Evaluate Voyager platform U.S. Bank and Verra Mobility for 2026 toll management in three steps: map your workflow, test a 30-day toll sample, and select the vendor that fully reconciles it.
The defensible choice changes if a vendor’s contract and 30-day sample reconciliation do not match your transaction count, exception ownership, and annual fee basis.

Map the toll workflow before any demo
The pre-demo workflow map is the control artifact for 2026 toll management: it fixes the current process before either Voyager platform U.S. Bank or Verra Mobility can frame it through a feature demo. Draw one swimlane for every toll lane, transponder or plate account, rental or lease exception, and dispute path, and mark the owner of each handoff for both vendors. The map should show who triggers the event, what data moves, where it lands, and who resolves the exception, so operational ownership is visible before pricing or features are discussed.
Name the systems of record that the pilot will test: fuel card, TMS, asset title, driver reimbursement, and the issuing toll authority portal. For each system, require the vendor to show exactly where its file lands and who edits exceptions. If a vendor cannot identify the landing point and the exception editor, that gap belongs in the comparison as an undefined ownership condition rather than as a future implementation detail.
| System of record | Required vendor proof |
|---|---|
| Fuel card | File landing point and exception editor |
| TMS | File landing point and exception editor |
| Asset title | File landing point and exception editor |
| Driver reimbursement | File landing point and exception editor |
| Issuing toll authority portal | File landing point and exception editor |
For each lane, account, exception, and dispute path, record the trigger, source record, handoff owner, exception type, and resolution evidence. Mark whether the owner is internal staff, Voyager platform U.S. Bank, Verra Mobility, or the toll authority. This prevents a vendor from absorbing a step in a demo while leaving the underlying exception with the customer.
Lock the minimum pilot data fields before any vendor receives data: vehicle ID, plate, transponder ID, toll authority, and transaction ID. Both vendors must accept the same field set and return the same identifiers, because the later reconciliation step depends on matching records without translation or manual repair. Treat any request to add, rename, or omit a field as a scope change that must be documented in the map.
Use the approved map as a gate: no demo, proposal, or fee discussion should proceed until the swimlanes, systems of record, owners, and minimum fields are documented and accepted by the internal toll, fleet, finance, and driver-reimbursement stakeholders. The map does not select a vendor; it defines the conditions under which the later evidence, sample, and cost steps can be compared fairly.

Demand three evidence sources, not slideware
A defensible 2026 toll-management decision rests on a three-source evidence file and a convergence check, not on interface demonstrations. This section specifies the three-source evidence file and the convergence check as the only basis for accepting vendor evidence. Each source answers a different question—what the system can record, what actually occurred, and what the toll network permits—and a claim is accepted only when the same transaction can be traced consistently across all three.
Source 1: vendor interface documentation. Request field-level examples from U.S. Bank Voyager and Verra Mobility for transaction ingest, exception codes, and dispute status. The examples must show how a posted toll is identified, how an exception is categorized, and how dispute progress is represented from submission through resolution. Treat documentation as a test artifact: if a required field is absent, ambiguously defined, or available only through manual export, record it as a fail rather than a follow-up item. A polished screen does not substitute for a documented data element.
Source 2: your own 30-day toll invoice sample. Reconcile every posted toll in the sample to a vehicle or plate and count unmatched items by reason, such as missing identifier, duplicate posting, unreadable plate, or unassigned account. This unmatched count becomes the denominator applied to both vendors, ensuring each is measured against the same operational reality. Require each vendor to show how its records would classify and resolve every unmatched item using the fields documented in Source 1.
