Haulage back-office admin: where the hours go, and how to get them back

Since April 2024, every UK haulier running loads to or from the EU has been filling in more paperwork than before. The Border Target Operating Model brought in documentary checks, risk-based physical inspections and sanitary and phytosanitary certification for a growing list of goods crossing the border - and none of that admin arrived with extra headcount to absorb it. It landed on the transport office that was already running the yard, the drivers and the compliance file.

That is the pattern behind most haulage admin. It rarely turns up as one big new job with its own line in the org chart - it arrives in fragments, bolted onto tasks that already existed, until the office is doing meaningfully more with the same number of people: BTOM paperwork stacked onto route planning, a new certification step stacked onto proof-of-delivery chasing, and a driver shortage that means fewer hands are around to help absorb any of it.

Almost none of it turns a wheel, but a load cannot be invoiced, a vehicle cannot be dispatched, and an inspector will not be satisfied the operating centre is compliant until it is done. If you run a haulage or logistics operation, you already know this much. What is harder to see from inside a working week is exactly how many hours it adds up to, and which parts are genuinely worth handing to a system first.

Where the transport office's hours actually go

Every operation is different depending on your mix of contract work, multi-drop and EU-facing loads, so treat the following as typical ranges rather than a fixed benchmark - the point is the shape of where the time goes, not a precise total.

Route and manifest planning: roughly 5-8 hours a week. Building the day's runs, sequencing multi-drop, and then rebuilding the plan every time an order lands late or a vehicle goes off the road. Most of this happens before a wheel turns and gets redone at least once a day when something changes.

Tachograph review and licence compliance: roughly 4-6 hours a week. Downloading and reviewing tachograph data, tracking driver hours against the Road Transport (Working Time) Regulations, and chasing Driver CPC, licence and medical expiries before they lapse. Miss one and it is a DVSA and O-licence problem, not just an admin one - the Traffic Commissioner can call an operator to a public inquiry over repeated tachograph infringements, so this file gets checked properly even when everything else is behind.

Chasing proof of delivery and matching it to invoices: roughly 6-10 hours a week. Collecting signed PODs off drivers and devices, matching each one to its job number, and hunting down the ones that went missing before an invoice can go out. Days of cash sit behind paperwork that has not caught up with the load.

BTOM and customs paperwork for EU-facing loads: roughly 3-5 hours a week for an operation with a meaningful share of EU traffic. Safety and security declarations, documentary checks and, for the relevant categories of goods, sanitary and phytosanitary certification - all sitting on top of a job that did not have a customs desk before 2021.

Add those together in a mixed operation and the back-office admin behind a working fleet can run to two full working days a week or more, spread across dispatch, the transport manager and whoever is nearest the phone when a customer calls asking where their load is.

What is safe to automate here, and what is not

Not all of that admin carries the same risk if it goes wrong, and that is the first thing worth being clear-eyed about before pointing anything at it.

Tachograph compliance and drivers' hours are a good example of where the line sits. Under GB drivers' hours rules, a driver's daily driving time is capped at 9 hours, extendable to 10 hours twice a week, with a weekly maximum of 56 hours and a fortnightly maximum of 90 - and the responsibility for staying inside those limits sits with the transport manager named on the operator's licence, not with a piece of software. AI can genuinely shrink the admin around this: flagging an approaching CPC or medical expiry, summarising an infringement report so it takes two minutes to read instead of twenty, drafting the reminder to a driver. It should not be making the call on whether an infringement is minor or serious, or deciding what happens next. That judgement, and the legal exposure that comes with it, stays exactly where it is now.

Route and manifest planning sits in similar territory. A system can build a first-pass plan and flag the conflicts. A dispatcher still needs to decide which customer gets bumped when a vehicle goes off the road on a Tuesday morning, because that decision carries a relationship, not just a sequence of postcodes.

Proof-of-delivery chasing and BTOM documentation are different. There is very little judgement in either. Matching a signed POD to a job number, sending a reminder when one is missing, checking a customs reference has been filed correctly against the load it belongs to - these are exactly the high-volume, low-variance, rule-based tasks that a well-set-up workflow handles reliably, freeing the office to spend its judgement where it is actually needed.

That split - what a system can safely take on versus what a person has to keep - is the first thing the AI workflow assessment for logistics and haulage works out for your specific operation, rather than assuming every haulier's mix looks the same.

One task, before and after: chasing proof of delivery

A useful example of what this looks like once it is fixed comes from a Birmingham-area haulage operator whose story has been reported publicly: an operations team member spotted that a high volume of daily pre-delivery calls followed the same script every time, automated that one specific task, and the business went on to cut missed-delivery rescheduling by 22% while growing towards £20m in revenue. We have written up what actually made that work, and what does not generalise from it, in more detail in our Birmingham haulage case study - the short version is that the result came from picking one measurable, repetitive task and fixing it properly, not from a platform rollout.

The same shape of fix applies to POD chasing. Before: a transport office manually working through a spreadsheet or a TMS export each afternoon, checking which jobs from the past few days still have no signed POD attached, phoning or messaging the driver, then manually matching the ones that come back to the right invoice line. In an operation running 40-60 jobs a week, that typically runs to several minutes per job once you count the checking, the chasing and the matching - which is where the 6-10 hours a week estimate above comes from.

After: the missing-POD check runs itself against the job list, a reminder goes to the driver automatically, and the match against the invoice happens as the POD comes back rather than in a separate batch job days later. The office's role shrinks to reviewing genuine exceptions - a POD that will not match, a driver who has not responded after two reminders - rather than doing the whole chase by hand every time. That is the difference between a job that eats a chunk of every afternoon and one that takes ten minutes of oversight.

Find your own number

The ranges above are typical, not yours. Your actual split between route planning, compliance, POD chasing and customs admin depends on your contract mix, how much of your traffic crosses into the EU, and how many people are already covering it.

If you want a rough first estimate before anything else, the admin cost calculator will give you a working figure for what your current admin load is costing in hours and pay, based on your own numbers rather than an industry average.

For a proper diagnostic of your operation specifically - which tasks are actually eating the week, which of them are safe to automate, and in what order - the AI workflow assessment for logistics and haulage is a 60-minute session that maps your transport office as it actually runs, not a generic template. Most operators we assess find 5 or more hours a week of recoverable office time, and if your report does not show at least that, the £799 fee comes straight back.

If you would rather start with something free, the 2-minute AI readiness quiz will give you a first read on where your own time leaks are, before you commit to anything.