There was a time when waste never got budgeted for, nor measured. Waste functioned like the utility bill nobody read before paying. It was the line item that lived two layers below “cost of goods sold” and never got a second look. The old system worked fine when a company had three locations and a guy named Pete who handled vendor calls out of a spreadsheet. That system stopped working the moment the same company had 300 locations across 40 states; each one ran a separate hauler contract, a separate pickup schedule, and a separate definition of “recycling.”
Zero waste, in the strict sense, means diverting at least 90% of material from landfill and incineration: that’s the TRUE and GBCI benchmark, not a marketing phrase, as RTS lays out. Hitting that number at a single restaurant is a logistics problem. Hitting the 90% mark across a national footprint becomes something closer to an infrastructure problem; infrastructure problems demand systems, not good intentions.
Zero waste makes the business case plainly: sustainability benchmarks and margin protection no longer operate as separate conversations. A hauler invoice that goes unaudited for five years overcharges a company for five years straight. A waste stream without standardized measurement across sites can’t be reported to investors without caveats. The companies getting this right aren’t the ones with the best intentions. Those companies are the ones that built the plumbing.
The Case for Centralized Infrastructure
Fragmentation is the default state of multi-location waste management. It costs real money in ways most finance teams never see. A regional operator in Ohio signs one contract. A regional operator in Texas signs another, at a different rate, with a different hauler, under different terms. Multiply those scattered contracts across 50 markets, and the result is a portfolio with no consistent service level agreement, no room to negotiate pricing, and no single person who knows what the company spent on waste last year without a week of phone calls.
Consolidation fixes the fragmentation problem. Moving from scattered regional contracts to a unified national procurement model gives a company one set of terms, one billing structure, and one point of accountability — an approach the industry calls waste brokerage, where a single vendor manages hauler relationships on a company’s behalf, as RTS describes its own service model (full disclosure: I’ve consulted for RTS on exactly this kind of deployment). The savings aren’t theoretical. WeWork’s portfolio-wide program stands as the case study RTS points to most often: dozens of locations, one administrative layer, and the elimination of the localized price gouging that creeps in when nobody compares invoices across markets. When a facilities director in Phoenix has no idea what the facilities director in Atlanta is paying for the same service, the hauler calls all the shots. Centralization takes it back.
There’s a second, less visible benefit: audit-ready reporting. Scope 3 emissions disclosure has moved from “nice to have” to a board-level requirement at most large companies, and waste diversion sits squarely inside that category. Inconsistent measurement leaves a company with nothing credible to report. Standardized metrics across every site, same categories, same units, same reporting cadence, let a sustainability team walk into an ESG review with numbers that hold up, as RTS’s own ESG guidance spells out. Whole Foods Market’s enterprise-wide diversion push offers a useful lesson here too: the aggressive benchmarks only worked because the underlying data stayed comparable store to store, as RTS’s case study on the program documents.
Infrastructure alone fails to reach 90% diversion. Infrastructure gets you organized. The diversion itself comes from somewhere else.
Where AI Actually Moves the Needle
For years, waste collection ran on a fixed calendar. Tuesday and Friday, every week, whether the dumpster was full or a quarter full. Haulers call this a static route. It means a lot of wasted diesel. Trucks were sent to pick up air.
Ultrasonic sensors changed that math. Devices like Pello — an RTS sensor product I’ve worked with directly, worth flagging upfront — sit inside a dumpster or compactor and measure fill level in real time, feeding that data into a dispatch system that triggers a pickup when the bin actually needs one, not when the calendar says so, as the product specs lay out. RTS has scaled this network to more than 50,000 units, which created its own problem: how do you keep that many connected devices running in the field, through heat, through compactor vibration, through whatever a loading dock throws at a sensor over a wet winter, as RTS’s own account of managing the network admits? The answer was automated triage: algorithms that flag hardware issues before a technician ever has to drive out and look.
Contamination makes up the other half of the AI story, and arguably it costs more than the first half. A single bag of food waste in a recycling stream can get an entire load rejected at the material recovery facility, turning a diversion win into a landfill loss plus a penalty fee. Computer vision inside the bin, or mounted on the collection truck itself, can flag the wrong material before the truck even leaves the site, as industry reporting on AI in waste management has tracked. Waste Management and the Recycling Partnership tested truck-mounted optical systems for exactly this: catching contamination at collection rather than discovering it three days later at the sorting facility. The timing trade-offs stay real, though; catching it at the curb still means the damage happened upstream, as WasteDive’s coverage of the pilot noted.
