Company Dispatch / AgriFood AI
FloVision
FloVision
Solutions
The AI that stands on the line, watches every piece of meat go by, and tells the plant exactly where the money is leaking out.
The AI that stands on the line, watches every piece of meat go by, and tells the plant exactly where the money is leaking out.
Walk into a beef or poultry plant FloVision works with and you might not notice the thing that changed. There is no new line, no robot arm, no banner announcing the future. There is a compact camera clamped above a conveyor, quietly scanning each cut with vision, depth, and a load cell as it slides past. By the time the meat reaches the next station, the system already knows what it weighs, how it was graded, whether a foreign body slipped through, and whether the last operator left yield on the table.
That is the whole pitch, and it is deceptively boring: keep your line, add our eyes, keep more of every animal. FloVision Solutions sells the unglamorous middle of the food industry - the place between the animal and the package where margin quietly evaporates - back to the people who run it. Roughly 31 people, split across the United States and Europe, now spend their days making the invisible measurable.
Above: not a tech demo, an actual processing line. The sensor does not care that the room is cold and loud.
Here is the number that keeps food executives up at night, or should: roughly one-third of everything produced is wasted. That waste costs producers north of $200 billion a year and pours greenhouse gases into the atmosphere for food nobody ever eats. Most of it is treated as a cost of doing business, which is a polite way of saying nobody was measuring it closely enough to argue.
On a protein line, waste hides in plain sight. A trimmer takes a little too much. A grade gets called by eye and called wrong. A piece that should have been recovered slides into the rejects. Each error is small. Multiplied by thousands of pieces an hour, across every shift, across every plant, it becomes the difference between a good year and a bad one. The problem was never that processors did not care. It was that they could not see.
Rian McDonnell grew up in rural Ireland, where his family worked in beef processing. He did not discover the kill floor in a pitch deck; he discovered it as a kid. The engineering came later, but the instinct - that this enormous, essential, low-margin industry was running on gut feel - came first. His co-founder, Elise Weimholt, arrived from the opposite direction: a technologist convinced that adapting the global economy to climate change meant going where the waste actually was. They decided it was the food supply chain.
Their first bet was wrong in a useful way. FloVision launched in 2020 aimed at cafeteria food waste - computer vision pointed at trays and trash. The Crunchbase page still carries the original handle, “flowaste,” like a baby photo. The pivot to meat processors was not a defeat. It was the moment the founders followed the data to where the kilograms - and the money - actually were.
Engineer raised in a rural Irish beef-processing family. Brought the ground-floor view of the problem and a stated goal that sounds absurd until you do the math: shave 1% off global greenhouse gas emissions.
Leads product development. Came to food waste through climate, not agriculture, and has been vocal about building a startup where women in tech are not the exception in the room.
FloVision's hardware is intentionally modest. There are two products, and they do not try to replace the line - they read it. The intelligence runs on edge hardware right there in the cold room, on chips like NVIDIA Jetson and Raspberry Pi, so the verdict arrives in real time instead of in next month's report.
A compact camera-and-sensor unit that integrates onto a conveyor, measuring each product and flagging defects, misgrades, and foreign materials for yield and QA analysis - without disrupting production.
A multi-camera workstation setup with dynamic laser projection that guides operators in real time, improving cut accuracy and cutting new-hire training time by up to half.
Per-piece vision, depth, and weight data rolled up into yield, quality, traceability, and staff-performance benchmarks that supervisors can actually act on mid-shift.
Skeptics are right to ask whether any of this moves the needle. FloVision's answer is a counter that keeps climbing: more than 23.2 million kilograms of food scanned across plants in three continents since 2020. On that food, customers report recovering up to 1.5% more yield - a figure that sounds trivial until you remember it lands on top of razor-thin margins, at industrial scale.
The money followed the kilograms. In July 2025, FloVision closed an $8.7M Series A led by software investor Insight Partners, with Serra Ventures, SOSV, and Rockstart joining and Insight's Madeleine Goldberg taking a board seat. It brought the company's total raised to roughly $11.6M. Under the hood, the vision models are built and deployed with tooling from Roboflow - the kind of partnership that does not make headlines but makes the product work.
It would be easy to file FloVision under cost-cutting and move on. The founders insist on a bigger frame, and for once the bigger frame is the more honest one. Food waste is not just lost margin; it is methane, land, water, and energy spent on meals that rot. Recover yield at the source and you are not only saving a processor money - you are keeping the carbon cost of that wasted food from ever being paid.
That is why Weimholt came at this from climate rather than agriculture, and why McDonnell talks in terms of a 1% cut to global emissions rather than basis points of yield. The two framings are the same lever pulled from different ends. The mission is to make the financially smart choice and the environmentally smart choice the exact same choice.
The Series A money is pointed outward: expansion across the Americas, Europe, Australia, and Asia, and hiring in engineering, AI/ML, and sales. The bet is that every protein line on earth is a line FloVision could eventually stand on. There are real competitors in industrial food vision - Marel, Baader, P&P Optica and others - and being a 31-person company chasing a global industry is its own kind of tension. FloVision's wager is that the clamp-on, keep-your-line approach beats the rip-and-replace one.
So return to that cold, loud room from the start. The conveyor still runs. The cuts still slide past. But now the line answers back - every piece weighed, graded, and accounted for the instant it passes. The waste that used to be a cost of doing business has become a number on a screen, and a number on a screen is something you can finally argue with. That is the change FloVision is selling: not a new factory, just a factory that can see.