Home · Solutions · Supply chain · Every tanker delivery reconciled to the litre at 15°C
Solution · Supply chainInvoice litres, delivery-note litres and tank litres stop being three different numbers
Every tanker delivery reconciled to the litre at 15°C
A robot joins the notification, the depot’s delivery note, gauge readings before and after the drop and the invoice, converts volumes to 15°C and compares every number for every delivery. An out-of-tolerance gap becomes a ready claim before the deadline passes.
Executive summary
Fuel is bought in litres converted to 15°C and received in actual litres. Comparing four numbers per delivery is human work, so it happens on a sample, and claim deadlines pass.
Every delivery reconciles itself: the depot note, tank measurements before and after, the invoice, all brought to the common 15°C base using density tables.
Excess transport losses stop vanishing. An out-of-tolerance gap is a complete document set and a calculation ready to claim in one click, within the contractual deadline.
depot delivery notes and wholesaler invoices; gauge readings before and after the drop; the ERP with deliveries; claim cases in Microsoft Teams
Business problem
Four numbers per delivery
Every fuel delivery produces at least four numbers that ought to meet: the litres on the depot’s delivery note (usually at actual temperature and converted to 15°C), the tank measurement before the drop, the measurement after, and the litres on the wholesaler’s invoice. Physics guarantees they will not be identical: fuel changes temperature in transit, and therefore volume, which is why the trade settles at the contractual 15°C base.
The conversion is not hard, only laborious: it needs density tables, the temperatures from the document and the gauge, and a quiet moment that receiving a delivery never offers. The station manager signs the document in the same minutes he serves customers, and head office first sees the delivery as a line to be paid.
So in practice every nth delivery gets compared, after the fact, once something has already raised suspicion. Meanwhile the right to claim losses runs on deadlines counted in days, and the evidence, the before-and-after measurement, the temperature, the seals, is cheapest on the day of the drop and priceless a week later, when it no longer exists.
The outcome is predictable: systematic small gaps against the buyer disappear into noise, claims are filed rarely and late, and the conversation with the supplier about losses rests on impressions rather than on a series of documented deliveries.
How it works today
Below is what the work looks like before anything is automated.
- PersonThe tanker arrives, the driver hands over the delivery note, the station manager signs between customers
- Risk of errorThe before-drop tank measurement gets skipped, because the drop is in a hurry and the gauge “will settle in a minute”
- PersonThe document travels to head office or waits in a binder until the end of the week
- WaitingThe wholesaler’s invoice arrives separately and is booked against the document, not the measurement
- Risk of errorConversion to 15°C happens on a sample, so systematic small gaps stay invisible
- Risk of errorWhen someone finally computes a loss, the claim deadline has passed and nobody recorded the before-and-after measurements
Why the current process costs more than it appears
The bill that never shows up in a budget.
- A loss small enough for the eye but not for arithmetic: half a per-mille systematically against you is real yearly money at volume, invisible in any single delivery.
- Claims are lost not because they are groundless but because on the day they could be proven nobody had time, and on the day there was time the evidence was gone.
- Accounting pays invoices against documents, not against the fuel that physically arrived. The difference has no owner, so it does not exist.
- Every supplier dispute starts from zero because there is no series: ten documented deliveries in a row say more than one argument a quarter.
Cost of inaction
The rows are illustrative and deliberately cautious: they assume gaps of per-milles, not per-cents. The point of the table is not the total but the proportion: after automation the per-delivery cost of control is close to zero, and each of the items above disappears or shrinks with the first month of full coverage.
There is also the opposite line worth remembering: some gaps run in the buyer’s favour. Automated reconciliation is not a weapon against the supplier; it is a way for both sides to look at the same numbers.
A model organisation with realistic proportions – the numbers exist so you can run the same maths on your own data; they are not a client result.
A forty-station network with its own fuel purchasing desk: deliveries from two fuel depots and one wholesaler, remote-read gauges, an ERP with a purchasing module, Microsoft 365 at head office.
260 tanker deliveries a month, around 2.8m litres a quarter, two people at head office responsible for purchase settlements.
Managers sign deliveries on the document, head office books invoices against documents, and temperature-compensated checks are done on a sample, a dozen or so a month.
A full check of one delivery takes about 25 minutes: gathering the document, the gauge readings, the tables and the invoice. Which is why it is done for one delivery in ten, and claim deadlines pass faster than the checking queue.
The robot assembles a complete set for every delivery: notification, delivery note, before-and-after gauge readings with temperatures, and the invoice; converts everything to 15°C, compares, and closes matching deliveries. An out-of-tolerance gap becomes a Teams case with the calculation and a full claim attachment set.
In the modelled case control covers one hundred percent of deliveries from the first week, and the first conversation with a supplier about a series of gaps happens on data, not impressions. Model numbers, not anyone’s ledgers.
