Every toll transaction captures a time, a location, a lane, a vehicle, and a plate. Yet at most fleets it arrives as a single line on a monthly invoice, and everything except the dollar figure gets set aside. The amount, it turns out, is the least useful part of the record.

Consider two fleets running similar lanes in the same region under identical published rates. One posts a predictable toll bill every month. The other watches toll spend climb and cannot say why. Because both pay the same rates, the gap comes down to how each fleet operates, and the evidence sits in toll data that neither of them reads.

Rates attract attention because they are easy to see. They are also the part nobody controls.

What fleet toll management can control

Toll rates went up again in 2026. The Pennsylvania Turnpike raised rates 4% in January. The MTA went up 7.5% across all of its facilities. The network keeps growing as well, with new capacity opening this year in Kansas, Washington, Virginia, and Texas, adding more places where routes and billing can change even when a fleet’s own plan does not.

All of it is announced in advance. Fleets can budget for it and move on. That is the part of fleet toll management a fleet controls.

What cannot be budgeted for is narrower and more specific: the one charge that looks wrong and takes half a day to run down, multiplied across a month of charges. On the invoice, it appears as nothing but a number.

What toll data records that the invoice leaves out

Charges like those are operational facts. They reach the fleet as amounts, and when toll spend climbs faster than rates do, the cause is usually in that detail. Across the toll activity PrePass® processes for fleets, a few patterns come up again and again.

• Express lane usage. Drivers taking premium lanes when the schedule did not require it. One charge is easy to justify. The same truck doing it every week is a policy question.
• Out-of-route activity. Detours, late route changes, and driver preference putting trucks on toll roads they did not need.
• Transponder mismatch and plate mismatch. A device moved between trucks, or a plate assigned to the wrong vehicle, so charges land on the wrong asset and cost allocation quietly stops being accurate.
• Max toll charges. An agency applying its highest rate when entry or exit information is missing.
• Unmatched location. A charge recorded at a plaza where vehicle GPS does not show the truck.
• Duplicate charges and classification errors.

Read as accounting, each charge is a small variance not worth chasing. Read as operations, together they describe how the fleet actually ran last month: which lanes drivers preferred, which trucks left the route, where equipment records drifted out of date, and which agencies billed for something nobody can verify.

That information already exists. A fleet pays to generate it every time a truck crosses a plaza. Almost none of it gets used.

Why toll analytics is harder than it sounds

The volume is real. A large operation can generate tens of thousands of toll transactions a month across a dozen or more agencies, each with its own file format and billing cycle. Someone reviews the biggest charges, the violations, and whatever a customer calls about. That process is good at finding isolated problems and poor at finding patterns. The larger costs are in the patterns.

Volume is only part of it. The bigger problem is that no single system holds the answer. Toll records sit with the agencies. GPS data sits with the telematics provider. Vehicle and plate assignments sit in the fleet records. Rental and lease information sits somewhere else again, often with a partner. Asking whether one specific charge was correct means pulling from at least three of those systems. With a thousand charges, nobody asks at all.

That is why the useful information only appears when the sources are read together. Take a charge recorded at a plaza a truck never passed through. In the toll record it looks ordinary. In the GPS record there is nothing, because nothing happened. The problem shows up only when the two are placed side by side.

That is the bar for toll analytics: checking charges against what the operation actually did. The comparison is easy to describe and tedious to run. Checking one charge against GPS takes a few minutes. Checking tens of thousands is arithmetic at a scale no one has time for.

This is what AI is good at. It reads every transaction, compares it against vehicle, plate, and location records, and flags the few that do not reconcile. The software decides nothing. It narrows a month of charges down to the ones worth a person's attention.

What toll management software changes

When those records are read together continuously, the question changes. Instead of asking what the fleet spent last month, an operations leader can ask why the same issue keeps appearing on the same lane, the same truck, or the same driver.

Charges get confirmed before payment rather than investigated after it. Costs land on the right vehicle, renter, or customer. Recurring issues appear as trends instead of one-offs.

Fleet operations has made this shift before. Fuel, maintenance, and routing all moved from monthly reporting to daily operating data. Toll is the last major area of fleet cost still handled mostly as an accounting task, even though the underlying records already support a more operational view.

Where to start

Fleets working on this have a few options. PrePass Toll Management approaches it by bringing toll activity from multiple agencies into one place through INFORM TM Tolling, so separate agency bills arrive as a single monthly invoice.

From there, GPS Toll Verification checks when and where charges occurred against route and GPS data, and AI Toll Insights connects toll, GPS, and vehicle data to surface transponder mismatch, plate mismatch, express lane usage, max toll charges, unmatched location events, and out-of-route activity. The intent is to point a team at what needs review rather than ask them to read every transaction. The Violation Prevention Service helps maintain continuous plate coverage, and the Violations Settlement Service works with participating agencies on outstanding violations.

Toll rates will keep rising, and that part is out of a fleet's hands. What toll data says about an operation is not.

PrePass Toll Management, including GPS Toll Verification and AI Toll Insights, is available at prepass.com/tollmanagement.

Sources

1. Pennsylvania Turnpike Commission, 2026 toll rates:
2. Metropolitan Transportation Authority, 2025 fare and toll changes:

New network capacity referenced generally by state, supported in these links:

1. Kansas Turnpike Authority, 69 Express:
2. Washington State DOT, Puget Sound Gateway Program (SR 167):
3. Virginia DOT, 495 Express Lanes Northern Extension:
4. TEXpress Lanes, I-635 East and I-30 projects:

Source: FreightWaves, https://www.freightwaves.com/news/the-most-useful-part-of-a-toll-charge-is-not-the-amount