From spreadsheets to automated routes: the power of an AI dispatch layer

Monday morning at 7:45. By now three new transport requests are in. One is in your inbox as an email. One is hidden inside a PDF from a customer's ERP system. And another sits buried five messages deep in a thread. Before planning can begin, someone has to open each request, pull out the key data, copy it into a sheet, work out what is missing, and decide which part can be assigned. This is what most forwarders live through every day.

The first hour of the day is not about finding better ways to optimize routes, improve utilization, or solve customer issues. It goes into gathering data that already exists but is scattered everywhere.

There is an alternative to traditional dispatch that does not ask you to give up your spreadsheet: an AI layer above your existing TMS that automatically processes new freight requests and drafts a dispatch plan. Your tool set stays the same. Only the transcribing of data and the hand matching of routes go away.

Why the spreadsheet persists in professional dispatch

Most software marketing files spreadsheets under professional failure. That view tells you more about the vendors than about the operators. At this level, Excel has survived in dispatching because it does well what it was designed to do. A dispatcher can add a column for a customer's quirk in seconds. The whole team can read the file, and it costs nothing beyond the office license. Nobody has to sit through an implementation. Alongside Excel, the company runs on its inbox, which is still the de facto EDI for mid-sized trucking firms. Customers keep sending transport requests by email, whatever a system vendor may suggest.

For a company with a stable customer base and a manageable fleet, this setup makes sense, because experienced dispatchers cover its gaps with judgment and memory. But the approach eventually hits a wall, and that wall is simply how many hours your dispatchers can physically work.

The costs that appear as volume grows

The first hidden cost is data entry. Every email or PDF has to be read and processed by hand before planning can start. So as your business grows, you are paying for more typing rather than better dispatching. Eventually the options narrow to hiring another dispatcher or letting customer response times slip. With Onyven we have cut manual data entry to a few minutes. Instead of acting as expensive typists, dispatchers now step in only to review and approve the occasional missing detail. Manual entry is also error prone. A typo in a weight or a reference number will not flash red in a spreadsheet. You find out about it when a truck is stuck at the loading dock.

The second cost is planning quality. A dispatcher has to match loads to vehicles from endless rows of data while fielding customer calls. In most cases they settle on a good enough plan rather than an optimal one. The difference shows up as empty kilometers: the missed backhaul from Hamburg, the two partial loads that could easily have shared a trailer, or the premium paid on the spot market when your own truck was actually free. In a business with razor thin margins, those missed connections are money left on the table.

The third cost is how late you catch mistakes. A spreadsheet does not verify your data. You realize too late that a truck is unfit for hazardous cargo, or that a driver is exceeding legal driving hours. Fixing it takes phone calls and detours, sometimes penalty fees. Had you caught the same conflict at 7:30, it would have been a quiet reassignment.

The final cost is the single point of failure. The most important dispatch logic sits in the head of your senior dispatcher. It is a huge risk if only one person on the team remembers that a specific customer needs a certain truck, or that a particular warehouse always turns away curtain-siders. That risk becomes obvious the moment the dispatcher goes on holiday, calls in sick, or hands in their resignation.

Why a full TMS migration stalls in the mid-market

If you go looking for an escape route, the standard answer is platform migration: replace your current system with a TMS and retrain your team. In the mid-market this underdelivers more often than the case studies suggest, for reasons that have little to do with the software itself. Integration and rollout take months. Processes get re-shaped to fit the system rather than the other way around. Operations depend on the judgment of experienced dispatchers, and those dispatchers are being asked to give up a tool they trust.

The industry even has a name for what happens next: the shadow spreadsheet. The company pays for the platform, but the actual planning still happens in Excel. The sheet bends to the workflow. The TMS does not.

A TMS does not solve the intake problem either. Even firms with a mature system of record still get requests by email, and the gap between inbox and system is still bridged by hand. (Read more about the software categories in our guide to dispatch software.)

A third route: an AI layer on top of the existing stack

The framing of spreadsheet versus TMS hides a third route: leave things as they are and add a smart layer on top to handle the coordination.

An AI dispatch layer plugs into your current stack. It reads the inbox and the PDFs, types the orders into your existing system automatically, and drafts the dispatch plan against the real constraints: vehicle capacity, equipment requirements, remaining driving time, traffic, customer specific rules. The best part is that there is no lengthy migration, and your dispatchers keep the system they already know.

What changes in your day to day

Data entry is off your plate. The three requests from earlier? The system extracts origin, destination, quantities, dates, and references from the email body or the attachment, double checks the data, and builds a dispatch plan.

The plan is ready for you. Validated loads are matched to the available fleet and drivers by checking constraints fast enough to surface combinations a person would need real time to find. It spots partial loads that can be bundled, lines up return trips with tomorrow's incoming freight, and uncovers hidden trucks that looked busy on the spreadsheet but are actually free. The output is a proposed plan built to raise utilization and cut empty kilometers. With Onyven you can reduce empty kilometers by 13% and shrink planning time from three or four hours to minutes.

Exceptions surface before departure. Before you approve the plan, the software highlights the potential risks. This is the core of the asymmetry. A constraint conflict caught during planning becomes a simple reassignment, while the same conflict caught on the A9 becomes an expensive afternoon. Moving detection to the start of the day is where a large share of the layer's economic value sits, alongside the found kilometers.

Dispatchers stay in control. The layer proposes, the dispatcher disposes. The person who carries responsibility for the plan approves it or changes it, and every automated step stays visible and reversible. For an operations leader this is a practical requirement, not a philosophical one. Override rights and a full audit trail are what make an automated plan defensible in front of customers and auditors.

Where a dispatch layer fits, and where it does not

A dispatch layer earns its keep where request volume has outgrown the morning routine: forwarders and carriers whose dispatch runs on Excel and email, with or without a TMS underneath, and whose realistic alternative is hiring another dispatcher. It is the wrong purchase for a company that has already decided on a full platform consolidation, because that is a TMS procurement, and a layer is a different decision with a different cost profile.

See it on one of your own workflows

The best way to understand an AI dispatch layer is to watch it run. Book a 30 minute demo and we will show you exactly what Onyven can do. We will walk you through a real workflow, taking a standard freight request and turning it into a clean, optimized dispatch plan right in front of you. More at onyven.com.