Tailored automation services from PointVAL
We connect systems, documents, payments and decisions into one measurable process built around the systems you already use. Technology handles the routine. People resolve the exceptions. Your business gets time back.
PointVAL is an automation service. It does not require a PointERP implementation. PointERP is a separate business system and our production-proven technology platform.
The PointVAL service
We design, implement and continually improve automation for a specific process in the client’s environment. We connect existing systems, add the missing capabilities and use proven PointERP components where they deliver measurable value.
Supplier invoice processing
From receipt to accounting-ready data
We connect KSeF and PDF invoices to supplier and product master data in the ERP system. We detect duplicates, retain the source document and prepare data for downstream processing or accounting import.
Payment matching and open-item clearing
From transaction to cleared open item
We retrieve transactions from banks, Revolut Business, Stripe and PayU. The system identifies the counterparty, proposes open items for clearing and learns from confirmed matches, leaving exceptions to the operator.
Spend and approvals
From purchase requisition to supplier invoice
We connect the purchase requisition, approval, purchase order, goods receipt and supplier invoice. Accountability and missing steps remain visible throughout.
Sales and order fulfilment
From lead to payment
We connect the lead, quotation, sales and purchase orders, delivery, sales invoice and payment. History and fulfilment status remain visible at line-item level, with data ready for accounting.
Sound familiar?
The same data is re-entered from a document into the ERP system, the accounting system and a spreadsheet
An invoice takes two weeks to move from receipt to accounting and passes through four pairs of hands
A bank transaction contains little identifying information, so someone manually searches for the counterparty and the open item it should clear
The status of an order line has to be reconstructed from messages, spreadsheets and several systems
The same counterparty exists in four master-data records, each with different details
When one person goes on leave, closing open work is delayed by a week
Each of these problems can be translated into hours and money. Almost no one has done that calculation – so the decision whether to change anything is currently made without evidence.
A good fit
Company size is not our qualification criterion. We look for a process where recurring volume, manual data hand-offs and exceptions can be measured.
Data passes through several places
A document, email, spreadsheet, ERP, bank or accounting system – and someone has to move the data between them.
Routine repeats; exceptions cause delays
Technology can handle standard cases. People should see only the situations that require a decision.
The process depends on one person
The current status and next step live in that person’s memory, messages or private notes.
The result can be quantified
We measure volume, staff time, the number and types of errors, and delays – before and after the change.
What we need from you
We need a process owner, access to representative data and the opportunity to observe the work as it is actually performed. You do not need complete documentation – we begin with the real process.
A shared starting point
We begin with a specific process, its owner and a business metric. Only then do we select the technology appropriate to the problem as defined.
A real client process
The same remote employee handles foreign supplier invoices and bank statements. Automation reduced the monthly workload for these two processes from approximately 11–12 hours to 1.5 hours.
11–12
hours of manual work in a typical month
1.5
hours per month after automating both processes
86–88%
less time spent on repetitive work
This is a conservative calculation based on typical monthly volumes and handling times. It represents 114–126 hours a year – roughly three working weeks recovered in a single role. It does not include the benefit of fewer errors or the shorter time needed to resolve exceptions.
Foreign PDF invoices
Manually entering an invoice – including supplier master-data checks, line-item mapping, date validation, the relevant National Bank of Poland exchange rate and currency conversion – takes an experienced operator approximately three minutes. A language model extracts data from a typical two-page invoice in 2–3 seconds. The system maps the supplier and line items, applies validation rules, and the operator approves the result. The entire process takes about 30 seconds per document.
Bank statements and payment matching
After importing a bank statement, the system identifies the counterparty and suggests the open items to be cleared. An operator decides whenever the match is uncertain. The confirmed decision is retained, allowing the next similar payment to be matched automatically. The resulting data is prepared in a format supported by the client’s accounting system or by its accounting firm.
A structured data flow
Management sees the current process status without separate reporting.
The employee can work remotely while the system maintains the current document and payment status, including cleared and outstanding items, as well as exceptions. The process no longer depends on private files or one person’s knowledge. Management information is created as part of everyday work.
Continuous improvement loop
Go-live starts the next measurement cycle.
We stay with the process. Together with the client, we monitor exceptions, recurring decisions and the steps that still consume time. We design the next improvements from real work and operational data, then measure their effect again.
