BeyondTrucks makes AI and human dispatch picks pass a fleet's rules first

Does the driver have facial hair? Has the driver been trained to handle a bear encounter in Alaska? At some specialized trucking operations, questions like these can help decide who can haul a load.

Neither question is hypothetical. BeyondTrucks CEO Hans Galland told Truck News that a fleet's rules can be specific enough to keep a driver with facial hair from delivering to a food plant, or to require certified bear attack training for drivers in Alaska.

BeyondTrucks, a San Francisco-based maker of transportation management and AI dispatch planning software, unveiled its Compatibility Engine on Oct. 17 in Charlotte, North Carolina, at the American Trucking Associations' 2026 Management Conference & Exhibition, The Trucker reported. Galland described the tool as an automatic guardrail. Before a load goes out, every assignment has to clear the fleet's rules, whether a dispatcher made it or the company's AI optimization engine suggested it.

So a software vendor selling AI dispatch has built a gate that its own AI has to pass. For a safety or operations director at a chemical, food-grade or other specialized carrier, that changes what a dispatch decision leaves behind: a rule that was applied, or a logged reason it wasn't.

Four places to write a rule

According to Heavy Duty Trucking editor Deborah Lockridge, who covered the launch, fleets can define requirements for a given stop, commodity, customer or piece of equipment. The engine then compares those requirements with the qualifications and current status of each driver, tractor and trailer. Rules may address what a driver is qualified to do, which features equipment has, who is available, whether regulations are met and whether the hauled commodity is a compatible fit.

CCJ's Jeff Crissey described the same setup from the fleet manager's side: managers can build their own sets of rules for each driver, truck and trailer, based on the hauling applications of each customer. The Trucker noted the system is designed for fleets with complex or specialized operations, which is where customer-by-customer rules pile up fastest.

The real examples are ordinary, and that's the point. Has this driver completed the customer's refinery orientation? Does this trailer meet the customer's specifications? What did the trailer haul on its last load? Each one is a small fact. Galland said dispatchers work through hundreds of loads a day, often against the clock, and match drivers, tractors, trailers, routes and times to each one. Every one of those assignments needs the small facts to line up.

When the engine flags or blocks an assignment, the dispatcher sees which rule fired and why. Dispatchers can still override, but each override is logged with the name of the person who made it and their reason. Lockridge reported that the log creates a record for safety, compliance and insurance teams. The same rules also become constraints inside the optimizer, so the AI considers fewer driver-and-equipment combinations from the start rather than proposing matches that will be rejected later.

Integrations with platforms such as GWC and DQM Connect let driver and equipment compliance data flow straight into the checks. During the conference demo, the system took a load's pickup and delivery requirements and compared them with a driver's chemical-hauling training, active endorsements, site requirements and availability. It then ranked drivers and equipment by qualification, with the best fits on top. Coverage did not describe the exact wording of a flag or a guided fix workflow, so fleets will want to see both on their own data.

A fixed rulebook above the optimizer

BeyondTrucks describes the engine's logic as deterministic: the same inputs always produce the same result. That matters because an optimization engine's output is a recommendation. A deterministic layer gives the fleet a way to confirm that an AI-built plan meets its operating requirements before anyone acts on it.

Galland described the dispatcher's work as two separate tasks. Finding the best plan is only one of them, he said. Dispatchers also have to confirm that each assignment is permitted, because a mistake carries high stakes. Set next to the product, the division is clear: the optimizer handles the best plan, and the Compatibility Engine decides whether that plan is allowed.

According to BeyondTrucks, as The Trucker relayed it, the deterministic approach lets fleets apply consistent standards whether an assignment came from a dispatcher or from AI-powered planning. That is the substance behind the headline. A human pick and a machine pick face the same test, and a failure looks the same in both cases.

The limit is obvious. A deterministic engine only knows what someone has written down. If the rule set leaves out a customer's refinery orientation, the engine passes that assignment as cleanly as a dispatcher who forgot about it. The override also exists on purpose: final judgment stays with the dispatcher, and the system's job is to record that judgment.

The design stands out against another way vendors are packaging AI for fleets this month. Samsara launched Agent Studio on Oct. 13, letting fleets build AI agents that draw on fleet data and company-specific information to flag hours-of-service risks, answer driver policy questions and compile reports, Heavy Duty Trucking reported. One customer, S. Bertram Foods, uses an agent that flags potential HOS violations three hours in advance, according to Samsara. Both products put a fleet's own rules to work. BeyondTrucks' version places them as a fixed checkpoint between an AI recommendation and a dispatched truck, which suggests the two approaches answer different questions for a buyer.

Getting the rulebook out of dispatchers' heads

Galland told Heavy Duty Trucking that dispatch planning often depends on cheat sheets, Post-it notes and the memory of dispatchers who have spent decades learning a single operation. "In most fleets today, compatibility rules live in the heads of experienced dispatchers," he said. Those veterans know which trailer can't go to which customer, which driver has finished which site orientation, and what each trailer carried last.

Speaking with CCJ's Crissey, he called this tribal knowledge, used to decide whether a driver is qualified to haul a certain load. The engine, he said, lets the dispatch team build that knowledge into the dispatch system so it becomes a guardrail in decision-making, turning tribal knowledge into institutional knowledge.

The practical case is staff turnover. Once requirements are recorded in the system, fleets can apply them consistently when dispatchers retire, leave or train their replacements. When one segment can carry 400 qualification types, the rulebook in a veteran dispatcher's head is a compliance system with no backup copy.

There's a screen-design angle too. Matias Oberpaur, BeyondTrucks' director of business development, told Truck News the company wants to give fleets more real-time data than ever while keeping it easy to navigate and avoiding information overload. BeyondTrucks demonstrated the engine at the show on a six-screen setup. Lockridge noted that many dispatchers, by contrast, have worked from a single crowded screen that isn't the most user-friendly.

That gap is the design problem in miniature: more data, less room to show it. A rule engine that surfaces only the conflict, plus a ranked list of qualified options, moves part of the memorization off the dispatcher and into the system.

What the override log is worth

The Trucker and CCJ named the same three costs of a poor match: a rejected load, a compliance violation or a safety incident. The Trucker's framing was that any of them can quickly turn a routine dispatch decision into a costly mistake. For carriers whose customers set their own driver and equipment requirements, the first cost sits at the level of each customer relationship.

Those three costs line up neatly with the three teams Lockridge said get the override record: safety, compliance and insurance. Galland said that for most fleets these rules will set their service apart and help them deliver service "consistently and safely, every time."

Blocks prevent errors. The log could end up being the more valuable piece for a safety manager or an insurance renewal, because it captures the judgment calls that used to disappear: who decided the rule didn't apply to this load, and why. For fleets with few customer-specific requirements, a deterministic layer adds less. For a chemical hauler juggling hundreds of qualification types, it's the part a dispatcher can't do from memory at scale.

The first real work comes before go-live: getting the rules written down. Fleets whose driver qualification records already sit in GWC or DQM Connect can feed those records straight into the checks. Everyone else starts by asking the most experienced dispatcher on staff to dictate the rulebook.