Service · Software development

Custom document workflow automation.

Custom software that automatically ingests your documents, classifies them, enriches them with data, routes them to the right people, gets them approved, archives them and passes them on to your core systems. For organisations where standard low-code runs up against its limits: deeper OCR, industry-specific rules, and integrations that cannot be handled with an off-the-shelf connector.

What exactly is document workflow automation?

Document workflow automation, also known as document workflow software, is the digital orchestration of the entire process a document goes through, from the moment it arrives to the moment it has been processed, archived and passed on to the downstream systems. In practice this means that an incoming invoice, contract, policy, complaint or HR document is automatically recognised, classified, enriched with the right data, sent to the right reviewer, approved according to your rules and finally passed on to your accounting system, ERP, CRM or case management.

It is important to be clear on this up front: this is different from a document management system (DMS). A document management system is the storage location: versions, access rights and searchability of your documents at rest. Document workflow automation is about who does what with a document while it moves through your organisation. The two complement each other and in modern implementations are almost always integrated, but they are different problems with different solutions.

We are called in when the volume of documents and the complexity of the rules grow beyond what a standard low-code tool can comfortably handle. For simple flows, such as a few hundred documents a month, a handful of fields and a single approver, a Power Automate, Make or Zapier flow is often good enough, and we actively advise against custom development. Our sweet spot is where you need OCR with industry-specific fields, a rules engine with dozens of conditions, or a deep integration with your core ERP, and where the volume or compliance requirements are serious enough to justify building a dedicated platform.

Our approach in four points.

Document-first
We model your document flows before we choose technology
OCR + AI
Industry-specific extraction, not generic template OCR
Deep integration
Integrations with ERP, CRM, accounting and case management
Audit-ready
Full audit trail and compliance reporting built in as standard

The document workflow projects we build.

Document workflow automation is broad. We see three patterns come up again and again, and most clients start with one and grow organically into the other two.

Engagement 01

Invoice and purchasing document automation

Incoming invoices, PO matching and accounting integration

Incoming invoices, whether by email, scan, Peppol or portal, are automatically recognised, processed by OCR and matched against the open purchase order. The flow then runs via the right budget holder, finance and, optionally, a second pair of eyes for amounts above a threshold, and pushes the completed invoice straight to your accounting system or ERP.

The value isn't in OCR alone, which any standard package can do, but in the logic on top of it: PO matching with tolerance rules, tax checks per supplier country, automatic routing to the right general ledger account and cost centre, and exception handling that doesn't bring all the manual work back the moment a single field deviates. We often combine this with a smart API integration to your accounting system, so journal entries are generated automatically instead of via CSV import.

OCR extractionPO matchingTax checksApproval flowERP pushPeppol
Engagement 02

Contract and compliance document workflow

Templates, redlining, sign-off and legal archiving

For organisations where contracts, NDAs, policies or permits are the core process: a platform that manages templates, generates draft contracts from CRM data, records redlining between parties, routes them for legal and commercial approval and finally has them signed with a digital signature, including a full version history and audit trail that meets what your regulator or auditor requires.

The custom layer lies in the rules: which clause commercial may negotiate without legal, which combination of fields requires a second approver, which type of contract must be renewed within which period. Standard contract management tools cover the basics, but for sector-specific rules, such as insurance underwriting, commercial property leases or grant applications with multiple dependent documents, we build the rules engine ourselves so that it fits your acceptance policy precisely.

Template engineRedliningMulti-level approvalDigital signatureAudit trailRenewal alerts
Engagement 03

Case and file workflow for service providers

Multi-document files with routing and compliance

For accountancy firms, law firms, notaries, insurers, municipalities and HR departments where a file consists of dozens of different documents, each taking its own route through the organisation. HR onboarding with CV, proof of ID, contract, NDA and GDPR consent; claims handling with claim form, photos, valuation and legal documents; grant applications with application form, annual accounts, project plan and verification report.

The difference from standalone document flows is that everything runs in conjunction: a file is only complete when all underlying documents have the right status, and the final step waits for the slowest document. That's why we build a file state model on top of the individual document workflows, with dashboards for case managers and automatic alerts when a file is at risk of stalling. For accountancy organisations this often sits alongside our accountancy software; for mid-market and larger organisations it is part of a broader enterprise software programme.

File stateMulti-document routingCompliance checksCase dashboardDeadline monitoringMulti-tenant

What you have at the end.

A working document workflow platform in production, plus everything around it so you can manage and extend it yourself.

Production + staging

Two environments, in your own cloud or hosted by us. Your choice.

OCR and rules engine

Trained on your documents, with a no-code admin UI for rules.

Codebase + docs

Full source code, build instructions and architecture overview.