Proposed solution
We start with the settlement rule: what counts as natural tolerance (transport, temperature, gauge accuracy) and what is a gap to explain, per supplier and per depot. The rule is written down with your purchasing desk and versioned; it, not intuition, decides every delivery.
For every notified delivery the robot completes the documents: the depot note (read from a PDF or a photo if that is how it arrives), gauge readings before and after the drop with temperatures, the invoice from the ERP. It converts volumes to 15°C using the density tables for the product and compares the full set of numbers. Deliveries within tolerance close themselves and feed the supplier’s statistics.
An out-of-tolerance gap becomes a Microsoft Teams case: with the calculation, all documents and a countdown of days to the claim deadline. A person decides whether to claim, and does it with a single approval, because the letter with attachments is already prepared. Statistics per supplier and per depot show where gaps are systematic and where they are random.
UiPath Orchestrator: a delivery queue, event triggers after the drop, retries and an audit trail; UiPath Document Understanding for reading delivery notes; UiPath Integration Service connectors for Microsoft Teams and Outlook 365; UiPath Action Center notifications
The tolerance rule per supplier and product with versioning, per-delivery document assembly, conversions to 15°C using density tables, the comparison engine, claim cases with a deadline countdown, and supplier statistics
Delivery notes from depots (PDF, photo or EDI), gauge readings via the tank monitoring system, notifications and invoices from the ERP via file interface or API
How the automated process works
- AutomationA notified delivery opens a case: the robot waits for the set, note, before-and-after measurements, invoice
- AutomationThe delivery note is read automatically, from PDF or photo too, with litres, temperature and density
- AutomationVolumes are converted to 15°C using the product’s tables and compared with the measurement and the invoice
- SystemDeliveries within tolerance close themselves and feed supplier and depot statistics
- PersonAn out-of-tolerance gap reaches purchasing in Microsoft Teams with the calculation, documents and days left to claim
- PersonThe claim decision is one approval; the letter with attachments is prepared, and the supplier’s reply returns to the case
Human-in-the-loop model
Automation handles
- Assembling the documents for every delivery and reading the note whatever its format
- Conversions to 15°C, comparison of the full number set and self-closing of in-tolerance deliveries
- Preparing claims with attachments, watching deadlines and keeping supplier gap statistics
People decide
- The tolerance rule per supplier, product and route, and every change to it
- The decision to file a claim and to escalate on a series of gaps
- Conversations with suppliers and depots, held on a series of documented deliveries
Before and after
Systems and integrations
The stack is short on purpose: one engine, one execution layer, one place where a person decides.
Inputs
- delivery notifications and invoices in the ERP
- delivery notes from fuel depots
- gauge readings before and after the drop, with temperature
- the tolerance rule per supplier and product
Automation layer
- UiPath Orchestrator
- UiPath Robots
- UiPath Integration Service
- UiPath Action Center
Target systems
- a register of settled deliveries with conversions
- claim cases in Microsoft Teams
- gap statistics per supplier and depot
Human touchpoints: UiPath Action Center cases in Microsoft Teams; a monthly supplier statistics review with purchasing
Technologies used
a delivery queue, post-drop triggers, retries, credential store, full audit
Areading delivery notes from PDFs and photos: actual and 15°C litres, temperature, density, seals
Acases for purchasing, claim dispatch, return correspondence into the case
Aone-click claim approval where purchasing already works
Astandard practice in settling liquid fuels; conversion parameters per petroleum product standards
BHow we deliver this
From this description to a running process
We deliver this solution through our process automation service: first we measure the process on company data, then we build, test and maintain the robots. SAP integration is handled by our SAP team.
Industries
Case studies
Delivered client work is described under Case studies.
Illustrative economic model
Numbers you can check against your own data.
The first two rows price a control nobody performs in full today because it is too expensive; automation makes it free per unit. The third row is a cautious recovery model: half a per-mille of systematic gap belongs to this scenario, not to a client’s ledgers. €26 an hour is an example full cost of a purchasing seat.
Run the numbers on your data
An illustrative estimate based on your inputs. It models freed capacity, not promised savings.
Business benefits
- Control covers every delivery, not a sample, and costs per unit about as much as a register entry
- Claims are filed on time, with before-and-after measurements and a conversion that is hard to dispute
- Systematic gaps surface as a series, not as an isolated suspicion
- Accounting pays for the fuel that arrived, not for the document that rode along with the driver
- Supplier and depot conversations run on shared, verifiable numbers
The management view
- The tightness of fuel purchasing stops depending on whether someone had 25 minutes at delivery time
- Every delivery has a record: documents, measurements, the conversion, the decision and the claim outcome
- Exposure to supplier disputes falls, because evidence is created on the day of the drop rather than reconstructed a week later
Board-level KPIs
share of deliveries settled with temperature compensation · out-of-tolerance gaps per hundred deliveries · claims filed within the deadline · amount recovered from claims per quarter · time from drop to settlement closure
Security and governance
The automation has exactly the permissions it needs. Not one more.
