Home · Solutions · Finance & accounting · Every station’s takings reconciled each morning
Solution · Finance & accountingThe day report, the terminal, the cash and the courier line up by themselves; people see only the exceptions
Every station’s takings reconciled each morning
Every morning a robot lines up all of yesterday’s streams per station and shift, posts the matching closures, and raises differences as cases with the amount, the shift and the history of similar incidents at that station.
Executive summary
A station’s daily close is several streams that accounting stitches together late and on a sample. Cash differences dissolve between shifts, and nobody has time to see the patterns.
All of yesterday’s streams reconcile in the morning, automatically, per station and per shift. Matching closures post themselves.
A difference is a case with the amount, the shift and that station’s history; a recurring pattern raises priority. Disciplinary decisions stay with people.
BOS/POS day reports; payment processor settlements; bank statements; cases and the difference board in Microsoft Teams
Business problem
Five streams, one desk
A station’s takings are not one number. Tills close the day with reports, terminals settle through processor files with their own timings and fees, cash goes to the bank via deposit machines or couriers, fleet customers buy on invoice, and the shop and car wash add their own lines. Every stream has its own rhythm and its own documents.
Accounting stitches it all together after the fact: station reports, processor files, bank statements. With dozens of stations the stitching takes days, so it happens on a sample and a week behind. A difference found on Tuesday concerns a shift from ten days ago that the manager no longer remembers, and the CCTV recording may be overwritten.
The expensive thing is patterns, not single amounts. Fifty zloty of difference now and then is the cost of running a station; the same fifty zloty on the same shift every Thursday is a phenomenon. Telling them apart needs a history per station and shift, which a manual process never builds because it barely keeps up with the present.
On top sits the daily nobody’s-work: assigning statement deposits to stations and days, explaining processor fees, chasing missing reports. Work that does not require an accountant, yet consumes accountants.
How it works today
Below is what the work looks like before anything is automated.
- PersonStations send day reports; some arrive late, some with corrections
- PersonSomeone in accounting matches statement deposits to stations and days, line by line
- Risk of errorTerminal settlements differ by fees and date shifts; differences are explained on a sample
- WaitingA full day reconciliation appears a week behind, or not at all
- Risk of errorA cash difference reaches the manager when the shift is a memory and the recording overwritten
- Risk of errorPer-shift difference patterns do not exist, because nobody has time to compute them
Why the current process costs more than it appears
The bill that never shows up in a budget.
- Unexplained differences get written off to costs, and their yearly total surprises everyone who computes it for the first time.
- A week of lag is a week in which the pattern keeps working: the Thursday difference will occur once more before anyone asks about the first.
- Accountants’ hours on matching deposits and fees: reproductive work in a team that owns closings, filings and audits.
- Station managers lose trust in settlements they do not understand and that arrive too late to check anything.
Cost of inaction
The total of unexplained differences differs per network and is almost always underestimated by management, because it is booked in small amounts all year. The model takes a moderate value; the first quarter of daily reconciliations will show the real one.
The process has a second side too: some differences are errors in the customer’s favour or processor mistakes nobody claims back. Daily control works both ways.
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 central accounting: a BOS collecting day reports, two payment processors, deposit machines at some stations and couriers at the rest, fleet sales on invoice, SAP FI, Microsoft 365.
Around 1,200 shift closures a month, 3,600 deposit and settlement lines to match, two accounting people dedicated to station settlements.
Reconciliations done in batches, two-three days a week, a week behind; differences emailed to managers; no per-shift difference history exists.
The backlog is structural: something is always unreconciled, and difference conversations concern events nobody remembers.
Every morning the robot pulls day reports, processor files and statements, matches the streams per station, day and shift, posts matching closures in SAP and raises differences as cases with full context and that station’s pattern history.
In the modelled case the difference conversation with a manager happens the next day, with the CCTV still available, and accounting recovers most of two roles’ time. Model numbers, not anyone’s ledgers.