Key-user training

Sessions for process owners and finance, video for end users.

Management (optional)

Monitoring, backups, model retraining and ongoing development.

When custom is the right choice.

Standard platforms such as Power Automate, Make, Zapier, Kissflow, Nintex, DocuWare and M-Files are excellent for many flows. We get called in when one or more of these patterns applies.

Extraction

OCR with industry-specific fields

Standard OCR handles generic invoice fields well but falls over on specialist documents: medical results, technical drawings with annotations, policy schedules with exclusions, or valuation reports with signed elements. We train models on your own documents so extraction is reliable enough to genuinely remove a large share of the manual work.

Lines

Complex rules engine that doesn't fit low-code

A Power Automate flow with ten steps is perfectly manageable. Fifty conditions that depend on amount, supplier country, contract type, customer history and sector-specific regulation make maintenance a nightmare. A custom rules engine with clear semantics, visual or in code, keeps scaling where low-code turns into spaghetti.

Integration

Deep integration with core systems

When documents aren't standalone output but part of a transaction in your ERP, CRM or case management system, such as generating journal entries in SAP, processing policy changes in an insurance core system or closing permit files in a municipal case management platform, standard connectors usually fall short. A custom integration then delivers true two-way sync instead of ad hoc imports.

Multi-tenant

One platform for multiple clients or branches

For accountancy, law and agency organisations that offer document flows to their own clients: one platform with separate data pools per client, their own branding and their own workflow templates per client segment. That kind of multi-tenant architecture is rarely built gracefully in low-code.

Audit

Compliance requirements that low-code can't handle

GDPR implementing legislation, the Wwft (Dutch anti-money laundering act), FATCA, Solvency II, GDPR for medical data, or sector-specific quality standards often call for immutable logging, electronic signatures in a specific form, or retention period enforcement that a standard package cannot handle, or only partly. Your own platform gives you the control to build in every compliance requirement exactly as needed.

Scale

Volumes that low-code can't handle

Hundreds of documents a day is no problem for most platforms. Thousands of documents a day with sub-second routing, parallel approval flows and peak load around month-end or quarter-end closing call for infrastructure you control yourself, with queues, retries and monitoring that have no hidden ceilings.

Not sure about a large project yet?

Test your idea first: a working prototype in 1 day

With OneDayBuild, we make your idea tangible in a single day for €1,150, so you know whether further development is worth the investment. Decide to go ahead with the full build? Then we deduct the cost in full.

View OneDayBuild →

How an engagement works.

01Introduction→ 02Document and process analysis→ 03Build in sprints→ 04Rollout and maintenance
Step 01

Introduction

A no-obligation conversation with your process owners and IT. Which document flows are causing friction, which systems are already involved, and what your volumes look like.

Step 02

Document and process analysis

We map out document types, variants, rules and exception paths. Workshops with finance, legal and the people on the shop floor. End result: a scope document with OCR targets, rules and integrations.

Step 03

Building in sprints

Working modules in 2-week sprints, starting with one document type end-to-end. Process owners test from the first sprint using real historical documents, giving direct feedback on extraction accuracy.

Step 04

Rollout and maintenance

A phased move to production: first one document type or one department, then scaling up. Ongoing maintenance, model retraining and extension with new document types.

What we integrate with and build on.

For each project we choose what fits your existing landscape. For the broader context of where this project fits, see our page on custom software; for the DMS layer where your documents ultimately end up, see our document management system page.

OCR & AI
Azure Document IntelligenceGoogle Document AIAWS TextractTesseractCustom LLM extractionHugging Face
Workflow & queues
TemporalCamundaBullMQRabbitMQPostgreSQLRedis
Integrations
SAPAFASExact OnlineTwinfieldPeppolMicrosoft 365SalesforceDocuSign

Frequently asked questions.

