News and knowledge base
Flowtion Insights
October edition · business processes, integration and AI
News and knowledge base
Technology news, from a business perspective.
The Flowtion Review is a monthly professional selection on business process digitalisation and Flowtion development. Alongside industry news, we explain what changes mean for everyday operations and how Flowtion solutions can use them.
Industry perspective
Business process digitalisation, artificial intelligence, document processing and system integration: an assessment of meaningful changes, their relevance to Flowtion and their implications for customers.
Flowtion updates
New features, product developments and practical use cases. We explain the business needs each development addresses and how it can fit into existing operations.
October edition
When data connectivity becomes business control.
The common question behind this month’s topics is how external regulation, AI, partner collaboration and field work become controlled, working business processes.
NAV and reporting
NAV M2M: when is the standard solution enough?
Current topic: M2M is not an end in itself: it becomes complex when several systems, missing localisation and corporate control workflows must work together.
Standard Hungarian accounting software may include M2M as a built-in capability. Flowtion is most valuable where data is spread across several systems, the ERP lacks suitable Hungarian localisation, or the NAV connection must be extended with a unified data model, internal validation, approvals and an auditable log. The NAV eÁFA-M2M 2.0 specification is a good opportunity to review source systems and controls.
Related professional article: NAV eÁFA-M2M 2.0 specification – changing data structures and controls make a review of source systems worthwhile.
Integration and ERP
The role of an integration layer in a fragmented system landscape
Current topic: Automation and AI can scale only when the company’s data and systems are connected reliably and under controlled rules.
Alongside the ERP, many companies operate specialist and finance systems, bank data sources and regulated external connections. The goal is not another isolated interface, but a unified data model, controlled data transfer and traceable error handling.
Flowtion Data Exchange receives, maps and forwards data from different sources in a unified structure. It handles scheduled and event-driven exchange, isolates faulty records, supports retries, and logs statuses and responses. With Studio, customer-specific business rules—such as general-ledger, partner, project or tax-code checks—can be built into the workflow without developing a separate interface. Flowtion can therefore connect an ERP, bank data source, NAV or another regulated external system as middleware, while the operation remains supervised in one place.
Related professional article: AI fails without a strong operational, data and ERP foundation – fragmented ERP and data environments make automation difficult to scale.
AI and automation
AI value is measured in the document workflow
Current topic: Document recognition alone is not a business result; value comes from validated, approved data embedded in a workflow.
McKinsey’s 2026 analysis highlights the need to continuously validate, monitor and evaluate AI outputs with appropriate human oversight. In enterprise document processing, recognition should therefore be followed by business rule checks, exception handling, approval and logging within the workflow.
The Flowtion layer built on Azure AI Document Intelligence extracts structured data from incoming PDFs, scanned documents and other files, then routes it into the appropriate business workflow. It can run field-level checks, request human review when values conflict or confidence is low, and after approval create an invoice, bank transaction, task or settlement item. The document and processing result remain traceable together, so AI becomes part of a controlled workflow rather than a separate experiment.
Related professional article: McKinsey: AI data readiness – without continuous validation and human oversight, AI output cannot form a reliable business process.
Partner digitalisation
A partner portal is part of operations and market presence
Current topic: Partner digitalisation reduces administration, improves transparency and makes the company’s external operations more professional.
Supplier and grant processes are increasingly moving toward shared, self-service interfaces. A portal gives external parties structured access to their own data and documents while improving transparency, confidentiality and the company’s professional presence. Würth’s procurement digitalisation illustrates this direction.
In Flowtion Connect, the portal is more than an online form. An applicant, supplier, subcontractor or customer uploads data and documents through a role-specific interface, follows status, supplies missing information and receives notifications. Behind the scenes, permissions, deadlines, checks, approvals, versioning and full logging are managed. Connect can forward the data to BusinessDesk, an ERP or another enterprise system, so information submitted through the portal does not remain in a separate database but can immediately support assessment, contracting, procurement or settlement.
Related professional article: Würth and SAP Business Network – a shared partner platform supports integrated procurement and customer experience.
Field operations
Labour shortages are also a scheduling and knowledge challenge
Current topic: Field efficiency depends not only on mobile work orders, but also on connecting capacity, skills and delivery to the business process.
As service and maintenance companies grow, dispatcher decision support, skills-based resource scheduling, mobile work orders and management of on-site evidence become critical.
In FieldOps, tasks can be generated from service contracts and registered facilities, equipment or technical systems. The dispatcher sees capacity, employee skills, location and required equipment in a calendar, then assigns and reschedules work accordingly. The technician records the work on a tablet, including materials, faults, photos, follow-up tasks and customer acceptance. Based on contract prices and service levels, the system automatically produces the basis for service acceptance and billing. FieldOps therefore connects the mobile work order to the contract, resource, acceptance and settlement.
Related professional article: How to pick HVAC field service management software in 2026 – scheduling, mobile work and billing integration support growth.
Featured Flowtion development direction
A new data collection and validation layer for NAV processes.
Flowtion’s M2M concept is not intended to replace standard accounting software. It adds an integration and control layer for complex environments where a unified transaction-level data model must be built from several source systems, or localisation and the company’s approval model need to be extended.
Data model
Unified data from different sources
Invoice and transaction data is placed in a unified structure, while source-system differences are handled in Flowtion. This is particularly useful when accounting, billing, project or bank data is generated in several systems. Flowtion builds a common transaction view from the source data while preserving the original source, transformation rule and processing status. The data can therefore support business and management review, not only submission.
Validation
Control before submission
The layer checks dates, domestic tax numbers, tax-code consistency, missing or duplicate items and agreement with NAV Online Invoice data. Differences do not disappear into a technical error log: the system shows which record, rule and source data caused the issue. Correctable errors can be returned to the responsible user, while disputed cases can be escalated for a decision before submission.
Approval and logging
Traceable operations
Submission can require approval; versions, NAV requests, responses, error messages and timestamps are logged for traceability. Roles can define who prepares the data, approves submission and handles differences. This separates automated processing from professional judgement and makes it possible to reconstruct what happened during a reporting cycle.
The actual implementation starts as a project tailored to the company’s source systems, data, tax processes and operating requirements.