Proposed solution
We start with the reconciliation rules: what makes up a station’s day, what tolerances each stream carries (processor fees, roundings, deposit date shifts), what is a difference to explain and what an accepted cost. The rules are written down with accounting and versioned.
Each morning the robot collects yesterday’s complete day: BOS day reports, processor settlements, bank statement lines, fleet sales from the ERP. It matches the streams per station and shift, allowing for processor settlement timing and deposit value dates. Matching closures post automatically in SAP under your posting scheme.
A difference becomes a Microsoft Teams case assigned to an owner: with the amount, the shift, the components and the history of similar differences at that station. A recurring pattern, same shift, same time, similar amount, raises the case’s priority and shows on the difference board. Decisions about people are always made by people; the robot delivers the facts while they are fresh.
UiPath Orchestrator: morning runs, queues per station and stream, retries and an audit trail; UiPath Integration Service connectors for Microsoft Teams and Outlook 365; UiPath Action Center cases; a Microsoft Power BI difference board
Reconciliation rules with per-stream tolerances, five-stream matching per station and shift, automatic posting of matching closures, difference cases with pattern history and the network board
Day reports from BOS/POS via file interface or API; payment processors’ settlement files; bank statements (MT940 or your bank’s format); fleet sales and the posting scheme in SAP FI
How the automated process works
- AutomationIn the morning the robot pulls day reports, processor files, statements and fleet sales for yesterday
- AutomationStreams are matched per station, day and shift, with fee and value-date tolerances
- SystemMatching closures post automatically in SAP under the agreed scheme
- PersonA difference reaches its owner as a case: amount, shift, components, that station’s history
- AutomationA recurring difference pattern raises priority and is marked on the network board
- PersonThe manager and accounting explain it the next day, with the recording and the shift’s memory still available
Human-in-the-loop model
Automation handles
- The morning stitch of all the day’s streams per station and shift, the same way every day
- Posting matching closures and running difference cases with full context
- The pattern history per station and shift, and the network difference board
People decide
- The reconciliation rules, tolerances and posting scheme, and every change to them
- Difference explanations and all decisions concerning employees
- Claims to processors and the bank, prepared by the robot, sent after a decision
Before and after
Systems and integrations
The stack is short on purpose: one engine, one execution layer, one place where a person decides.
Inputs
- till day reports from BOS/POS
- payment processors’ settlement files
- bank statements with deposits
- fleet sales and the posting scheme in the ERP
Automation layer
- UiPath Orchestrator
- UiPath Robots
- UiPath Integration Service
- UiPath Action Center
Target systems
- automatic closure postings in SAP FI
- difference cases in Microsoft Teams
- the difference and pattern board in Power BI
Human touchpoints: UiPath Action Center cases in Microsoft Teams; the difference board as a Teams tab; a weekly pattern review with operations
Technologies used
morning runs, per-station queues, retries, credential store, audit
Adifference cases, missing-report reminders, processor claims
Aexplanations and decisions where accounting and regional managers already work
Aautomatic closure postings under your scheme, via documented interfaces
Athe difference board: network, stations, shifts, patterns and case age
AHow 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.
Halving the differences is a model assumption based on a mechanism, not a promise: differences shrink because the question is asked the next day and patterns stop working for weeks. The first quarter will show the real value; it comes out lower at some networks and markedly higher at others, depending on what the “cash differences” line hides today.
Run the numbers on your data
An illustrative estimate based on your inputs. It models freed capacity, not promised savings.
Business benefits
- Every day of every station is reconciled by morning, with no structural backlog
- Differences are explained while evidence exists: the recording, the shift’s memory, yesterday’s documents
- Patterns stop being invisible; the recurring Thursday difference gets priority automatically
- Accounting stops being a deposit-matching machine; exceptions and decisions remain
- Processor and bank mistakes surface daily and are claimed within deadlines
The management view
- The total of unexplained differences becomes a managed number with a trend, not a yearly surprise
- Every difference has a case, an owner and a recorded resolution
- Internal control receives a pattern history instead of anecdotes
Board-level KPIs
days reconciled by next morning · monthly total of unexplained differences · time from difference to resolution · recurring patterns detected and closed · processor claims filed within deadlines
Security and governance
The automation has exactly the permissions it needs. Not one more.
