sometimes it will involve better education, a different process, or calling somebody who already understands the problem.

And sometimes, apparently, it involves paying someone to mow your lawn.

Thanks to Mark for the conversation, for helping bring people together through SICK and AEM, and for reminding us that the point is to get the doing done.

https://www.linkedin.com/in/markkennedyautonomyeconomy/ https://www.sick.com/us/en/catalog/products/training/c/g568262