Capacity for growth
For a role heavily burdened with similar tasks, a very conservative potential of 300–500 recovered hours per year frees up capacity equivalent to 0.15–0.25 FTE.
The recovered time can be used for management reports and analysis, customer service and exceptions that genuinely require the employee’s judgement, expertise and accountability.
How we work
We do not begin by transforming the entire company. We design and launch one measured process. After each stage, you decide independently whether the result justifies continuing.
Measurement and design
2–4 weeksWe establish the process time and cost, as well as the number and types of errors and exceptions. This gives us the basis for designing the first automation scope and the method for measuring its result.
First process
6–12 weeksWe automate one process and put it into production, with the improvement threshold written into the contract. Initially, it runs alongside the existing workflow, so you can compare the results directly.
Systems integration
3–6 monthsOnly after the first process has proved its value do we extend the foundation: integrations, reference data, validation at source and event logging. Subsequent processes build on the same foundation.
Subsequent processes
–Once the solution is live, we analyse the data, exceptions and recurring decisions. We design further improvements only where the results justify more work – and we do not start from scratch each time.
Performance commitment
At the end of the measurement stage, we jointly sign off the baseline: a specific metric, a specific definition and a specific value. For the first process implementation, the contract defines the improvement threshold, measured using the same method before and after go-live.
If we do not reach the threshold, we refine the solution at no additional charge – within the scope and under the terms set out in the contract. We can make that commitment because measurement starts on day one.
This commitment covers the first process and applies when the client-side conditions listed in the contract are met: access to data, a designated process owner and no change of scope during delivery.
Without a baseline, there is nothing to guarantee
Most projects of this kind go off track not because the technology fails, but because no one recorded the starting point. Without it, improvement cannot be proved – and a lack of improvement cannot be recognised early enough.
That is why measurement is a separate paid deliverable, not a free pre-sales exercise.
Where we begin
Process measurement
When you know that something consumes too much of your time, but not how much or where it is being lost.
Over two to four weeks, we measure the time, cost, errors and exceptions in one process. Your team contributes through a small number of interviews.
You receive a quantified process map, an agreed baseline and a prioritised list of three to five automation opportunities with estimated costs and returns. The measurement fee is credited against subsequent work, and the document remains yours regardless of your decision.
Discovery and solution design
When you already know what needs to be built and the question is how to deliver it – and at what cost.
We define the current and target process, metrics, preferred integration approach and fallback, first-release scope, key-screen wireframes, solution architecture and security design.
The result is an implementation plan, cost estimate and risk register specific enough to support a budget decision.
We begin with a short qualification call. A description of the process, its scale and the systems in use is enough for us to prepare a proposal for a paid discovery and solution-design engagement.
You do not need to know where to begin.
In 20 minutes, we will determine whether process measurement or discovery and solution design is the better first step.
Free 20-minute online call. Choose a convenient time.
Where the model stops
A language model can read an unstructured document and turn it into structured data. Matching is supported by rules, public registers and previous confirmations, while uncertain cases are routed to an operator. Financial outcomes remain under the control of explicit, testable rules. In our solutions, this boundary is defined precisely:
Every automated decision with a financial impact can be explained and reproduced. During an audit, “the model decided” is not an acceptable answer.
PointERP - a separate business system
PointERP can be implemented as a complete business system. PointVAL automation can also use selected, production-proven PointERP components - without requiring implementation of the entire system.
PointERP demonstrates that we can connect the entire flow – from lead and quotation through sales order and purchase order to sales invoice and payment. The system preserves the process history, provides line-level fulfilment status and prepares import-ready data for the client’s accounting system.
The process and integration approach are tailored to the client. The technology foundation remains proven.
2012
We have been developing the foundation of today’s PointERP in production since that year. We know its architecture, rules and dependencies because we own and maintain them ourselves.
Owning the technology lets us deliver improvements and new capabilities quickly as the client’s work evolves – without waiting for an external vendor’s roadmap, replacing a working system or losing knowledge built up over years.
Who we are
PointVAL is a brand of COMTRONIC Sp. z o.o.. We have been implementing business systems since 1991 and have helped our clients navigate every major regulatory change affecting Polish companies.
Let us begin with one process
During the call, we will establish where manual work occurs, the scale of the process and which systems it touches. We will tell you whether we see a sound case for measurement and automation, and what the next step should be. No product demonstration and no obligation.
Free 20-minute online call. Choose a convenient time.
Prefer to write first? office@pointval.com