What is the difference between document management software and document workflow automation?
A document management system (DMS) is the repository: it controls where documents are stored, who can access them, how versions are managed and how you find them again. Document workflow automation controls who does what with those documents as they move through your process: arriving, being classified, having data extracted, being routed to the right people, being approved and being passed on to downstream systems. In practice, a workflow platform is often linked to a DMS: once a document has completed its workflow, it is automatically archived in the DMS with the right metadata and version information. The two systems therefore complement each other, but they solve different problems and need to be designed separately.
What exactly is document workflow software and how does it work in practice?
Document workflow software automates a document's path through your organisation. An incoming invoice or contract is picked up by a trigger (mailbox, upload, scan, API), automatically classified by type, and the relevant fields are extracted using OCR or AI. Based on content, amount or project, the document is routed to the right reviewer. It then moves through an approval chain, for example manager, finance and, where needed, legal, which can run in parallel or in sequence. At the end, the document is archived in a DMS with an audit trail, and the next action follows automatically: posting to the accounting system, preparing the payment, registering the contract or closing the case file. Good document workflow software handles all of this without manual steps for the exceptions.
Do you replace Power Automate, Nintex or Kissflow for our flows?
For simple flows with a few hundred documents a month and a handful of rules: no, and we would actively advise against it. Power Automate, Make, Zapier, Kissflow and Process Street work well for such use cases, are quick to set up and cheaper to maintain than a custom platform. Our added value comes into play when one of these limits is crossed: OCR with industry-specific fields, a rules engine with dozens of conditions, a deep two-way integration with your core ERP or CRM, multi-tenancy for your own customer base, or compliance requirements that the standard platforms do not cover, or only partially. In the introductory call, we will tell you openly whether your situation is better served by low-code or by a custom solution.
How reliable is OCR on our documents?
For standard document types such as invoices, packing slips and receipts, modern OCR engines from Azure, Google and AWS are approaching high reliability, especially on printed documents. For specialist documents, such as handwritten notes, technical drawings, policy schedules with tables and exclusions, or medical results, generic OCR falls short. In that case we train a dedicated model on your historical documents, measure extraction accuracy per field, and build a review step for fields below the threshold so that uncertain extractions are always confirmed by a person. To be honest: 100% automation is rarely realistic, nor is it desirable. Your exceptions are exactly where human expertise adds value.
How do you integrate the document workflow platform with our ERP, CRM or case management system?
Integrations are almost always part of a project, as pure document processing with no link to your core systems delivers little. We integrate with SAP via S/4HANA REST, BAPI or OData; AFAS via Profit connectors; Exact and Twinfield via their official APIs; and Salesforce, Dynamics 365 and HubSpot via their cloud APIs. For sector-specific case management systems, such as insurance core systems, municipal case systems or legal practice tools, we build the integration to order, based on the integration options the system offers. We document all integrations so that your IT team can later manage or extend them itself.
What about GDPR, audit trails and compliance?
An audit trail is a standard part of every document workflow platform we build: immutable logging of who viewed, edited, approved or rejected which document, and when. GDPR-compliant data processing is our starting point, with data residency in an EU region, encryption in transit and at rest, and clear retention-period enforcement per document type. For sectors with additional requirements, such as the Wwft (Dutch anti-money laundering act) for accountancy and legal practices, Solvency II and EIOPA for insurers, and GDPR for medical data in healthcare organisations, we build in the specific requirements from the start, not as an afterthought. For larger projects, a DPIA and optionally an external pen test are standard.
What determines the cost of a document workflow automation project?
Mainly four things. First, the breadth of the scope: invoice processing only, or the full range of documents including contracts, HR and case files. Second, the OCR complexity: standard invoice fields are cheaper than industry-specific models that need to be trained on your own data. Third, the number of integrations with external systems, since every ERP, CRM or case management integration is a mini-project in its own right. Fourth, whether we take care of monitoring, model retraining and ongoing development on a structural basis, or whether you handle that in-house. After the scope document, we give you an honest estimate with fixed sprint prices, so there are no surprises.
How long does it take to go live with a first document type?
A well-defined first document type, for example incoming invoices with a few suppliers and one accounting integration, can be in production within a few sprints. A broader platform with multiple document types, a complex rules engine and deeper ERP integration takes a project of several sprints. We always phase the work so that after the first few sprints, one document type is actually running in production. The business can then start phasing out manual work while we continue building the broader scope. Only once the first flow is running stably do we add the next document types.
Who benefits most from document workflow automation?
In practice, most enquiries come from mid-market and larger organisations with more than a hundred documents a day in workflow: accountancy firms with client-file flows, insurers with claims and policy-change flows, HR departments with onboarding and absence management, local and central government with subsidy, objection and permit processes, notaries and law firms with case administration, and real estate organisations with rental and sales administration. For smaller organisations with lower volumes, a standard low-code solution is usually sufficient and more proportionate. We're happy to give you honest advice on which side of that line your situation falls.
Do you work with our in-house IT department or existing supplier?
Almost always. A document workflow project will almost certainly touch existing systems: accounting, ERP, CRM, a legacy DMS, and sometimes a Power Automate layer built earlier that we take parts of over. We work transparently with your IT team and, where needed, your existing suppliers. Clear API contracts, documented integrations and an incident runbook are standard. Code, infrastructure as code and knowledge stay with you, so you never become dependent on a single agency.

Talk to us about your document workflow.

A no-obligation introductory call of half an hour. We listen to your document flows, the rules around them and your existing systems, and give you direction you can actually use, even if it turns out a standard platform is the better route.

Edit content