Front-of-house technology closes the loop. Reverse vending machines (the kind that scan a bottle, verify material type, and sort it before a human ever touches a bin) are showing up at high-traffic retail and event locations, driven partly by Coca-Cola’s push for high-purity recycling streams, as WasteDive reported. Citi Field deployed Pello’s sensor network across its concourses for exactly this reason: a stadium generates enormous waste volume in short, unpredictable bursts, and static schedules simply can’t keep up, as RTS documented in its own launch announcement.
Innovative programming and technology finally gave leaders full visibility into the waste program.
The Part Nobody Puts in the Deck
None of this happens without friction. I’d lie if I said otherwise.
California’s SB 1383 mandates organics diversion and sets procurement targets for recycled content, with compliance burdens that fall differently on every municipality inside the state, as WasteDive’s reporting on the rollout has shown. New York City runs an entirely different model: Commercial Waste Zones that restrict which haulers operate in a given district at all, which means a national contract that works everywhere else turns legally unusable in Manhattan, as the city’s own zone map lays out. And state legislatures keep adding to the pile — extended producer responsibility laws, battery disposal rules, new organics mandates arriving on a rolling basis through 2026, as WasteDive’s roundup of incoming state laws tracks. I’ve watched a single national policy document fail to survive contact with fifty different regulators. It has to flex at the local level while still rolling up to one reporting standard, which is a harder engineering problem than it sounds.
Then there’s geography of a different kind: infrastructure disparity. A location in a major metro might have five composting facilities within a 30-mile radius. A location in a secondary or tertiary market might have zero. Packaging designed for recyclability in California turns functionally un-recyclable in a town where the material recovery facility lacks the sorting equipment to process it, as panelists at a recent sustainability packaging conference described. A company can standardize a program on paper. A company has no way to standardize the downstream capacity that program depends on; pretending otherwise leaves companies with beautiful color-coded bins and nowhere for the organics to actually go.
Data fragmentation presents the quieter challenge, an operational one: independent regional haulers run on each hauler’s own software, each hauler’s own invoicing cycle, each hauler’s own definition of what counts as “diverted.” Centralizing that data into one layer amounts to unglamorous, unsexy integration work; that work also determines whether the ESG report holds up as defensible or reads as decorative.
Hardware comes with a built-in tax. A sensor built for a clean indoor environment doesn’t necessarily survive a compactor yard in August in Phoenix, or a loading dock that floods every spring. Durability questions sound trivial until a sensor actually fails: I watched a year of fill-level data vanish from one program, silently, because a handful of units died in the field and nobody noticed until the annual report was already overdue.
Working Smarter, Not Just Bigger
Businesses must have a willingness to go upstream, before any waste reaches the bin entirely.
Source reduction belongs at the top of that list. Every pound of packaging that never arrives on site is a pound a facility never has to sort, haul, or report on. That means pushing the conversation back to suppliers: can a vendor ship in bulk instead of individually wrapped units? Does a distributor have room to swap a non-recyclable film for something the local MRF actually processes? Restaurants did versions of this for years out of necessity, running waste audits specifically to identify where packaging decisions upstream created downstream cost, as RTS’s own restaurant guidance describes.
Back-of-house standardization makes up the unglamorous second half: identical color-coded streams at every location, staff training that holds steady across markets, and, these days, image-based feedback that shows employees exactly what went wrong when a bin ends up contaminated. 1 Hotels built its zero-waste program around precisely this combination of certification, staff training, and continuous data tracking across properties, as RTS’s profile of the brand lays out, and the hospitality sector generally has had to get disciplined about it given how many hands touch a waste stream in a single hotel shift.
Surplus recovery closes the loop on the most expensive category of waste a company can generate: edible food. The Bill Emerson Good Samaritan Food Donation Act protects businesses that donate food in good faith, which removes the liability excuse that’s kept a lot of edible surplus heading straight to a dumpster. Restaurants alone lose an estimated $25 billion a year to food waste that lands in a landfill instead of a donation bin or a compost pile, generating methane instead of meals, as RTS’s breakdown of restaurant food waste details. Donating it isn’t charity accounting. It’s avoided tipping fees, full stop.
The Roadmap, Stripped Down
Here’s what the actual sequence looks like, once you strip away the vendor pitch decks: conduct a baseline waste audit at a representative sample of locations before buying anything. Install bin-level telemetry so the data exists before you try to optimize against it. Centralize hauler contracts under one procurement umbrella so pricing and reporting live in one place. Standardize upstream procurement with suppliers so less waste shows up in the first place.
I’ve watched that order matter every time. The companies closing in on 90% diversion didn’t start with AI-enabled sensors. They started with structure, added the sensors once that structure could use the data, and kept pushing the hardest work (procurement, supplier relationships, staff behavior) further upstream every year.
The infrastructure gets you organized. The intelligence layer gets you precise.
