You finance a specific VIN.
You get your data in aggregate.
You carry the maintenance and residual value risk on a single vehicle. You authorise repairs with no reference point, the invoice arrives as one document for the entire service network, and you learn the truth about the vehicle at return — when nothing can be changed any more.
FleetFlow works in the only moment when the money has not left yet: the cost authorisation.
A pilot on a carved-out part of your fleet. Alongside your process, without replacing the system you run.
vehicle financing in Poland in 2025 — 73.8 % of the entire leasing market
vehicles in the fleets of Polish fleet management companies (30 June 2026)
new fleet management registrations in H1 2026 — against 12.4 % fleet growth
year-on-year increase in repair costs in 2024, against 4.7 % inflation
Managed fleets grow by 12.4 %, new registrations by 2 %, and the share of fleet management in corporate purchases is falling. The portfolio tail is growing, not sales — and that is exactly the moment when margin is found in the cost of service, not in growth.
Where the data breaks
Accounting produces a supplier and an amount. Controlling needs something entirely different — and nobody assembles that chain.
- 1event: repair, service, damage
- 2the garage issues an estimate
- 3authorisation — usually without a benchmarkthis is where FleetFlow works
- 4work is carried out
- 5one invoice for the whole network and period
- 6cost lands on the contract 30–60 days later
The chain accounting does not produce
VIN → event → part or labour → mileage → authorisation → contract → customer → cost
Authorisation without a reference point
The approver sees an estimate and a monetary threshold. They do not see what this line cost at this garage a quarter ago, what it costs elsewhere, whether this vehicle had the same repair six months ago, or whether the part is still under warranty.
The spread between authorised dealers and independent garages: 40–50 % for the same job.
A single invoice with no per-VIN breakdown
One document, several hundred lines, the entire service network. Splitting it is manual work — the industry itself talks about many extra hours a month.
Part of the cost never reaches the right vehicle, only the „fleet costs” bucket.
A 30–60 day delay
By the time the cost is assigned to the contract, the vehicle has done another five thousand kilometres. You can describe the trend. You cannot react to it.
Reconciliation against the contract budget: quarterly, or at the end of the term.
Pricing feedback loop severed
The service budget in the contract was calculated three years ago, at different part and labour prices. The deviation surfaces in the final settlement.
Which means: in a dispute with the customer, at renewal.
The question no system in a Polish leasing company answers today
„Show me every VIN where expected margin has dropped below the threshold, explain why, and propose an action.”
Answering it means joining layers that sit in separate silos: the contract core, ERP, service, invoices, damages, tyres, telematics and remarketing.
Five mechanisms, not a list of modules
Each one addresses a different leak. The second is the most important — and every deployment starts there.
Cost reaches the vehicle by itself, on the day the invoice arrives
A network invoice arrives as one document with several hundred lines. Someone sits down and splits it manually, part of the cost lands on a pooled account, and controlling sees the full picture a quarter later.
The document is split into lines, each tied to a vehicle by a cascade: supplier cost account → registration number → fuel card or customer account. Classified (part, labour, material, tyres, damage, assistance) and posted to the cost allocation valid on the day of the event. Uncertain lines go to an exception queue with a suggested match.
- vehicle cost is complete, not „the part we managed to assign”
- controlling has the data on the day the document arrives, not after a quarter
- a human resolves exceptions, not every line
The decision before the cost is approved
The core of the offerRepair authorisation is the only moment in the chain when you can still not spend the money. Everything after it is bookkeeping. And the approver sees an estimate and a threshold — not the vehicle history, not the garage's pricing history, not the warranty.
For every estimate the agent returns a recommendation with reasoning: is the repair plausible for this mileage and age, is it a repeat of one six months ago, should it go through warranty, does the price match this garage's and the network's history, does the scope fit the service contract. Below threshold and without deviations — approved automatically.
- routine passes on its own; human attention goes to the cases that deserve it
- every recommendation carries reasoning, a confidence level and an audit trail
- a shorter decision means shorter downtime — on a truck, downtime costs more than the repair
Approved ≠ carried out ≠ invoiced
Three documents describe the same event: the authorisation, the work order and the invoice. In most organisations nobody reconciles them systematically.
FleetFlow reconciles them automatically and shows the differences: lines beyond the approved scope, price drift against the authorisation, duplicates, work invoiced without any consent at all.
- the fastest and most auditable proof of value — either there is a difference or there isn't
- a garage that knows every line is reconciled against the authorisation issues different invoices
- procurement gets hard material for renegotiating rates
Vehicle memory and garage memory as an asset
Vehicle history is scattered across the system, e-mail and people's memory. At return the discussion becomes about who remembers what — and return is, as the industry itself says, the most common source of end-of-contract disputes.
Every event stays with the vehicle: cost, scope, garage, mileage, decision and its reasoning. After a year you have a full vehicle history and a cost profile for every garage in the network.
- end-of-contract charges defended with a documented history, not an argument
- input for residual value pricing and a deviation signal before the vehicle comes back
- a feedback loop into pricing new contracts: the real cost curve instead of four-year-old assumptions
- an early risk signal: a vehicle that suddenly stops being serviced on schedule
An agent that works — instead of a screen you have to open
Fleet systems and BI give you a screen. Someone has to open it, know what to look for and have the time. In practice they open it once there is already a problem.