Source 3: issuing toll authority interoperability guidance. Verify the supported plate or transponder types, account rules, and interoperability conditions published by the relevant toll authorities for the roads in your sample. This source confirms whether a transaction failure originates in vendor handling or in an authority-level limitation, preventing misattribution during review.
| Evidence source | Required artifact | Convergence check |
|---|---|---|
| Vendor documentation | Field-level examples for ingest, exception codes, and dispute status | Every sampled transaction and exception can be represented without a missing field |
| 30-day toll invoice sample | Posted tolls reconciled to vehicle or plate with unmatched reasons counted | Vendor records explain each matched and unmatched item against the same denominator |
| Toll authority guidance | Published interoperability and supported plate or transponder rules | Observed transaction behavior is consistent with authority-supported types and conditions |
Apply the convergence check by tracing sampled transactions across the file: a toll posted by the authority should appear in the invoice sample, map to documented vendor fields, and conform to stated interoperability rules. Any break in that chain is a documented evidence gap that must be resolved with source material before vendor claims are accepted.

Pick the reconciliation pilot over the demo
The 30-day reconciliation pilot is the comparison winner because it tests Voyager platform U.S. Bank and Verra Mobility against your own toll sample instead of a prepared demonstration. A demo can show screens; a pilot shows whether the vendor can match, explain, and price your actual transactions.
| Option | What it proves | Why it is not enough |
|---|---|---|
| A: vendor demo | User interface and feature walkthrough | Proves UI only; it does not prove transaction match, exception ownership, or fee basis on your sample |
| B: contract redline | Terms, service levels, and commercial language | Verifies terms but not data; it cannot show how the vendor handles your exceptions |
| C: 30-day toll-sample pilot | Transaction match, exception ownership, and fee basis on your own sample | Winner: it tests the actual workflow with your transactions and your exception rules |
Run both vendors on the identical sample and score four things: matched transactions, unresolved exceptions, manual minutes, and fee math. Keep the sample, input files, and exception rules unchanged so the score reflects vendor performance rather than preparation differences. Require each vendor to return the same reconciliation fields for every transaction, including transaction identifier, match status, exception reason, owner, and calculated charge.
Apply the elimination rule before comparing scores: any vendor that cannot name the exception owner for an unresolved item is eliminated. A pilot that leaves ownership unnamed has not reconciled the sample; it has only moved the work to your team. Record the owner as a named role or named person, not as a queue, department, or shared responsibility.
Do not compare a monthly platform fee to a per-transaction fee directly. Score fee math only after both vendors' charges are expressed through the same sample-based calculation, so the comparison reflects cost per reconciled transaction rather than billing cadence. If one vendor cannot show the fee basis behind each line, treat that fee as unverified.
Choose the vendor that reconciles the complete sample with the fewest unresolved exceptions and manual minutes, after fee math is normalized to the same basis. If two vendors tie on match rate, break the tie by exception ownership clarity and by the lower manual minutes, not by interface preference.

Normalize every fee to annual cost
To compare Voyager platform U.S. Bank and Verra Mobility on equal terms in 2026, normalize every quoted charge to a single annual-cost figure before any selection decision. Use this formula: annual cost = platform or account fees + (transaction count × per-transaction fee) + (exception count × manual-review minutes × loaded labor rate) + missed-toll penalties + implementation and integration cost. Pull each input from a signed fee schedule; if a vendor will not commit an input in writing, mark it “quote required” and treat the annual total as incomplete rather than estimated.
Each term captures a different cost driver. Platform or account fees and per-transaction fees are the visible charges, but exception handling converts operational friction into labor spend, and missed-toll penalties transfer risk back to your budget. Implementation and integration belong in the same year-one view because, as OneIO’s 2026 integration total-cost-of-ownership comparison notes, the visible build or deployment cost is only part of the larger total cost of an integration. Keeping these components in one annual model prevents a low line-item price from masking downstream expense.
The lowest per-transaction fee is not the lowest annual cost. Run this check on your own sample: multiply the per-transaction fee gap between the two vendors by your monthly transaction count, then by 12, to get the annual gap. That difference can be erased or reversed by higher platform fees, a larger exception burden, greater manual-review time, missed-toll penalties, or one-time implementation and integration charges. A vendor that wins on unit price can lose on annual cost once every committed and quote-required input is normalized.
