By Luis Lopez, AI transportation consultant, CEO of Go Hub.io Holdings Corp and subsidiaries, and host of the Freight Guru Podcast
IFTA reporting is one of those back-office jobs that nobody notices until it goes wrong. Every quarter, a fleet that operates qualified motor vehicles across jurisdictions has to report miles driven and fuel purchased in each state or province, and the base jurisdiction redistributes the tax. The math is simple. The record keeping behind it is not. That is where AI IFTA reporting tools come in, and it is also where fleets need to be careful about what they hand off.
As an AI transportation consultant, the first question I ask about any fuel tax process is simple: if an auditor asked for the source records behind last quarter’s return, how long would it take to produce them? The answer usually tells you more about the process than the return itself.
What IFTA actually requires you to prove
An IFTA return is a summary. Behind it, a carrier is expected to keep detailed distance records and fuel records that support every number. In general terms that means trip-level records showing dates, origin and destination, routes, jurisdictions crossed, miles in each jurisdiction and the vehicle involved, plus fuel receipts showing date, seller, location, gallons, fuel type and the vehicle that was fueled. Retention periods are set by the agreement and your base jurisdiction, and they run for years, so check your base jurisdiction’s current requirements rather than relying on memory.
The point is that the return is only as good as the records. Automation that produces a clean return from messy records has not solved the problem. It has hidden it.
What fleets can reasonably automate
Jurisdiction miles from ELD and GPS data
Most fleets already collect location data through electronic logging devices and telematics. Software can take those GPS points, determine when a truck crossed a state or provincial line, and calculate miles in each jurisdiction for each trip. This is far more consistent than drivers filling out trip sheets by hand, and it produces the trip-level detail that supports the return. If you want to understand what your telematics data does and does not contain, see questions to ask about AI and ELD telematics data.
Fuel receipt capture
Fuel card transactions can usually be imported directly. Receipts paid another way, such as cash or a company card that is not tied to fuel, are the gap. Document extraction tools let drivers photograph a receipt and have the date, location, gallons and fuel type pulled out automatically and matched to the right truck. That turns a shoebox of paper into structured records.
Matching fuel to vehicles and trips
Once miles and fuel are in the same system, software can match each purchase to the truck and trip it belongs to, using time, location and vehicle identifiers. A purchase in one state while the truck’s GPS shows it hundreds of miles away is an obvious flag.
Quarterly totals and draft returns
With clean trip and fuel records, generating jurisdiction totals and a draft return is straightforward. This is the part most people think of as “IFTA software,” but it is the least risky step if the inputs are right.
Gap detection is the real value
The most useful thing AI can do for fuel tax reporting is not calculation. It is finding what is missing or does not make sense. Good tools should flag:
- GPS gaps. Periods where a truck moved but the device stopped reporting, which leave miles unassigned to any jurisdiction.
- Odometer and GPS disagreement. When calculated miles drift from odometer readings over a quarter, something needs explaining.
- Fuel without miles, or miles without fuel. A truck that bought fuel but shows no trips, or ran long trips with no recorded purchases.
- Unusual fuel economy. Miles per gallon well outside what that truck normally gets can point to missing receipts, duplicate receipts or a fuel card used for the wrong vehicle.
- Duplicate receipts. The same purchase entered from a card feed and a photographed receipt.
- Personal conveyance and off-highway use. Movements that may need to be classified differently depending on how your base jurisdiction treats them.
Each of these is a question a person must answer. The tool’s job is to put the question in front of someone before the return is filed, not after an audit letter arrives. Fuel data also feeds broader cost work, which I covered in AI fuel management for truck fleets.
Building an audit trail
An audit trail means anyone can trace a number on the return back to the trip records and receipts that produced it, and see every change made along the way. When evaluating a tool, look for:
- Retention of the raw GPS data, not just the calculated jurisdiction miles.
- Original receipt images stored alongside the extracted data.
- A log of every manual edit, showing who changed what and why.
- The ability to export trip-level and receipt-level detail in a format an auditor can read without your software.
- Clear documentation of how the tool assigns miles near borders and handles gaps.
If a tool fills GPS gaps automatically, you need to know how it does that and be able to show it. An estimated route between two points may be reasonable, but it should be labeled as estimated, not presented as recorded data.
What fleets must still check
Automation reduces the work. It does not transfer the responsibility. The licensee signs the return. Before filing each quarter, someone should:
- Review every flag the system raised and document how it was resolved.
- Confirm that every qualified vehicle in the fleet appears in the data, including trucks added or sold during the quarter.
- Compare total fuel purchases against fuel card statements and other payment records.
- Spot-check a sample of trips by comparing calculated jurisdiction miles to a map.
- Confirm tax rates and any rule changes for the quarter with the base jurisdiction rather than assuming the software was updated.
- Review fleet-wide and per-truck fuel economy for anything unusual.
If your team is using a general-purpose AI assistant to summarize data or answer questions about fuel tax rules, verify those answers against official guidance. Language models can state incorrect rules confidently. See how to catch AI hallucinations in freight for practical checks.
A sensible way to start
Begin with data you already have. Pull one quarter of ELD or GPS data and fuel card transactions into a single view and see how much reconciles on its own. The places it does not reconcile are where your process needs work, and they are also a good test of any tool you are evaluating. A vendor demo with clean sample data proves very little. Your own messy quarter proves a lot.
Done well, AI turns IFTA from a quarterly scramble into a steady process where problems are found while they are still easy to fix. Done carelessly, it produces confident numbers that nobody can support. The difference is human review and an audit trail.
For more conversations on freight and AI, subscribe to the Freight Guru Podcast.
About the author: Luis Lopez is a Miami-based AI transportation consultant and logistics entrepreneur, the CEO of Go Hub.io Holdings Corp and subsidiaries, and host of the Freight Guru Podcast.