- The robot reads the ERP and the gauge system with a minimal-rights technical account; it does not post or change invoices, it keeps its own settlement register
- Every settlement keeps its source documents, readings, table version and tolerance rule, so the result can be reproduced to the litre
- The tolerance rule is versioned and changed only by the designated purchasing owner
- Personal data is limited to the driver details as they appear on the delivery note; statistics deal in suppliers and depots
- Cases and registers stay in your Microsoft 365 tenant, and the robots run in the EU region of UiPath Automation Cloud
Why now
Fuel retail margins do not forgive purchasing losses: half a per-mille of systematic gap is a line visible in the annual result and invisible in any single delivery.
The data already exists: gauges measure before and after every drop, the ERP knows notifications and invoices, and delivery notes can be machine-read. The only missing piece is the layer that joins them on delivery day.
Contractual claim deadlines are short and do not negotiate backwards. The only way to use them is to have complete evidence before they pass, which means automatically.
Relevant executive roles
One hundred percent of deliveries under control, and gap statistics per supplier instead of sampled suspicions
We pay for 15°C litres that arrived; the gap between document and fuel has an owner and a procedure
Receiving a delivery stops requiring the station manager to be an inspector he never had time to be
Common questions and objections
Most likely, and that is exactly why it is worth measuring. The tolerance rule separates physics from anomaly: what is natural closes itself. If after three months the statistics show a clean distribution around zero, that is a result too: peace of mind backed by data rather than by absence of control.
They do not have to do anything beyond what they do: measurements come from the gauges automatically, keyed to notification times and the drop event. Where there are no gauges, the rule works in a reduced variant, on documents and the invoice, and still catches gaps between those.
A 15°C document is half the settlement. The other half is how much fuel physically arrived in the tank, and only that pair says whether the transport loss is within the norm. Besides, the invoice is often issued from a different system than the note, and those two numbers can differ as well.
When this is not the right solution
- A dozen deliveries a month from a single depot: a spreadsheet with a density table and measurement discipline will do, the automation will not pay back
- No remote-read gauges and no willingness to install them: without before-and-after measurement what remains is a document-versus-invoice check, poorer though still sensible
- Purchasing exclusively on terms without a right to claim losses: settlement then has informational value but no recovery value
A question for the next management meeting
How many of last quarter’s fuel deliveries were settled with temperature compensation and a before-and-after measurement, and how many did we pay against the document alone?
Implementation approach
Scope without ambiguity, before anything is signed.
We deliver
- The tolerance rule per supplier, product and route, written down with purchasing and versioned
- Automatic assembly and reading of every delivery’s documents, PDFs and photos included
- Conversions to 15°C, comparison of the full number set and a settlement register with full history
- Claim cases in Microsoft Teams with a deadline countdown and a letter ready to send
- Gap statistics per supplier and depot, and a parallel run through the first month of deliveries
We need from you
- Three months of delivery history: delivery notes, invoices, gauge readings, past claims
- A designated purchasing owner to approve tolerances and own the rule
- Access to the ERP (notifications, invoices), the gauge system and the mailbox where depot documents arrive
Stages
Discovery
A month of deliveries read document by document: note formats, measurement completeness, the actual gaps
Rules
Tolerances per supplier and product and claim principles approved by purchasing
Build
Document reading, conversions, comparison, Teams cases and supplier statistics
Parallel run
The robot settles all deliveries, current practice runs alongside, differences explained delivery by delivery
Go-live
Full coverage with post-go-live care and a tolerance review after the first month
A quick win. Effort depends on the number of delivery-note formats, on gauge readings being available with a drop timestamp, and on whether wholesaler invoices arrive electronically.
Tuesday’s delivery: 32,000 litres at 15°C on the invoice, 31,700 arrived in the tank. Nobody noticed, because nobody counted.
Send us the documents of your last ten deliveries: notes, invoices and gauge readings if you have them. We return a full settlement of each with temperature compensation, and a note of which would have qualified for a claim.
Send your last ten deliveries for a checkThe neighbouring process usually has the same problem
The gauges say one thing, the tills another. A daily balance says why, station by station.
See the solution Finance & accountingThree-way match exceptions resolved before the payment runThree documents, three numbers, and an invoice waiting for someone to find the fourth.
See the solution Finance & accountingSupplier statements reconciled the day they arriveYour suppliers tell you every month what they think you owe. Most of those letters go unread.
See the solution