# Automated workflows. Secure and traceable.

Source: https://vetosec.at/en/services/automation/

Collecting invoices, reading documents, connecting systems: what runs the same way every week becomes a workflow. We build it the way we build everything: with permissions, logging and approvals. Your data stays in your infrastructure or in the EU.

- **Rules** and AI
- **Own models** on your hardware
- **Human** approves where needed

## Examples

What we often automate

- [**Collect invoices**](https://vetosec.at/en/solutions/rechnungseingang-automatisieren/): Fetch invoices from mailboxes and supplier portals, name them, file them and pass them to accounting.
- [**Read documents**](https://vetosec.at/en/solutions/dokumente-auslesen/): Recognise forms, delivery notes or contracts and extract the important fields as data.
- [**Filing into the file cloud**](https://vetosec.at/en/solutions/managed-nextcloud/): Attachments land in the right Nextcloud folder automatically, with a meaningful file name.
- **Sync data between systems**: Shop, ERP and accounting stay in sync without retyping.
- **Generate reports**: Figures from several systems, automatically every week as PDF or mail.
- **Approvals and notifications**: A workflow pauses until a person approves and notifies the right people.

## From simple to complex

The right level for each task

1. **Fixed rules**: When the format is clear, scripts and workflows are enough. Predictable, cheap and easy to verify.
2. **AI workflows**: Language models read, summarise and classify. A person approves where it matters.
3. [**Own models**](https://vetosec.at/en/solutions/lokale-ki/): Self-hosted language models when data must not leave the building. Specialised ML models when a task is clearly defined.
4. **Specialists**: For research-grade questions we bring in experts from our network.

## Consulting

**A specialised model is often more accurate than one language model for everything.**

For a clearly defined task such as classifying receipts, a small own model is usually faster and cheaper to run and works on your hardware. We advise when a large language model makes sense and when an own model does.

## When it pays off

When automation pays off

### A good fit if

- The same task is done by hand every week
- Data is retyped between systems
- Sensitive data should not go to external AI services

### What we do not do

- No website chatbots
- No automation without oversight where mistakes are costly

## Frequently asked questions

### Does our data stay with us?

Yes, if you want. Automations and models run in your infrastructure or in a data centre in the EU.

### Do we need our own hardware for AI?

Not for simple workflows. Self-hosted language models need suitable GPUs; we handle the sizing.

### What does an automation cost?

After the first call you get a quote with fixed scope. Small workflows are often done in a few days.

## Related

- [Solutions](https://vetosec.at/en/solutions/): Nextcloud, SSO, backup, clusters
- [Managed services](https://vetosec.at/en/services/managed-services/): fixed price per service
- [Gap analysis & consulting](https://vetosec.at/en/services/gap-analysis/): ISO 27001, NIS2, GDPR

## Process automation, AI workflows and self-hosted LLMs in Austria

vetosec automates business processes in Austria, from invoice processing to self-hosted AI.

For defined tasks vetosec advises on specialised ML models, including hardware sizing and GDPR-compliant self-hosting.

Updated: October 19, 2026

Provider: vetosec, Dipl.-Ing. Samuel Ziegler, BSc, Berndorf, Lower Austria, Austria. Contact: https://vetosec.at/en/contact/