Apply the formula consistently across both vendors using the same transaction count, exception count, review-time assumption, and loaded labor rate so the comparison reflects contract terms rather than presentation choices. When reconciliation requirements are satisfied, select only the vendor whose signed schedules and documented quote-required items produce the lower defensible annual cost for your operation.

What a 30-day pilot cannot prove
This section alone lists the edge cases where pilot evidence breaks. A 30-day pilot can prove only what happened inside its own 30-day toll sample; it cannot prove that the same result will hold when the toll population changes.
| Edge case | Why pilot evidence breaks | Required proof before signing |
|---|---|---|
| Seasonal toll spikes | The pilot can miss your peak month, so fee math at peak-month transaction count may differ. | Rerun the fee math at peak-month transaction count; the vendor still wins only if it holds the same fee basis. |
| New toll authority, plate, or transponder type | Support for that lane may be unproven in the sample. | Require a dated support document; if it is missing, require a written exception and re-score that lane. |
| Fleet mix or volume change | The transaction profile no longer matches the sample. | Re-run the reconciliation test at the changed mix or volume and confirm exception ownership stays assigned. |
For seasonal spikes, apply the proof requirement in the edge-case table above: rerun the fee math at peak-month transaction count before signing, and keep the vendor only if it holds the same fee basis.
For new toll authorities, plates, or transponder types, the pilot breaks unless the vendor supplies a dated support document. If that document is missing, require a written exception and re-score that lane before the contract is signed.
For fleet mix or volume change, the pilot breaks when the transaction profile no longer matches the sample. Re-run the same reconciliation test at the changed mix or volume and confirm that exception ownership remains assigned to the same party.
If an edge case cannot be tested before signing, convert it into a contract condition with a named owner and a dated review. Do not let a clean pilot override a missing document or an untested peak month.
Keep these edge cases in the contract file as conditions, not as assumptions; a pilot that reconciles 100% of one 30-day sample does not by itself reconcile a changed toll population.

Work a 30-day reconciliation worksheet
This section provides the worked worksheet artifact and checkpoint arithmetic for your 30-day toll sample. On day 0, lock every input in one file and label each value as Observed (pulled from your records) or Quoted (provided by Voyager platform U.S. Bank or Verra Mobility). Do not substitute demo summaries for these fields; the worksheet is only valid when every cell traces to your sample or a written vendor quote.
Run the checkpoint arithmetic on the same sample before comparing vendors. First, confirm transaction closure: matched count plus exception count must equal invoice transaction count. If the sum does not equal the invoice count, stop and resolve the gap in the source records before proceeding.
Next, calculate manual exception effort for the 30-day period. Multiply exceptions by minutes per exception, then convert to hours: 540 × 6 = 3,240 minutes; 3,240 ÷ 60 = 54 hours. Apply the loaded labor rate to quantify the sample-period labor burden: 54 × $45 = $2,430. Record this result alongside each vendor’s quoted annual fee in the worksheet so the comparison uses the same transaction, exception, and labor inputs.
Replace fleet count, transaction volume, exception mix, handling time, labor rate, and both annual fee quotes with your own observed and quoted values before deciding. A defensible 2026 selection depends on your worksheet reconciling 100% of sample transactions with documented exception reasons and vendor inputs you can verify in writing.
Apply five switch-or-stay decision rules
Apply these five rules only after the 30-day sample has been run through both vendors and every fee has been normalized to annual cost. Treat them as ordered gates rather than a menu: reconciliation and ownership checks come before cost, and a vendor eliminated by an earlier gate cannot be revived by a later fee advantage. This ordering prevents a lower quote from masking unmatched tolls.
| Rule | If... | Then... |
|---|---|---|
| 1 | Both vendors reconcile the same sample at 100% and annual fee bases are equal | Stay with the incumbent and negotiate an exception-review SLA |
| 2 | One vendor has unmatched transactions with no named owner | Eliminate that vendor regardless of a lower quoted fee |
| 3 | Verra Mobility's signed fee schedule wins on annual cost after manual review | Run a 90-day parallel switch using the same exception log |
| 4 | Voyager platform U.S. Bank wins on that same annual-cost math | Keep it and close the evaluation without a parallel switch |
| 5 | Neither vendor reconciles 100% of the sample | Reject both for this cycle and retain the current process |
Apply rules 1 and 2 from the table above in order: equal 100% reconciliation and equal annual fee bases means stay with the incumbent and negotiate an exception-review SLA, while unmatched transactions with no named owner eliminate a vendor regardless of a lower quoted fee, because an exception without an accountable owner has no correction path and no audit trail.
