The Smartest AI Strategy in Hardware Belongs to a Tractor Company

Image by ChatGPT

John Deere is one of the most advanced AI companies on earth. Vision systems that tell a weed from a crop. Autonomous machines. Data streaming from millions of acres.

And they still call themselves the green tractor company.

Calling themselves the green tractor company is the entire strategy, and it’s the part everyone misses. Deere has more AI than most companies in Silicon Valley, and they never once forgot who they are at their core. They didn’t set up a separate AI business or chase a new identity. They put the intelligence into the work they had already done.

Every season, Deere’s machines harvest data the same way they harvest crops. That data trains the next season’s AI, which makes the next season’s equipment better, which throws off better data the season after that. Run that loop for a decade, and no competitor catches up without their own decade of fields, machines, and seasons behind them. The moat is years of data nobody else has, and the algorithms are downstream of it.

And it didn’t stay in one machine. Deere took the know-how buried in its equipment, its agronomists, and its dealers, digitized it, and built it into software that runs across the whole company. Expertise stopped living only in people’s heads. It became a layer that the entire operation runs on.

The long-game version of the same move

Rolls-Royce ran this play sixty years before AI was a phrase anyone used. In 1962, they stopped selling jet engines and started selling flight hours. You pay for the time the engine runs. Then, decade after decade, they wired every engine with sensors and layered intelligence on top. The engines got smarter the longer they flew. Rolls-Royce is still, today, an engine company, but the engines now pay them every hour they’re in the air.

The question on every manufacturer’s desk

Every manufacturer is being told to become an AI company. The harder and more useful question is where the AI actually goes. Inside the work you already do, or off in a corner as a separate thing that calls itself the future.

If you make networking gear, your switches and access points are Deere’s tractors. They’re the core of the business, and they’re what you’ve spent decades getting right. You’re not going to stop being that, and you shouldn’t want to. The question is whether the AI you build this year makes those products better, or whether it sits beside them in its own building with its own culture, telling itself a different story about what the company is.

The expensive way to get it wrong

A decade ago, General Electric decided its future was software. The CEO announced GE would become a top-ten software company. They built an industrial software platform, stood it up as a separate business in a separate building with its own culture, hired thousands of engineers, and set out to sell that platform to everyone. The investment ran into the billions.

By 2018, it was being broken up and sold for parts.

GE built the software beside the company instead of inside it, detached from the machines that were the actual reason GE mattered. The new group never absorbed the operational expertise it needed, and the core never trusted what it built. The losses showed up on the balance sheet years later, but the structural mistake was made the day they put the software in a separate building.

What the Deere move looks like inside a networking company

In our AI Labs, we run exactly the play Deere ran. One level in. We find the expertise locked inside our best people and our hardest-won processes, digitize it, and build it into solutions that run across the whole operation.

Take the kind of call our best people make a hundred times a quarter — the read that says which opportunities are real, which look promising but aren’t, which look small but will compound. That judgment lived in a handful of heads. We took it apart, digitized the signals it ran on, and built it into a system that scores every situation the way the best of them would.

When the instinct our best people spent twenty years earning stops living in one person’s head, a new hire can start their first day with it already in hand. The judgment that once belonged to one expert becomes something the whole company draws on. The work gets faster, the answers get sharper, and the value we create for each other and our customers compounds every month rather than resetting whenever someone retires or teams move.

That’s not AI replacing people. It’s AI taking the best of what our people already know and lifting everyone with it.

It is the Deere Operations Center, run inside a networking company. For any manufacturer asking where AI belongs, that’s the soundest bet on the board.

The temptation worth resisting

The trap is seductive right now, and it’s making a comeback, because it looks like progress.

The pressure is to make AI a separate thing. Stand up an initiative in the corner. Buy a stack of disconnected tools. Announce you’re an “AI company” and chase an identity that photographs well on a slide. It feels like thought leadership, but it’s the GE move at any scale. Loud, expensive, and detached from the only thing that ever made you matter.

Build AI that makes the work you already do better. The kind that lives inside the tractor, not bolted to the side of it.

The close

You don’t need to become an AI company. You need to become the most formidable version of the one you already are, where the best of what your people know runs through everything you do.

Deere ran that play in agriculture and built a moat no Silicon Valley company can copy. Rolls-Royce ran it for sixty years and turned an engine business into something that pays them every hour an engine flies.

The decision facing any manufacturer this year is whether the AI you build lives within the work or alongside it. Make that call deliberately. The companies that get it right will spend the next decade compounding while everyone else explains why their AI division is being wound down.

Joe Flynn is VP of AI Strategy at RUCKUS Networks, where he leads the company’s Operational AI. His focus is the AI that runs a company, not the AI it sells. His AI Labs team builds and runs a hybrid agentic platform on self-hosted fine-tuned models. Before RUCKUS he was a fintech CEO; he built a workflow-automation company, patented the core technology, and sold it, back before any of this got called AI. He uses AI as a tool, not a ghostwriter, so the ideas and the words here are his own. He writes for business leaders who want AI that actually ships.