The agent reviews the data itself and escalates only what needs a human. You ask in plain language — „show me vehicles where service cost exceeded the assumption by more than 30 %” — and get an answer, not instructions on how to build a report.
- the value reaches someone who has no time to open the system
- operations on fleet data behind a per-company switch, role control and a full audit log
- the agent stops and asks a human when the decision deserves it
Run the numbers on your own fleet
There are no Polish benchmarks for cost leakage — nobody publishes them. What exists is arithmetic you can fill with your own numbers in a minute.
This is arithmetic, not a promise. The defaults are illustrative. On the first call we replace them with your numbers — and if the result does not justify a deployment, we say so instead of selling you a pilot.
Evidence that this money can be recovered
Five deployments from markets where this layer has been running for years. Measured and published.
These are not our results and we do not present them as a forecast for your portfolio. We present them as evidence that the problem has scale and can be addressed.
Customers rejected USD 41.6 m of unnecessary repair lines, +112 % year on year. 2.5 h saved per repair, 3 days faster closing.
AI scores an estimate against policies, limits, rate cards, history and warranty: approve / negotiate / reject / escalate. 1.5 m vehicles, USD 1.6 bn of identified savings in a year.
Scores every request 0–100 on 40 million data points from 5 years: propensity to authorise, historical price, frequency of the line at that garage.
An audit layer on invoices: −60 % processing time, −15–20 % maintenance cost in the tested areas.
70 % invoice automation, processing from 9 days to 4, on-time payments from 60 % to 90 %, 60 group entities.
e-invoicing: a window, not an advantage
From 2026 every garage, dealer and fuel station in Poland issues a structured document with line items, available through a single state API. The „build OCR and two hundred integrations” barrier is gone.
Invoice import alone is now standard with every vendor. The advantage begins only past it.
- matching lines to a VIN and a work order, with a confidence score
- reconciling the authorisation against the work order and the invoice
- catching duplicates and lines beyond the approved scope
- a decision taken before approval, not a report after the fact
Who this is for
We sell where you carry the maintenance cost risk yourself — because only there does every badly approved repair eat directly into your result.
Very high service costs, expensive downtime, high mileage. Every badly approved repair shows up in the result immediately.
A fixed instalment, with the maintenance and residual value risk on your side. Contract margin depends on service decisions.
Cost per VIN, authorisations, invoices, warranties and mileage — the same problem across a narrower service scope.
High rotation, damages, cost per day and resale. Vehicle history decides the result on the way out.
Pure financial leasing with no exposure to maintenance costs is not our customer today. If you do not carry the service risk, the recovered-margin argument does not apply to you — and we would rather say that on the first call than on the third.
Three different conversations
There is no single offer for „a leasing company”. It helps to know which conversation we are in.
“Stop overpaying for repairs and get your team's time back.”
Metrics: value of detected deviations · share of automatic decisions · authorisation time
“Give your customer what the competition does not offer — and see their fleet.”
Metrics: retention at renewal · service upsell · vehicle usage data
“Know earlier what is coming back and what it will be worth.”
Metrics: success rate in defending charges · number of disputes · residual value accuracy
You can resell the same tool to your own customer
FleetFlow runs under your brand as part of the contract. The fleet customer gets cost allocations and company structure, a controlling report and an accounting posting file, fuel budgets, recurring tasks, driver campaigns, an EV audit and a carbon footprint — plus an assistant they can simply talk to.
an argument the competition does not have as standard
a harder argument at contract renewal
telematics sits with the customer today, while the residual risk sits with you
What we do not promise
A shorter list than the feature one, and a more important one. A financial institution buys from someone who knows the limits of their own product.
We do not price residual value
We supply the structured data the pricing stands on, and an earlier deviation signal. The pricing is yours.
We do not replace the core system
And we do not want to replace it as a first step. We come in through the API as a decision layer and hand enriched data back.
We do not start with a market benchmark of repair prices
The first version compares a vehicle against its own history and a garage against its own — plus contract rules, duplicates and repeats. The cross-customer benchmark grows with every new client.
We do not present numbers from foreign deployments as our result
We quote them as a reference for the scale of the problem. Your result gets calculated on your data, in the pilot.
How we start
No system replacement, no process change and no quarter-long IT project. The first result is a number, not a presentation.
Diagnostic call
45 minutes. Ten questions about your authorisation and invoicing process. You leave with an order of magnitude calculated.
Three months of your data
Service invoices, work orders and authorisations from a carved-out part of the fleet. No integration, no access to production systems.
Two to three weeks in parallel
We work alongside your process. We compare our recommendations against the decisions that were actually made.
A report with one metric
The value of detected deviations. Plus the share of automatically matched lines and authorisation time.
Metrics are agreed before the start, not after. Four to six: unassigned cost, share of automatically matched invoices, number and value of exceptions, average authorisation time, share of automatic decisions, repeats and warranty cases caught.
The questions that always come up
We will show, on your data, how much service cost is not controlled today
At the level of the vehicle, the authorisation and the invoice. Without replacing the system you run. After a few weeks you have a measurable report — and you decide whether there is anything worth discussing further.