- The robot posts only closures matching the rules, under a scheme approved by accounting; everything else waits for a person
- Every reconciliation keeps all streams’ components, so every amount is traceable to the report, the file and the statement line
- Rules and tolerances are versioned; only the designated accounting owner changes them
- Difference cases deal in stations and shifts; employee data appears only where an explanation requires it, and only for the authorised
- Cases and boards stay in your Microsoft 365 tenant, and the robots run in the EU region of UiPath Automation Cloud
Why now
Payments have splintered into ever more streams: terminals, mobile schemes, apps, fleets. Manual stitching handles more files every year with the same team.
Cost pressure in retail leaves no room for roles spent on reproductive work an automation does better, because daily.
Evidence has a short shelf life: recordings get overwritten and shift memory does not reach ten days back. A reconciliation’s value depends directly on its speed.
Relevant executive roles
Station settlements stop producing a structural backlog, and the difference total has a trend instead of a yearly surprise
Per-shift difference patterns are visible systemically, with a history that previously did not exist
Difference conversations concern yesterday, not last month, and rest on components rather than suspicions
Common questions and objections
And that is half the work: the report says how much there should be. The reconciliation says how much actually arrived: through the terminal, the bank and the courier, after fees and date shifts. The automation stitches exactly the half the BOS cannot see.
The explaining, yes, and it stays that way. What changes is that the manager receives the case the next day, with components and history, instead of an email after two weeks. The same accountability, with tools that make it bearable.
Only closures matching the rules your accounting wrote and approved post automatically, under your scheme. Everything outside the rule waits for a person, and every entry has a full lineage. That is less discretion, not more.
When this is not the right solution
- A single owner-run station: one person closes the day in twenty minutes and sees everything herself
- No electronic day reports: first a BOS/POS with exports, then reconciliation
- The expectation that the system will name the culprit: the system shows facts and patterns, decisions about people remain people’s decisions
A question for the next management meeting
What did unexplained cash differences total across the network last year, and who watches that number’s trend month by month?
Implementation approach
Scope without ambiguity, before anything is signed.
We deliver
- Reconciliation rules with per-stream tolerances, written down with accounting and versioned
- The morning reconciliation of all stations and shifts with automatic posting of matching closures
- Difference cases with pattern history and the network board in Power BI
- Processor and bank claims prepared automatically for decision
- A two-week parallel run and post-go-live care
We need from you
- A month of history: day reports, processor files, statements, current reconciliations
- A designated accounting owner for the rules and the posting scheme
- Access to BOS/POS, processor files, banking and SAP FI
Stages
Discovery
A month of closures reconstructed stream by stream: formats, timings, the actual differences and their fates
Rules
Tolerances, the definition of a difference and the posting scheme approved by accounting
Build
Collectors, matching, postings, cases, patterns and the board
Parallel run
Two weeks: the robot reconciles in parallel, differences against the manual process explained to the zloty
Go-live
One region first, then the network, with a rules review after a month
Effort depends on the number of day report and processor formats and on the cleanliness of the posting scheme; single-BOS networks implement fastest.
Thursday’s difference: minus 180 zloty, night shift, station twelve. Same amount, same shift, fourth time this quarter.
Send us one station’s week of closures: day reports, processor files and the statement. We return each day reconciled by our rules, the differences with their components, and the first patterns, if any appear.
Send one station’s week of closuresThe neighbouring process usually has the same problem
The gauges say one thing, the tills another. A daily balance says why.
See the solution Finance & accountingFleet cards reconciled, anomalies visible at onceA fill-up 300 km off route stops waiting for month-end.
See the solution Finance & accountingSupplier statements reconciled the day they arriveYour suppliers tell you every month what they think you owe.
See the solution