# Process automation

> Camunda 8 workflows for back office and manufacturing. Document AI for incoming invoices, approvals and supplier workflows.

Recurring steps, approvals and data transfers run automatically instead of by hand.

## When it helps

Useful when a workflow currently consists of emails, spreadsheets and manual re-entry.

## Key points

- Fewer manual steps in repeating processes
- Clear approvals instead of email ping-pong
- Better traceability and auditability

## Details

We digitalise workflows that today run manually through your case handling and manufacturing. Camunda 8 as backbone, Kafka for eventing, document AI for unstructured inputs.

## When do you need this?

- **Your back office runs in Excel and email chains** — status is nowhere reliable, ownership is unclear.
- **You have a process** with clear steps that runs manually only because no off-the-shelf tool fits.
- **You have BPMN diagrams**, but nobody has turned them into productive code.
- **You want to measure** how long each step takes — and where the bottleneck sits.

## What we deliver

- **Camunda 8 workflows**: from BPMN diagram to productive application with webhooks, service tasks and human tasks.
- **Document AI** for classification and data extraction from incoming invoices, delivery notes, applications.
- **Event-driven architectures** with Apache Kafka for scalable, decoupled steps.
- **Escalation and SLA logic**: SLA breaches escalate automatically.
- **Dashboards** for business and management — lead times, bottlenecks, SLA fulfilment.
- **Integration with legacy systems**: SAP, Microsoft Dynamics, ProAlpha, Salesforce, custom case-management.

## Methodology

1. **Analysis call**: 60 minutes on one concrete process — where time is lost and which step is the biggest lever.
2. **Next step**: if the scope is clear, you receive an offer. For more complex processes we model the as-is and target state first (Domain Storytelling, BPMN) in a paid concept phase — including which steps get automated, which are supported and which stay manual.
3. **Implementation**: once commissioned, we build the workflows including happy path and escalation paths, delivered in 2-week sprints.
4. **Rollout & operations**: go-live in the real process, then operations and further development on request.

## Typical use cases

- **Incoming-invoice workflow** with document AI + Camunda + ERP integration (industry back office)
- **Claims processes for insurers** — from intake via triage to payout
- **KYC onboarding for banks** — automated pre-check, escalated compliance step
- **Supplier onboarding for industry** — master data, NDA, order confirmation as end-to-end workflow
- **Citizen-request management for public sector** — intake, classification, processing

## Camunda experience

We have used Camunda 8 (cloud or self-managed) productively since the first versions and have experience with:

- **Camunda Cloud** for lightweight workflows without infrastructure overhead
- **Self-managed** for data-residency requirements (EU AI Act, NIS-2, DORA)
- **Migration from Camunda 7 to 8** in productive environments
- **Integration into Spring Boot back-ends** and React front-ends
- **Tasklist UI** and custom front-ends for case handling

## Frequent questions

**Do we need Camunda Cloud or self-managed?**
Depends on data residency and compliance. We assess this in the analysis call and recommend the fitting variant.

**Can you take over existing BPMN diagrams?**
Yes. We import existing diagrams and check them for completeness and implementability.

**How do you integrate document AI into the workflow?**
As a service task in the Camunda process. Document-AI service returns structured data, the workflow continues based on it.

**What happens with model errors?**
Every AI-driven step has a human escalation path. Confidence thresholds and audit logs are standard.

## Technologies

- Camunda 8
- Apache Kafka
- Spring Boot
- REST APIs
- Java
- Python
- OpenAI API
- LangChain
- PostgreSQL

## Relevant links

- [HTML version](https://monadion.com/en/services/prozessautomatisierung)
- [English overview (llms.txt)](https://monadion.com/en/llms.txt)
- [Full AI context](https://monadion.com/en/llms-full.txt)
- [Book an analysis call](https://monadion.com/en/booking)
- [Contact](https://monadion.com/en/contact)

## Related

- [All services](https://monadion.com/en/services.md)
- [Industries](https://monadion.com/en/branchen.md)