If Verra Mobility's signed fee schedule wins on annual cost after manual review, run a 90-day parallel switch using the same exception log and the same owner assignments, so the switch is measured against the same control baseline. If Voyager platform U.S. Bank wins on that same annual-cost math, keep it and close the evaluation without a parallel switch.
If neither vendor reconciles 100% of the sample, reject both for this cycle and retain the current process until transaction matching and exception ownership are corrected. A lower annual fee cannot offset unmatched tolls that no named owner can resolve, and it cannot create evidence that the pilot did not produce.
Record which rule triggered, which vendor was eliminated or retained, and the owner of every exception in the same log used for the pilot. The decision is defensible only when the rule, the sample result, the annual fee basis, and the owner assignment can be traced to one record without reopening the workflow map or fee formula.
Use the exception-review SLA to set response times and closure evidence for unmatched tolls, but do not treat SLA promises as proof of reconciliation. The pilot result remains the controlling evidence; the SLA only governs how exceptions are handled after the decision.
Frequently Asked Questions
What artifact should be completed before scheduling a vendor demo?
The pre-demo workflow map is the control artifact that fixes the current process before either vendor can frame it through a feature demo.
Under what condition does the defensible vendor choice change?
The defensible choice changes if a vendor’s contract and 30-day sample reconciliation do not match your transaction count, exception ownership, and annual fee basis.
Which process elements each get their own swimlane in the workflow map?
Draw one swimlane for every toll lane, transponder or plate account, rental or lease exception, and dispute path, and mark the owner of each handoff for both vendors.
What four ownership details must the workflow map make visible?
The map should show who triggers the event, what data moves, where it lands, and who resolves the exception.
Which five systems of record must the pilot test?
Name the systems of record that the pilot will test: fuel card, TMS, asset title, driver reimbursement, and the issuing toll authority portal.
How should a vendor's inability to identify a file landing point be recorded?
If a vendor cannot identify the landing point and the exception editor, that gap belongs in the comparison as an undefined ownership condition rather than as a future implementation detail.
Quick answers
| What are the three steps to evaluate Voyager platform U.S. Bank and Verra Mobility for toll management? | The three steps are to map your workflow, test a toll sample over the specified period, and select the vendor that fully reconciles it. |
| When might the defensible choice between the two vendors change? | The defensible choice changes if a vendor’s contract and sample reconciliation over the specified period do not match your transaction count, exception ownership, and annual fee basis. |
| What is the purpose of mapping the toll workflow before any demo? | The pre-demo workflow map is the control artifact for toll management in the specified year, fixing the current process before either vendor can frame it through a feature demo. |
| What details should the toll workflow map include? | The map should draw one swimlane for every toll lane, transponder or plate account, rental or lease exception, and dispute path, and mark the owner of each handoff for both vendors, showing who triggers the event, what data moves, where it lands, and who resolves the exception. |
| How should a gap in vendor information be handled during the comparison? | If a vendor cannot identify the landing point and the exception editor, that gap belongs in the comparison as an undefined ownership condition rather than as a future implementation detail. |
Also worth reading: Fleet toll reconciliation cost 2026: U.S. Bank Verra Mobility vs standalone accounts: Fleet toll reconciliation cost 2026: · Fleet Management Software Pricing: 1,800 vs. 2,160 Units—Vehicle or Seat?: Fleet Management Software Pricing: 1,800 · Fleet TCO 2026: Cloud vs On-Prem for a 50-Vehicle Fleet: Fleet TCO 2026: Cloud vs